| File: | numpy/core/src/_simd/_simd_convert.inc |
| Warning: | line 87, column 25 PyObject ownership leak with reference count of 1 |
Press '?' to see keyboard shortcuts
Keyboard shortcuts:
| 1 | #line 1 "numpy/core/src/_simd/_simd.dispatch.c.src" | |||
| 2 | ||||
| 3 | /* | |||
| 4 | ***************************************************************************** | |||
| 5 | ** This file was autogenerated from a template DO NOT EDIT!!!! ** | |||
| 6 | ** Changes should be made to the original source (.src) file ** | |||
| 7 | ***************************************************************************** | |||
| 8 | */ | |||
| 9 | ||||
| 10 | #line 1 | |||
| 11 | /*@targets #simd_test*/ | |||
| 12 | #include "_simd.h" | |||
| 13 | #include "_simd_inc.h" | |||
| 14 | ||||
| 15 | #if NPY_SIMD128 | |||
| 16 | #include "_simd_data.inc" | |||
| 17 | #include "_simd_convert.inc" | |||
| 18 | #include "_simd_vector.inc" | |||
| 19 | #include "_simd_arg.inc" | |||
| 20 | #include "_simd_easyintrin.inc" | |||
| 21 | ||||
| 22 | //######################################################################### | |||
| 23 | //## Defining NPYV intrinsics as module functions | |||
| 24 | //######################################################################### | |||
| 25 | #line 34 | |||
| 26 | #if 1 | |||
| 27 | /*************************** | |||
| 28 | * Memory | |||
| 29 | ***************************/ | |||
| 30 | #line 41 | |||
| 31 | SIMD_IMPL_INTRIN_1(load_u8, vu8, qu8)static PyObject *simd__intrin_load_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu8}; if (!PyArg_ParseTuple( args, "O&:" "load_u8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu8 = npyv_load_u8( arg.data.qu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 32 | ||||
| 33 | #line 41 | |||
| 34 | SIMD_IMPL_INTRIN_1(loada_u8, vu8, qu8)static PyObject *simd__intrin_loada_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu8}; if (!PyArg_ParseTuple( args, "O&:" "loada_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu8 = npyv_loada_u8( arg.data.qu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 35 | ||||
| 36 | #line 41 | |||
| 37 | SIMD_IMPL_INTRIN_1(loads_u8, vu8, qu8)static PyObject *simd__intrin_loads_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu8}; if (!PyArg_ParseTuple( args, "O&:" "loads_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu8 = npyv_loads_u8( arg.data.qu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 38 | ||||
| 39 | #line 41 | |||
| 40 | SIMD_IMPL_INTRIN_1(loadl_u8, vu8, qu8)static PyObject *simd__intrin_loadl_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu8}; if (!PyArg_ParseTuple( args, "O&:" "loadl_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu8 = npyv_loadl_u8( arg.data.qu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 41 | ||||
| 42 | #line 46 | |||
| 43 | // special definition due to the nature of store | |||
| 44 | static PyObject * | |||
| 45 | simd__intrin_store_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 46 | { | |||
| 47 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 48 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 49 | if (!PyArg_ParseTuple( | |||
| 50 | args, "O&O&:store_u8", | |||
| 51 | simd_arg_converter, &seq_arg, | |||
| 52 | simd_arg_converter, &vec_arg | |||
| 53 | )) { | |||
| 54 | return NULL((void*)0); | |||
| 55 | } | |||
| 56 | npyv_store_u8(seq_arg.data.qu8, vec_arg.data.vu8); | |||
| 57 | // write-back | |||
| 58 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 59 | simd_arg_free(&seq_arg); | |||
| 60 | return NULL((void*)0); | |||
| 61 | } | |||
| 62 | simd_arg_free(&seq_arg); | |||
| 63 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 64 | } | |||
| 65 | ||||
| 66 | #line 46 | |||
| 67 | // special definition due to the nature of storea | |||
| 68 | static PyObject * | |||
| 69 | simd__intrin_storea_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 70 | { | |||
| 71 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 72 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 73 | if (!PyArg_ParseTuple( | |||
| 74 | args, "O&O&:storea_u8", | |||
| 75 | simd_arg_converter, &seq_arg, | |||
| 76 | simd_arg_converter, &vec_arg | |||
| 77 | )) { | |||
| 78 | return NULL((void*)0); | |||
| 79 | } | |||
| 80 | npyv_storea_u8(seq_arg.data.qu8, vec_arg.data.vu8); | |||
| 81 | // write-back | |||
| 82 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 83 | simd_arg_free(&seq_arg); | |||
| 84 | return NULL((void*)0); | |||
| 85 | } | |||
| 86 | simd_arg_free(&seq_arg); | |||
| 87 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 88 | } | |||
| 89 | ||||
| 90 | #line 46 | |||
| 91 | // special definition due to the nature of stores | |||
| 92 | static PyObject * | |||
| 93 | simd__intrin_stores_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 94 | { | |||
| 95 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 96 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 97 | if (!PyArg_ParseTuple( | |||
| 98 | args, "O&O&:stores_u8", | |||
| 99 | simd_arg_converter, &seq_arg, | |||
| 100 | simd_arg_converter, &vec_arg | |||
| 101 | )) { | |||
| 102 | return NULL((void*)0); | |||
| 103 | } | |||
| 104 | npyv_stores_u8(seq_arg.data.qu8, vec_arg.data.vu8); | |||
| 105 | // write-back | |||
| 106 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 107 | simd_arg_free(&seq_arg); | |||
| 108 | return NULL((void*)0); | |||
| 109 | } | |||
| 110 | simd_arg_free(&seq_arg); | |||
| 111 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 112 | } | |||
| 113 | ||||
| 114 | #line 46 | |||
| 115 | // special definition due to the nature of storel | |||
| 116 | static PyObject * | |||
| 117 | simd__intrin_storel_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 118 | { | |||
| 119 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 120 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 121 | if (!PyArg_ParseTuple( | |||
| 122 | args, "O&O&:storel_u8", | |||
| 123 | simd_arg_converter, &seq_arg, | |||
| 124 | simd_arg_converter, &vec_arg | |||
| 125 | )) { | |||
| 126 | return NULL((void*)0); | |||
| 127 | } | |||
| 128 | npyv_storel_u8(seq_arg.data.qu8, vec_arg.data.vu8); | |||
| 129 | // write-back | |||
| 130 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 131 | simd_arg_free(&seq_arg); | |||
| 132 | return NULL((void*)0); | |||
| 133 | } | |||
| 134 | simd_arg_free(&seq_arg); | |||
| 135 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 136 | } | |||
| 137 | ||||
| 138 | #line 46 | |||
| 139 | // special definition due to the nature of storeh | |||
| 140 | static PyObject * | |||
| 141 | simd__intrin_storeh_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 142 | { | |||
| 143 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 144 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 145 | if (!PyArg_ParseTuple( | |||
| 146 | args, "O&O&:storeh_u8", | |||
| 147 | simd_arg_converter, &seq_arg, | |||
| 148 | simd_arg_converter, &vec_arg | |||
| 149 | )) { | |||
| 150 | return NULL((void*)0); | |||
| 151 | } | |||
| 152 | npyv_storeh_u8(seq_arg.data.qu8, vec_arg.data.vu8); | |||
| 153 | // write-back | |||
| 154 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 155 | simd_arg_free(&seq_arg); | |||
| 156 | return NULL((void*)0); | |||
| 157 | } | |||
| 158 | simd_arg_free(&seq_arg); | |||
| 159 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 160 | } | |||
| 161 | ||||
| 162 | ||||
| 163 | /**************************************** | |||
| 164 | * Non-contiguous/Partial Memory access | |||
| 165 | ****************************************/ | |||
| 166 | #if 0 | |||
| 167 | // Partial Load | |||
| 168 | SIMD_IMPL_INTRIN_3(load_till_u8, vu8, qu8, u32, u8)static PyObject *simd__intrin_load_till_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu8}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_u8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_load_till_u8 ( arg1.data.qu8, arg2.data.u32, arg3.data.u8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 169 | SIMD_IMPL_INTRIN_2(load_tillz_u8, vu8, qu8, u32)static PyObject *simd__intrin_load_tillz_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu8}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_u8" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu8 = npyv_load_tillz_u8 ( arg1.data.qu8, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 170 | ||||
| 171 | // Partial Store | |||
| 172 | static PyObject * | |||
| 173 | simd__intrin_store_till_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 174 | { | |||
| 175 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 176 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 177 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 178 | if (!PyArg_ParseTuple( | |||
| 179 | args, "O&O&O&:store_till_u8", | |||
| 180 | simd_arg_converter, &seq_arg, | |||
| 181 | simd_arg_converter, &nlane_arg, | |||
| 182 | simd_arg_converter, &vec_arg | |||
| 183 | )) { | |||
| 184 | return NULL((void*)0); | |||
| 185 | } | |||
| 186 | npyv_store_till_u8( | |||
| 187 | seq_arg.data.qu8, nlane_arg.data.u32, vec_arg.data.vu8 | |||
| 188 | ); | |||
| 189 | // write-back | |||
| 190 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 191 | simd_arg_free(&seq_arg); | |||
| 192 | return NULL((void*)0); | |||
| 193 | } | |||
| 194 | simd_arg_free(&seq_arg); | |||
| 195 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 196 | } | |||
| 197 | ||||
| 198 | // Non-contiguous Load | |||
| 199 | #line 112 | |||
| 200 | static PyObject * | |||
| 201 | simd__intrin_loadn_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 202 | { | |||
| 203 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 204 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 205 | #if 0 | |||
| 206 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 207 | #endif // till | |||
| 208 | #if 0 | |||
| 209 | simd_arg fill_arg = {.dtype = simd_data_u8}; | |||
| 210 | #endif | |||
| 211 | if (!PyArg_ParseTuple( | |||
| 212 | args, "O&O&:loadn_u8", | |||
| 213 | simd_arg_converter, &seq_arg, | |||
| 214 | simd_arg_converter, &stride_arg | |||
| 215 | #if 0 | |||
| 216 | ,simd_arg_converter, &nlane_arg | |||
| 217 | #endif | |||
| 218 | #if 0 | |||
| 219 | ,simd_arg_converter, &fill_arg | |||
| 220 | #endif | |||
| 221 | )) { | |||
| 222 | return NULL((void*)0); | |||
| 223 | } | |||
| 224 | npyv_lanetype_u8 *seq_ptr = seq_arg.data.qu8; | |||
| 225 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 226 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 227 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u816; | |||
| 228 | if (stride < 0) { | |||
| 229 | seq_ptr += cur_seq_len -1; | |||
| 230 | min_seq_len = -min_seq_len; | |||
| 231 | } | |||
| 232 | if (cur_seq_len < min_seq_len) { | |||
| 233 | PyErr_Format(PyExc_ValueError, | |||
| 234 | "loadn_u8(), according to provided stride %d, the " | |||
| 235 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 236 | stride, min_seq_len, cur_seq_len | |||
| 237 | ); | |||
| 238 | goto err; | |||
| 239 | } | |||
| 240 | npyv_u8 rvec = npyv_loadn_u8( | |||
| 241 | seq_ptr, stride | |||
| 242 | #if 0 | |||
| 243 | , nlane_arg.data.u32 | |||
| 244 | #endif | |||
| 245 | #if 0 | |||
| 246 | , fill_arg.data.u8 | |||
| 247 | #endif | |||
| 248 | ); | |||
| 249 | simd_arg ret = { | |||
| 250 | .dtype = simd_data_vu8, .data = {.vu8=rvec} | |||
| 251 | }; | |||
| 252 | simd_arg_free(&seq_arg); | |||
| 253 | return simd_arg_to_obj(&ret); | |||
| 254 | err: | |||
| 255 | simd_arg_free(&seq_arg); | |||
| 256 | return NULL((void*)0); | |||
| 257 | } | |||
| 258 | ||||
| 259 | #line 112 | |||
| 260 | static PyObject * | |||
| 261 | simd__intrin_loadn_till_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 262 | { | |||
| 263 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 264 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 265 | #if 1 | |||
| 266 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 267 | #endif // till | |||
| 268 | #if 1 | |||
| 269 | simd_arg fill_arg = {.dtype = simd_data_u8}; | |||
| 270 | #endif | |||
| 271 | if (!PyArg_ParseTuple( | |||
| 272 | args, "O&O&O&O&:loadn_till_u8", | |||
| 273 | simd_arg_converter, &seq_arg, | |||
| 274 | simd_arg_converter, &stride_arg | |||
| 275 | #if 1 | |||
| 276 | ,simd_arg_converter, &nlane_arg | |||
| 277 | #endif | |||
| 278 | #if 1 | |||
| 279 | ,simd_arg_converter, &fill_arg | |||
| 280 | #endif | |||
| 281 | )) { | |||
| 282 | return NULL((void*)0); | |||
| 283 | } | |||
| 284 | npyv_lanetype_u8 *seq_ptr = seq_arg.data.qu8; | |||
| 285 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 286 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 287 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u816; | |||
| 288 | if (stride < 0) { | |||
| 289 | seq_ptr += cur_seq_len -1; | |||
| 290 | min_seq_len = -min_seq_len; | |||
| 291 | } | |||
| 292 | if (cur_seq_len < min_seq_len) { | |||
| 293 | PyErr_Format(PyExc_ValueError, | |||
| 294 | "loadn_till_u8(), according to provided stride %d, the " | |||
| 295 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 296 | stride, min_seq_len, cur_seq_len | |||
| 297 | ); | |||
| 298 | goto err; | |||
| 299 | } | |||
| 300 | npyv_u8 rvec = npyv_loadn_till_u8( | |||
| 301 | seq_ptr, stride | |||
| 302 | #if 1 | |||
| 303 | , nlane_arg.data.u32 | |||
| 304 | #endif | |||
| 305 | #if 1 | |||
| 306 | , fill_arg.data.u8 | |||
| 307 | #endif | |||
| 308 | ); | |||
| 309 | simd_arg ret = { | |||
| 310 | .dtype = simd_data_vu8, .data = {.vu8=rvec} | |||
| 311 | }; | |||
| 312 | simd_arg_free(&seq_arg); | |||
| 313 | return simd_arg_to_obj(&ret); | |||
| 314 | err: | |||
| 315 | simd_arg_free(&seq_arg); | |||
| 316 | return NULL((void*)0); | |||
| 317 | } | |||
| 318 | ||||
| 319 | #line 112 | |||
| 320 | static PyObject * | |||
| 321 | simd__intrin_loadn_tillz_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 322 | { | |||
| 323 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 324 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 325 | #if 1 | |||
| 326 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 327 | #endif // till | |||
| 328 | #if 0 | |||
| 329 | simd_arg fill_arg = {.dtype = simd_data_u8}; | |||
| 330 | #endif | |||
| 331 | if (!PyArg_ParseTuple( | |||
| 332 | args, "O&O&O&:loadn_tillz_u8", | |||
| 333 | simd_arg_converter, &seq_arg, | |||
| 334 | simd_arg_converter, &stride_arg | |||
| 335 | #if 1 | |||
| 336 | ,simd_arg_converter, &nlane_arg | |||
| 337 | #endif | |||
| 338 | #if 0 | |||
| 339 | ,simd_arg_converter, &fill_arg | |||
| 340 | #endif | |||
| 341 | )) { | |||
| 342 | return NULL((void*)0); | |||
| 343 | } | |||
| 344 | npyv_lanetype_u8 *seq_ptr = seq_arg.data.qu8; | |||
| 345 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 346 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 347 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u816; | |||
| 348 | if (stride < 0) { | |||
| 349 | seq_ptr += cur_seq_len -1; | |||
| 350 | min_seq_len = -min_seq_len; | |||
| 351 | } | |||
| 352 | if (cur_seq_len < min_seq_len) { | |||
| 353 | PyErr_Format(PyExc_ValueError, | |||
| 354 | "loadn_tillz_u8(), according to provided stride %d, the " | |||
| 355 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 356 | stride, min_seq_len, cur_seq_len | |||
| 357 | ); | |||
| 358 | goto err; | |||
| 359 | } | |||
| 360 | npyv_u8 rvec = npyv_loadn_tillz_u8( | |||
| 361 | seq_ptr, stride | |||
| 362 | #if 1 | |||
| 363 | , nlane_arg.data.u32 | |||
| 364 | #endif | |||
| 365 | #if 0 | |||
| 366 | , fill_arg.data.u8 | |||
| 367 | #endif | |||
| 368 | ); | |||
| 369 | simd_arg ret = { | |||
| 370 | .dtype = simd_data_vu8, .data = {.vu8=rvec} | |||
| 371 | }; | |||
| 372 | simd_arg_free(&seq_arg); | |||
| 373 | return simd_arg_to_obj(&ret); | |||
| 374 | err: | |||
| 375 | simd_arg_free(&seq_arg); | |||
| 376 | return NULL((void*)0); | |||
| 377 | } | |||
| 378 | ||||
| 379 | ||||
| 380 | // Non-contiguous Store | |||
| 381 | #line 178 | |||
| 382 | static PyObject * | |||
| 383 | simd__intrin_storen_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 384 | { | |||
| 385 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 386 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 387 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 388 | #if 0 | |||
| 389 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 390 | #endif | |||
| 391 | if (!PyArg_ParseTuple( | |||
| 392 | args, "O&O&O&:storen_u8", | |||
| 393 | simd_arg_converter, &seq_arg, | |||
| 394 | simd_arg_converter, &stride_arg | |||
| 395 | #if 0 | |||
| 396 | ,simd_arg_converter, &nlane_arg | |||
| 397 | #endif | |||
| 398 | ,simd_arg_converter, &vec_arg | |||
| 399 | )) { | |||
| 400 | return NULL((void*)0); | |||
| 401 | } | |||
| 402 | npyv_lanetype_u8 *seq_ptr = seq_arg.data.qu8; | |||
| 403 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 404 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 405 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u816; | |||
| 406 | if (stride < 0) { | |||
| 407 | seq_ptr += cur_seq_len -1; | |||
| 408 | min_seq_len = -min_seq_len; | |||
| 409 | } | |||
| 410 | // overflow guard | |||
| 411 | if (cur_seq_len < min_seq_len) { | |||
| 412 | PyErr_Format(PyExc_ValueError, | |||
| 413 | "storen_u8(), according to provided stride %d, the" | |||
| 414 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 415 | stride, min_seq_len, cur_seq_len | |||
| 416 | ); | |||
| 417 | goto err; | |||
| 418 | } | |||
| 419 | npyv_storen_u8( | |||
| 420 | seq_ptr, stride | |||
| 421 | #if 0 | |||
| 422 | ,nlane_arg.data.u32 | |||
| 423 | #endif | |||
| 424 | ,vec_arg.data.vu8 | |||
| 425 | ); | |||
| 426 | // write-back | |||
| 427 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 428 | goto err; | |||
| 429 | } | |||
| 430 | simd_arg_free(&seq_arg); | |||
| 431 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 432 | err: | |||
| 433 | simd_arg_free(&seq_arg); | |||
| 434 | return NULL((void*)0); | |||
| 435 | } | |||
| 436 | ||||
| 437 | #line 178 | |||
| 438 | static PyObject * | |||
| 439 | simd__intrin_storen_till_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 440 | { | |||
| 441 | simd_arg seq_arg = {.dtype = simd_data_qu8}; | |||
| 442 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 443 | simd_arg vec_arg = {.dtype = simd_data_vu8}; | |||
| 444 | #if 1 | |||
| 445 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 446 | #endif | |||
| 447 | if (!PyArg_ParseTuple( | |||
| 448 | args, "O&O&O&O&:storen_u8", | |||
| 449 | simd_arg_converter, &seq_arg, | |||
| 450 | simd_arg_converter, &stride_arg | |||
| 451 | #if 1 | |||
| 452 | ,simd_arg_converter, &nlane_arg | |||
| 453 | #endif | |||
| 454 | ,simd_arg_converter, &vec_arg | |||
| 455 | )) { | |||
| 456 | return NULL((void*)0); | |||
| 457 | } | |||
| 458 | npyv_lanetype_u8 *seq_ptr = seq_arg.data.qu8; | |||
| 459 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 460 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 461 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u816; | |||
| 462 | if (stride < 0) { | |||
| 463 | seq_ptr += cur_seq_len -1; | |||
| 464 | min_seq_len = -min_seq_len; | |||
| 465 | } | |||
| 466 | // overflow guard | |||
| 467 | if (cur_seq_len < min_seq_len) { | |||
| 468 | PyErr_Format(PyExc_ValueError, | |||
| 469 | "storen_till_u8(), according to provided stride %d, the" | |||
| 470 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 471 | stride, min_seq_len, cur_seq_len | |||
| 472 | ); | |||
| 473 | goto err; | |||
| 474 | } | |||
| 475 | npyv_storen_till_u8( | |||
| 476 | seq_ptr, stride | |||
| 477 | #if 1 | |||
| 478 | ,nlane_arg.data.u32 | |||
| 479 | #endif | |||
| 480 | ,vec_arg.data.vu8 | |||
| 481 | ); | |||
| 482 | // write-back | |||
| 483 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu8, simd_data_qu8)) { | |||
| 484 | goto err; | |||
| 485 | } | |||
| 486 | simd_arg_free(&seq_arg); | |||
| 487 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 488 | err: | |||
| 489 | simd_arg_free(&seq_arg); | |||
| 490 | return NULL((void*)0); | |||
| 491 | } | |||
| 492 | ||||
| 493 | #endif // 0 | |||
| 494 | ||||
| 495 | /*************************** | |||
| 496 | * Misc | |||
| 497 | ***************************/ | |||
| 498 | SIMD_IMPL_INTRIN_0(zero_u8, vu8)static PyObject *simd__intrin_zero_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_u8") ) return ((void*)0); simd_arg a = { .dtype = simd_data_vu8, .data = {.vu8 = _mm_setzero_si128()}, }; return simd_arg_to_obj(&a); } | |||
| 499 | SIMD_IMPL_INTRIN_1(setall_u8, vu8, u8)static PyObject *simd__intrin_setall_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u8}; if (!PyArg_ParseTuple( args, "O&:" "setall_u8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu8 = _mm_set1_epi8((char)(arg.data.u8) )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 500 | SIMD_IMPL_INTRIN_3(select_u8, vu8, vb8, vu8, vu8)static PyObject *simd__intrin_select_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_u8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_select_u8 ( arg1.data.vb8, arg2.data.vu8, arg3.data.vu8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 501 | ||||
| 502 | #line 246 | |||
| 503 | #if 1 | |||
| 504 | SIMD_IMPL_INTRIN_1(reinterpret_u8_u8, vu8, vu8)static PyObject *simd__intrin_reinterpret_u8_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "reinterpret_u8_u8", simd_arg_converter, &arg )) return ( (void*)0); simd_data data = {.vu8 = arg.data.vu8}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 505 | #endif // simd_sup2 | |||
| 506 | ||||
| 507 | #line 246 | |||
| 508 | #if 1 | |||
| 509 | SIMD_IMPL_INTRIN_1(reinterpret_s8_u8, vs8, vu8)static PyObject *simd__intrin_reinterpret_s8_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "reinterpret_s8_u8", simd_arg_converter, &arg )) return ( (void*)0); simd_data data = {.vs8 = arg.data.vu8}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 510 | #endif // simd_sup2 | |||
| 511 | ||||
| 512 | #line 246 | |||
| 513 | #if 1 | |||
| 514 | SIMD_IMPL_INTRIN_1(reinterpret_u16_u8, vu16, vu8)static PyObject *simd__intrin_reinterpret_u16_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 515 | #endif // simd_sup2 | |||
| 516 | ||||
| 517 | #line 246 | |||
| 518 | #if 1 | |||
| 519 | SIMD_IMPL_INTRIN_1(reinterpret_s16_u8, vs16, vu8)static PyObject *simd__intrin_reinterpret_s16_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 520 | #endif // simd_sup2 | |||
| 521 | ||||
| 522 | #line 246 | |||
| 523 | #if 1 | |||
| 524 | SIMD_IMPL_INTRIN_1(reinterpret_u32_u8, vu32, vu8)static PyObject *simd__intrin_reinterpret_u32_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 525 | #endif // simd_sup2 | |||
| 526 | ||||
| 527 | #line 246 | |||
| 528 | #if 1 | |||
| 529 | SIMD_IMPL_INTRIN_1(reinterpret_s32_u8, vs32, vu8)static PyObject *simd__intrin_reinterpret_s32_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 530 | #endif // simd_sup2 | |||
| 531 | ||||
| 532 | #line 246 | |||
| 533 | #if 1 | |||
| 534 | SIMD_IMPL_INTRIN_1(reinterpret_u64_u8, vu64, vu8)static PyObject *simd__intrin_reinterpret_u64_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 535 | #endif // simd_sup2 | |||
| 536 | ||||
| 537 | #line 246 | |||
| 538 | #if 1 | |||
| 539 | SIMD_IMPL_INTRIN_1(reinterpret_s64_u8, vs64, vu8)static PyObject *simd__intrin_reinterpret_s64_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vu8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 540 | #endif // simd_sup2 | |||
| 541 | ||||
| 542 | #line 246 | |||
| 543 | #if 1 | |||
| 544 | SIMD_IMPL_INTRIN_1(reinterpret_f32_u8, vf32, vu8)static PyObject *simd__intrin_reinterpret_f32_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vu8 )}; simd_arg_free(&arg); simd_arg ret = { . data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 545 | #endif // simd_sup2 | |||
| 546 | ||||
| 547 | #line 246 | |||
| 548 | #if NPY_SIMD_F641 | |||
| 549 | SIMD_IMPL_INTRIN_1(reinterpret_f64_u8, vf64, vu8)static PyObject *simd__intrin_reinterpret_f64_u8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_u8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vu8 )}; simd_arg_free(&arg); simd_arg ret = { . data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 550 | #endif // simd_sup2 | |||
| 551 | ||||
| 552 | ||||
| 553 | /** | |||
| 554 | * special definition due to the nature of intrinsics | |||
| 555 | * npyv_setf_u8 and npy_set_u8. | |||
| 556 | */ | |||
| 557 | #line 258 | |||
| 558 | static PyObject * | |||
| 559 | simd__intrin_setf_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 560 | { | |||
| 561 | npyv_lanetype_u8 *data = simd_sequence_from_iterable(args, simd_data_qu8, npyv_nlanes_u816); | |||
| 562 | if (data == NULL((void*)0)) { | |||
| 563 | return NULL((void*)0); | |||
| 564 | } | |||
| 565 | simd_data r = {.vu8 = npyv_setf_u8(npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 566 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 567 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 568 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 569 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 570 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 571 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 572 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 573 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 574 | data[64] // for setfnpyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 575 | )npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16]))}; | |||
| 576 | simd_sequence_free(data); | |||
| 577 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu8); | |||
| 578 | } | |||
| 579 | ||||
| 580 | #line 258 | |||
| 581 | static PyObject * | |||
| 582 | simd__intrin_set_u8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 583 | { | |||
| 584 | npyv_lanetype_u8 *data = simd_sequence_from_iterable(args, simd_data_qu8, npyv_nlanes_u816); | |||
| 585 | if (data == NULL((void*)0)) { | |||
| 586 | return NULL((void*)0); | |||
| 587 | } | |||
| 588 | simd_data r = {.vu8 = npyv_set_u8(npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 589 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 590 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 591 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 592 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 593 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 594 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 595 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 596 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 597 | data[64] // for setfnpyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 598 | )npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15]))}; | |||
| 599 | simd_sequence_free(data); | |||
| 600 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu8); | |||
| 601 | } | |||
| 602 | ||||
| 603 | ||||
| 604 | /*************************** | |||
| 605 | * Reorder | |||
| 606 | ***************************/ | |||
| 607 | #line 287 | |||
| 608 | SIMD_IMPL_INTRIN_2(combinel_u8, vu8, vu8, vu8)static PyObject *simd__intrin_combinel_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_u8", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu8 = _mm_unpacklo_epi64 ( arg1.data.vu8, arg2.data.vu8 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 609 | ||||
| 610 | #line 287 | |||
| 611 | SIMD_IMPL_INTRIN_2(combineh_u8, vu8, vu8, vu8)static PyObject *simd__intrin_combineh_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_u8", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu8 = _mm_unpackhi_epi64 ( arg1.data.vu8, arg2.data.vu8 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 612 | ||||
| 613 | ||||
| 614 | #line 293 | |||
| 615 | SIMD_IMPL_INTRIN_2(combine_u8, vu8x2, vu8, vu8)static PyObject *simd__intrin_combine_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8x2 = npyv__combine( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8x2 }; return simd_arg_to_obj(&ret); } | |||
| 616 | ||||
| 617 | #line 293 | |||
| 618 | SIMD_IMPL_INTRIN_2(zip_u8, vu8x2, vu8, vu8)static PyObject *simd__intrin_zip_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8x2 = npyv_zip_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8x2 }; return simd_arg_to_obj(&ret); } | |||
| 619 | ||||
| 620 | ||||
| 621 | #if 1 | |||
| 622 | SIMD_IMPL_INTRIN_1(rev64_u8, vu8, vu8)static PyObject *simd__intrin_rev64_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "rev64_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu8 = npyv_rev64_u8( arg.data.vu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 623 | #endif | |||
| 624 | ||||
| 625 | /*************************** | |||
| 626 | * Operators | |||
| 627 | ***************************/ | |||
| 628 | #if 0 > 0 | |||
| 629 | SIMD_IMPL_INTRIN_2(shl_u8, vu8, vu8, u8)static PyObject *simd__intrin_shl_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_shl_u8( arg1.data.vu8, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 630 | SIMD_IMPL_INTRIN_2(shr_u8, vu8, vu8, u8)static PyObject *simd__intrin_shr_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_shr_u8( arg1.data.vu8, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 631 | // immediate constant | |||
| 632 | SIMD_IMPL_INTRIN_2IMM(shli_u8, vu8, vu8, 0)static PyObject *simd__intrin_shli_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu8 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shli_u8, vu8 ) data.vu8; simd_arg_free(& arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu8 } ; return simd_arg_to_obj(&ret); } | |||
| 633 | SIMD_IMPL_INTRIN_2IMM(shri_u8, vu8, vu8, 0)static PyObject *simd__intrin_shri_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu8 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shri_u8, vu8 ) data.vu8; simd_arg_free(& arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu8 } ; return simd_arg_to_obj(&ret); } | |||
| 634 | #endif // shl_imm | |||
| 635 | ||||
| 636 | #line 314 | |||
| 637 | SIMD_IMPL_INTRIN_2(and_u8, vu8, vu8, vu8)static PyObject *simd__intrin_and_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""and_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_and_si128( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 638 | ||||
| 639 | #line 314 | |||
| 640 | SIMD_IMPL_INTRIN_2(or_u8, vu8, vu8, vu8)static PyObject *simd__intrin_or_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""or_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_or_si128( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 641 | ||||
| 642 | #line 314 | |||
| 643 | SIMD_IMPL_INTRIN_2(xor_u8, vu8, vu8, vu8)static PyObject *simd__intrin_xor_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_xor_si128( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 644 | ||||
| 645 | ||||
| 646 | SIMD_IMPL_INTRIN_1(not_u8, vu8, vu8)static PyObject *simd__intrin_not_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "not_u8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu8 = _mm_xor_si128(arg.data.vu8, _mm_set1_epi32(-1 ))}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 647 | ||||
| 648 | #line 322 | |||
| 649 | SIMD_IMPL_INTRIN_2(cmpeq_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmpeq_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_cmpeq_epi8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 650 | ||||
| 651 | #line 322 | |||
| 652 | SIMD_IMPL_INTRIN_2(cmpneq_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmpneq_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_xor_si128(_mm_cmpeq_epi8(arg1 .data.vu8, arg2.data.vu8), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(& ret); } | |||
| 653 | ||||
| 654 | #line 322 | |||
| 655 | SIMD_IMPL_INTRIN_2(cmpgt_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmpgt_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = npyv_cmpgt_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 656 | ||||
| 657 | #line 322 | |||
| 658 | SIMD_IMPL_INTRIN_2(cmpge_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmpge_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = npyv_cmpge_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 659 | ||||
| 660 | #line 322 | |||
| 661 | SIMD_IMPL_INTRIN_2(cmplt_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmplt_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = npyv_cmpgt_u8(arg2.data.vu8, arg1 .data.vu8)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 662 | ||||
| 663 | #line 322 | |||
| 664 | SIMD_IMPL_INTRIN_2(cmple_u8, vb8, vu8, vu8)static PyObject *simd__intrin_cmple_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = npyv_cmpge_u8(arg2.data.vu8, arg1 .data.vu8)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 665 | ||||
| 666 | ||||
| 667 | /*************************** | |||
| 668 | * Conversion | |||
| 669 | ***************************/ | |||
| 670 | SIMD_IMPL_INTRIN_1(cvt_u8_b8, vu8, vb8)static PyObject *simd__intrin_cvt_u8_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb8}; if (!PyArg_ParseTuple( args, "O&:" "cvt_u8_b8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu8 = arg.data.vb8}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 } ; return simd_arg_to_obj(&ret); } | |||
| 671 | SIMD_IMPL_INTRIN_1(cvt_b8_u8, vb8, vu8)static PyObject *simd__intrin_cvt_b8_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b8_u8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vb8 = arg.data.vu8}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vb8 } ; return simd_arg_to_obj(&ret); } | |||
| 672 | #if 1 | |||
| 673 | SIMD_IMPL_INTRIN_1(expand_u16_u8, vu16x2, vu8)static PyObject *simd__intrin_expand_u16_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "expand_u16_u8", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vu16x2 = npyv_expand_u16_u8( arg.data .vu8 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16x2 }; return simd_arg_to_obj(&ret ); } | |||
| 674 | #endif // expand_sup | |||
| 675 | /*************************** | |||
| 676 | * Arithmetic | |||
| 677 | ***************************/ | |||
| 678 | #line 339 | |||
| 679 | SIMD_IMPL_INTRIN_2(add_u8, vu8, vu8, vu8)static PyObject *simd__intrin_add_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""add_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_add_epi8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 680 | ||||
| 681 | #line 339 | |||
| 682 | SIMD_IMPL_INTRIN_2(sub_u8, vu8, vu8, vu8)static PyObject *simd__intrin_sub_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_sub_epi8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 683 | ||||
| 684 | ||||
| 685 | #if 1 | |||
| 686 | #line 346 | |||
| 687 | SIMD_IMPL_INTRIN_2(adds_u8, vu8, vu8, vu8)static PyObject *simd__intrin_adds_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_adds_epu8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 688 | ||||
| 689 | #line 346 | |||
| 690 | SIMD_IMPL_INTRIN_2(subs_u8, vu8, vu8, vu8)static PyObject *simd__intrin_subs_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_subs_epu8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 691 | ||||
| 692 | #endif // sat_sup | |||
| 693 | ||||
| 694 | #if 1 | |||
| 695 | SIMD_IMPL_INTRIN_2(mul_u8, vu8, vu8, vu8)static PyObject *simd__intrin_mul_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_mul_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 696 | #endif // mul_sup | |||
| 697 | ||||
| 698 | #if 0 | |||
| 699 | SIMD_IMPL_INTRIN_2(div_u8, vu8, vu8, vu8)static PyObject *simd__intrin_div_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""div_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_div_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 700 | #endif // div_sup | |||
| 701 | ||||
| 702 | #if 1 | |||
| 703 | SIMD_IMPL_INTRIN_1(divisor_u8, vu8x3, u8)static PyObject *simd__intrin_divisor_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u8}; if (!PyArg_ParseTuple( args, "O&:" "divisor_u8", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu8x3 = npyv_divisor_u8( arg.data.u8 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8x3 }; return simd_arg_to_obj(&ret); } | |||
| 704 | SIMD_IMPL_INTRIN_2(divc_u8, vu8, vu8, vu8x3)static PyObject *simd__intrin_divc_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_divc_u8( arg1.data.vu8, arg2 .data.vu8x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu8 } ; return simd_arg_to_obj(&ret); } | |||
| 705 | #endif // intdiv_sup | |||
| 706 | ||||
| 707 | #if 0 | |||
| 708 | #line 367 | |||
| 709 | SIMD_IMPL_INTRIN_3(muladd_u8, vu8, vu8, vu8, vu8)static PyObject *simd__intrin_muladd_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_u8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_muladd_u8 ( arg1.data.vu8, arg2.data.vu8, arg3.data.vu8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 710 | ||||
| 711 | #line 367 | |||
| 712 | SIMD_IMPL_INTRIN_3(mulsub_u8, vu8, vu8, vu8, vu8)static PyObject *simd__intrin_mulsub_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_u8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_mulsub_u8 ( arg1.data.vu8, arg2.data.vu8, arg3.data.vu8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 713 | ||||
| 714 | #line 367 | |||
| 715 | SIMD_IMPL_INTRIN_3(nmuladd_u8, vu8, vu8, vu8, vu8)static PyObject *simd__intrin_nmuladd_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_nmuladd_u8 ( arg1.data.vu8, arg2.data.vu8, arg3.data.vu8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 716 | ||||
| 717 | #line 367 | |||
| 718 | SIMD_IMPL_INTRIN_3(nmulsub_u8, vu8, vu8, vu8, vu8)static PyObject *simd__intrin_nmulsub_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu8 = npyv_nmulsub_u8 ( arg1.data.vu8, arg2.data.vu8, arg3.data.vu8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 719 | ||||
| 720 | #endif // fused_sup | |||
| 721 | ||||
| 722 | #if 0 | |||
| 723 | SIMD_IMPL_INTRIN_1(sum_u8, u8, vu8)static PyObject *simd__intrin_sum_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "sum_u8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.u8 = npyv_sum_u8( arg.data.vu8 )}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_u8 }; return simd_arg_to_obj(&ret); } | |||
| 724 | #endif // sum_sup | |||
| 725 | ||||
| 726 | #if 1 | |||
| 727 | SIMD_IMPL_INTRIN_1(sumup_u8, u16, vu8)static PyObject *simd__intrin_sumup_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "sumup_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.u16 = npyv_sumup_u8( arg.data.vu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_u16 }; return simd_arg_to_obj(&ret); } | |||
| 728 | #endif // sumup_sup | |||
| 729 | ||||
| 730 | /*************************** | |||
| 731 | * Math | |||
| 732 | ***************************/ | |||
| 733 | #if 0 | |||
| 734 | #line 386 | |||
| 735 | SIMD_IMPL_INTRIN_1(sqrt_u8, vu8, vu8)static PyObject *simd__intrin_sqrt_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_u8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu8 = npyv_sqrt_u8( arg.data.vu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 736 | ||||
| 737 | #line 386 | |||
| 738 | SIMD_IMPL_INTRIN_1(recip_u8, vu8, vu8)static PyObject *simd__intrin_recip_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "recip_u8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu8 = npyv_recip_u8( arg.data.vu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 739 | ||||
| 740 | #line 386 | |||
| 741 | SIMD_IMPL_INTRIN_1(abs_u8, vu8, vu8)static PyObject *simd__intrin_abs_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "abs_u8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu8 = npyv_abs_u8( arg.data.vu8 )}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 } ; return simd_arg_to_obj(&ret); } | |||
| 742 | ||||
| 743 | #line 386 | |||
| 744 | SIMD_IMPL_INTRIN_1(square_u8, vu8, vu8)static PyObject *simd__intrin_square_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&:" "square_u8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu8 = npyv_square_u8( arg.data.vu8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 745 | ||||
| 746 | #endif | |||
| 747 | ||||
| 748 | #line 393 | |||
| 749 | SIMD_IMPL_INTRIN_2(max_u8, vu8, vu8, vu8)static PyObject *simd__intrin_max_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""max_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_max_epu8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 750 | ||||
| 751 | #line 393 | |||
| 752 | SIMD_IMPL_INTRIN_2(min_u8, vu8, vu8, vu8)static PyObject *simd__intrin_min_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""min_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = _mm_min_epu8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 753 | ||||
| 754 | ||||
| 755 | #if 0 | |||
| 756 | #line 400 | |||
| 757 | SIMD_IMPL_INTRIN_2(maxp_u8, vu8, vu8, vu8)static PyObject *simd__intrin_maxp_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_maxp_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 758 | ||||
| 759 | #line 400 | |||
| 760 | SIMD_IMPL_INTRIN_2(minp_u8, vu8, vu8, vu8)static PyObject *simd__intrin_minp_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_u8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu8 = npyv_minp_u8( arg1.data.vu8, arg2 .data.vu8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 761 | ||||
| 762 | #endif | |||
| 763 | ||||
| 764 | /*************************** | |||
| 765 | * Mask operations | |||
| 766 | ***************************/ | |||
| 767 | #line 410 | |||
| 768 | SIMD_IMPL_INTRIN_4(ifadd_u8, vu8, vb8, vu8, vu8, vu8)static PyObject *simd__intrin_ifadd_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; simd_arg arg4 = { .dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_u8", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu8 = npyv_ifadd_u8 ( arg1.data.vb8, arg2.data.vu8, arg3.data.vu8, arg4.data.vu8 ) }; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(& ret); } | |||
| 769 | ||||
| 770 | #line 410 | |||
| 771 | SIMD_IMPL_INTRIN_4(ifsub_u8, vu8, vb8, vu8, vu8, vu8)static PyObject *simd__intrin_ifsub_u8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vu8 }; simd_arg arg3 = {.dtype = simd_data_vu8}; simd_arg arg4 = { .dtype = simd_data_vu8}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_u8", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu8 = npyv_ifsub_u8 ( arg1.data.vb8, arg2.data.vu8, arg3.data.vu8, arg4.data.vu8 ) }; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(& ret); } | |||
| 772 | ||||
| 773 | ||||
| 774 | #endif // simd_sup | |||
| 775 | ||||
| 776 | #line 34 | |||
| 777 | #if 1 | |||
| 778 | /*************************** | |||
| 779 | * Memory | |||
| 780 | ***************************/ | |||
| 781 | #line 41 | |||
| 782 | SIMD_IMPL_INTRIN_1(load_s8, vs8, qs8)static PyObject *simd__intrin_load_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs8}; if (!PyArg_ParseTuple( args, "O&:" "load_s8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs8 = npyv_load_s8( arg.data.qs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 783 | ||||
| 784 | #line 41 | |||
| 785 | SIMD_IMPL_INTRIN_1(loada_s8, vs8, qs8)static PyObject *simd__intrin_loada_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs8}; if (!PyArg_ParseTuple( args, "O&:" "loada_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs8 = npyv_loada_s8( arg.data.qs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 786 | ||||
| 787 | #line 41 | |||
| 788 | SIMD_IMPL_INTRIN_1(loads_s8, vs8, qs8)static PyObject *simd__intrin_loads_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs8}; if (!PyArg_ParseTuple( args, "O&:" "loads_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs8 = npyv_loads_s8( arg.data.qs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 789 | ||||
| 790 | #line 41 | |||
| 791 | SIMD_IMPL_INTRIN_1(loadl_s8, vs8, qs8)static PyObject *simd__intrin_loadl_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs8}; if (!PyArg_ParseTuple( args, "O&:" "loadl_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs8 = npyv_loadl_s8( arg.data.qs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 792 | ||||
| 793 | #line 46 | |||
| 794 | // special definition due to the nature of store | |||
| 795 | static PyObject * | |||
| 796 | simd__intrin_store_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 797 | { | |||
| 798 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 799 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 800 | if (!PyArg_ParseTuple( | |||
| 801 | args, "O&O&:store_s8", | |||
| 802 | simd_arg_converter, &seq_arg, | |||
| 803 | simd_arg_converter, &vec_arg | |||
| 804 | )) { | |||
| 805 | return NULL((void*)0); | |||
| 806 | } | |||
| 807 | npyv_store_s8(seq_arg.data.qs8, vec_arg.data.vs8); | |||
| 808 | // write-back | |||
| 809 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 810 | simd_arg_free(&seq_arg); | |||
| 811 | return NULL((void*)0); | |||
| 812 | } | |||
| 813 | simd_arg_free(&seq_arg); | |||
| 814 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 815 | } | |||
| 816 | ||||
| 817 | #line 46 | |||
| 818 | // special definition due to the nature of storea | |||
| 819 | static PyObject * | |||
| 820 | simd__intrin_storea_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 821 | { | |||
| 822 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 823 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 824 | if (!PyArg_ParseTuple( | |||
| 825 | args, "O&O&:storea_s8", | |||
| 826 | simd_arg_converter, &seq_arg, | |||
| 827 | simd_arg_converter, &vec_arg | |||
| 828 | )) { | |||
| 829 | return NULL((void*)0); | |||
| 830 | } | |||
| 831 | npyv_storea_s8(seq_arg.data.qs8, vec_arg.data.vs8); | |||
| 832 | // write-back | |||
| 833 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 834 | simd_arg_free(&seq_arg); | |||
| 835 | return NULL((void*)0); | |||
| 836 | } | |||
| 837 | simd_arg_free(&seq_arg); | |||
| 838 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 839 | } | |||
| 840 | ||||
| 841 | #line 46 | |||
| 842 | // special definition due to the nature of stores | |||
| 843 | static PyObject * | |||
| 844 | simd__intrin_stores_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 845 | { | |||
| 846 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 847 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 848 | if (!PyArg_ParseTuple( | |||
| 849 | args, "O&O&:stores_s8", | |||
| 850 | simd_arg_converter, &seq_arg, | |||
| 851 | simd_arg_converter, &vec_arg | |||
| 852 | )) { | |||
| 853 | return NULL((void*)0); | |||
| 854 | } | |||
| 855 | npyv_stores_s8(seq_arg.data.qs8, vec_arg.data.vs8); | |||
| 856 | // write-back | |||
| 857 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 858 | simd_arg_free(&seq_arg); | |||
| 859 | return NULL((void*)0); | |||
| 860 | } | |||
| 861 | simd_arg_free(&seq_arg); | |||
| 862 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 863 | } | |||
| 864 | ||||
| 865 | #line 46 | |||
| 866 | // special definition due to the nature of storel | |||
| 867 | static PyObject * | |||
| 868 | simd__intrin_storel_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 869 | { | |||
| 870 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 871 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 872 | if (!PyArg_ParseTuple( | |||
| 873 | args, "O&O&:storel_s8", | |||
| 874 | simd_arg_converter, &seq_arg, | |||
| 875 | simd_arg_converter, &vec_arg | |||
| 876 | )) { | |||
| 877 | return NULL((void*)0); | |||
| 878 | } | |||
| 879 | npyv_storel_s8(seq_arg.data.qs8, vec_arg.data.vs8); | |||
| 880 | // write-back | |||
| 881 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 882 | simd_arg_free(&seq_arg); | |||
| 883 | return NULL((void*)0); | |||
| 884 | } | |||
| 885 | simd_arg_free(&seq_arg); | |||
| 886 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 887 | } | |||
| 888 | ||||
| 889 | #line 46 | |||
| 890 | // special definition due to the nature of storeh | |||
| 891 | static PyObject * | |||
| 892 | simd__intrin_storeh_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 893 | { | |||
| 894 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 895 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 896 | if (!PyArg_ParseTuple( | |||
| 897 | args, "O&O&:storeh_s8", | |||
| 898 | simd_arg_converter, &seq_arg, | |||
| 899 | simd_arg_converter, &vec_arg | |||
| 900 | )) { | |||
| 901 | return NULL((void*)0); | |||
| 902 | } | |||
| 903 | npyv_storeh_s8(seq_arg.data.qs8, vec_arg.data.vs8); | |||
| 904 | // write-back | |||
| 905 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 906 | simd_arg_free(&seq_arg); | |||
| 907 | return NULL((void*)0); | |||
| 908 | } | |||
| 909 | simd_arg_free(&seq_arg); | |||
| 910 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 911 | } | |||
| 912 | ||||
| 913 | ||||
| 914 | /**************************************** | |||
| 915 | * Non-contiguous/Partial Memory access | |||
| 916 | ****************************************/ | |||
| 917 | #if 0 | |||
| 918 | // Partial Load | |||
| 919 | SIMD_IMPL_INTRIN_3(load_till_s8, vs8, qs8, u32, s8)static PyObject *simd__intrin_load_till_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs8}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_s8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_load_till_s8 ( arg1.data.qs8, arg2.data.u32, arg3.data.s8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 920 | SIMD_IMPL_INTRIN_2(load_tillz_s8, vs8, qs8, u32)static PyObject *simd__intrin_load_tillz_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs8}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_s8" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs8 = npyv_load_tillz_s8 ( arg1.data.qs8, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 921 | ||||
| 922 | // Partial Store | |||
| 923 | static PyObject * | |||
| 924 | simd__intrin_store_till_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 925 | { | |||
| 926 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 927 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 928 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 929 | if (!PyArg_ParseTuple( | |||
| 930 | args, "O&O&O&:store_till_s8", | |||
| 931 | simd_arg_converter, &seq_arg, | |||
| 932 | simd_arg_converter, &nlane_arg, | |||
| 933 | simd_arg_converter, &vec_arg | |||
| 934 | )) { | |||
| 935 | return NULL((void*)0); | |||
| 936 | } | |||
| 937 | npyv_store_till_s8( | |||
| 938 | seq_arg.data.qs8, nlane_arg.data.u32, vec_arg.data.vs8 | |||
| 939 | ); | |||
| 940 | // write-back | |||
| 941 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 942 | simd_arg_free(&seq_arg); | |||
| 943 | return NULL((void*)0); | |||
| 944 | } | |||
| 945 | simd_arg_free(&seq_arg); | |||
| 946 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 947 | } | |||
| 948 | ||||
| 949 | // Non-contiguous Load | |||
| 950 | #line 112 | |||
| 951 | static PyObject * | |||
| 952 | simd__intrin_loadn_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 953 | { | |||
| 954 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 955 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 956 | #if 0 | |||
| 957 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 958 | #endif // till | |||
| 959 | #if 0 | |||
| 960 | simd_arg fill_arg = {.dtype = simd_data_s8}; | |||
| 961 | #endif | |||
| 962 | if (!PyArg_ParseTuple( | |||
| 963 | args, "O&O&:loadn_s8", | |||
| 964 | simd_arg_converter, &seq_arg, | |||
| 965 | simd_arg_converter, &stride_arg | |||
| 966 | #if 0 | |||
| 967 | ,simd_arg_converter, &nlane_arg | |||
| 968 | #endif | |||
| 969 | #if 0 | |||
| 970 | ,simd_arg_converter, &fill_arg | |||
| 971 | #endif | |||
| 972 | )) { | |||
| 973 | return NULL((void*)0); | |||
| 974 | } | |||
| 975 | npyv_lanetype_s8 *seq_ptr = seq_arg.data.qs8; | |||
| 976 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 977 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 978 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s816; | |||
| 979 | if (stride < 0) { | |||
| 980 | seq_ptr += cur_seq_len -1; | |||
| 981 | min_seq_len = -min_seq_len; | |||
| 982 | } | |||
| 983 | if (cur_seq_len < min_seq_len) { | |||
| 984 | PyErr_Format(PyExc_ValueError, | |||
| 985 | "loadn_s8(), according to provided stride %d, the " | |||
| 986 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 987 | stride, min_seq_len, cur_seq_len | |||
| 988 | ); | |||
| 989 | goto err; | |||
| 990 | } | |||
| 991 | npyv_s8 rvec = npyv_loadn_s8( | |||
| 992 | seq_ptr, stride | |||
| 993 | #if 0 | |||
| 994 | , nlane_arg.data.u32 | |||
| 995 | #endif | |||
| 996 | #if 0 | |||
| 997 | , fill_arg.data.s8 | |||
| 998 | #endif | |||
| 999 | ); | |||
| 1000 | simd_arg ret = { | |||
| 1001 | .dtype = simd_data_vs8, .data = {.vs8=rvec} | |||
| 1002 | }; | |||
| 1003 | simd_arg_free(&seq_arg); | |||
| 1004 | return simd_arg_to_obj(&ret); | |||
| 1005 | err: | |||
| 1006 | simd_arg_free(&seq_arg); | |||
| 1007 | return NULL((void*)0); | |||
| 1008 | } | |||
| 1009 | ||||
| 1010 | #line 112 | |||
| 1011 | static PyObject * | |||
| 1012 | simd__intrin_loadn_till_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1013 | { | |||
| 1014 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 1015 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1016 | #if 1 | |||
| 1017 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1018 | #endif // till | |||
| 1019 | #if 1 | |||
| 1020 | simd_arg fill_arg = {.dtype = simd_data_s8}; | |||
| 1021 | #endif | |||
| 1022 | if (!PyArg_ParseTuple( | |||
| 1023 | args, "O&O&O&O&:loadn_till_s8", | |||
| 1024 | simd_arg_converter, &seq_arg, | |||
| 1025 | simd_arg_converter, &stride_arg | |||
| 1026 | #if 1 | |||
| 1027 | ,simd_arg_converter, &nlane_arg | |||
| 1028 | #endif | |||
| 1029 | #if 1 | |||
| 1030 | ,simd_arg_converter, &fill_arg | |||
| 1031 | #endif | |||
| 1032 | )) { | |||
| 1033 | return NULL((void*)0); | |||
| 1034 | } | |||
| 1035 | npyv_lanetype_s8 *seq_ptr = seq_arg.data.qs8; | |||
| 1036 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1037 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1038 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s816; | |||
| 1039 | if (stride < 0) { | |||
| 1040 | seq_ptr += cur_seq_len -1; | |||
| 1041 | min_seq_len = -min_seq_len; | |||
| 1042 | } | |||
| 1043 | if (cur_seq_len < min_seq_len) { | |||
| 1044 | PyErr_Format(PyExc_ValueError, | |||
| 1045 | "loadn_till_s8(), according to provided stride %d, the " | |||
| 1046 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1047 | stride, min_seq_len, cur_seq_len | |||
| 1048 | ); | |||
| 1049 | goto err; | |||
| 1050 | } | |||
| 1051 | npyv_s8 rvec = npyv_loadn_till_s8( | |||
| 1052 | seq_ptr, stride | |||
| 1053 | #if 1 | |||
| 1054 | , nlane_arg.data.u32 | |||
| 1055 | #endif | |||
| 1056 | #if 1 | |||
| 1057 | , fill_arg.data.s8 | |||
| 1058 | #endif | |||
| 1059 | ); | |||
| 1060 | simd_arg ret = { | |||
| 1061 | .dtype = simd_data_vs8, .data = {.vs8=rvec} | |||
| 1062 | }; | |||
| 1063 | simd_arg_free(&seq_arg); | |||
| 1064 | return simd_arg_to_obj(&ret); | |||
| 1065 | err: | |||
| 1066 | simd_arg_free(&seq_arg); | |||
| 1067 | return NULL((void*)0); | |||
| 1068 | } | |||
| 1069 | ||||
| 1070 | #line 112 | |||
| 1071 | static PyObject * | |||
| 1072 | simd__intrin_loadn_tillz_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1073 | { | |||
| 1074 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 1075 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1076 | #if 1 | |||
| 1077 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1078 | #endif // till | |||
| 1079 | #if 0 | |||
| 1080 | simd_arg fill_arg = {.dtype = simd_data_s8}; | |||
| 1081 | #endif | |||
| 1082 | if (!PyArg_ParseTuple( | |||
| 1083 | args, "O&O&O&:loadn_tillz_s8", | |||
| 1084 | simd_arg_converter, &seq_arg, | |||
| 1085 | simd_arg_converter, &stride_arg | |||
| 1086 | #if 1 | |||
| 1087 | ,simd_arg_converter, &nlane_arg | |||
| 1088 | #endif | |||
| 1089 | #if 0 | |||
| 1090 | ,simd_arg_converter, &fill_arg | |||
| 1091 | #endif | |||
| 1092 | )) { | |||
| 1093 | return NULL((void*)0); | |||
| 1094 | } | |||
| 1095 | npyv_lanetype_s8 *seq_ptr = seq_arg.data.qs8; | |||
| 1096 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1097 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1098 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s816; | |||
| 1099 | if (stride < 0) { | |||
| 1100 | seq_ptr += cur_seq_len -1; | |||
| 1101 | min_seq_len = -min_seq_len; | |||
| 1102 | } | |||
| 1103 | if (cur_seq_len < min_seq_len) { | |||
| 1104 | PyErr_Format(PyExc_ValueError, | |||
| 1105 | "loadn_tillz_s8(), according to provided stride %d, the " | |||
| 1106 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1107 | stride, min_seq_len, cur_seq_len | |||
| 1108 | ); | |||
| 1109 | goto err; | |||
| 1110 | } | |||
| 1111 | npyv_s8 rvec = npyv_loadn_tillz_s8( | |||
| 1112 | seq_ptr, stride | |||
| 1113 | #if 1 | |||
| 1114 | , nlane_arg.data.u32 | |||
| 1115 | #endif | |||
| 1116 | #if 0 | |||
| 1117 | , fill_arg.data.s8 | |||
| 1118 | #endif | |||
| 1119 | ); | |||
| 1120 | simd_arg ret = { | |||
| 1121 | .dtype = simd_data_vs8, .data = {.vs8=rvec} | |||
| 1122 | }; | |||
| 1123 | simd_arg_free(&seq_arg); | |||
| 1124 | return simd_arg_to_obj(&ret); | |||
| 1125 | err: | |||
| 1126 | simd_arg_free(&seq_arg); | |||
| 1127 | return NULL((void*)0); | |||
| 1128 | } | |||
| 1129 | ||||
| 1130 | ||||
| 1131 | // Non-contiguous Store | |||
| 1132 | #line 178 | |||
| 1133 | static PyObject * | |||
| 1134 | simd__intrin_storen_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1135 | { | |||
| 1136 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 1137 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1138 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 1139 | #if 0 | |||
| 1140 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1141 | #endif | |||
| 1142 | if (!PyArg_ParseTuple( | |||
| 1143 | args, "O&O&O&:storen_s8", | |||
| 1144 | simd_arg_converter, &seq_arg, | |||
| 1145 | simd_arg_converter, &stride_arg | |||
| 1146 | #if 0 | |||
| 1147 | ,simd_arg_converter, &nlane_arg | |||
| 1148 | #endif | |||
| 1149 | ,simd_arg_converter, &vec_arg | |||
| 1150 | )) { | |||
| 1151 | return NULL((void*)0); | |||
| 1152 | } | |||
| 1153 | npyv_lanetype_s8 *seq_ptr = seq_arg.data.qs8; | |||
| 1154 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1155 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1156 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s816; | |||
| 1157 | if (stride < 0) { | |||
| 1158 | seq_ptr += cur_seq_len -1; | |||
| 1159 | min_seq_len = -min_seq_len; | |||
| 1160 | } | |||
| 1161 | // overflow guard | |||
| 1162 | if (cur_seq_len < min_seq_len) { | |||
| 1163 | PyErr_Format(PyExc_ValueError, | |||
| 1164 | "storen_s8(), according to provided stride %d, the" | |||
| 1165 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1166 | stride, min_seq_len, cur_seq_len | |||
| 1167 | ); | |||
| 1168 | goto err; | |||
| 1169 | } | |||
| 1170 | npyv_storen_s8( | |||
| 1171 | seq_ptr, stride | |||
| 1172 | #if 0 | |||
| 1173 | ,nlane_arg.data.u32 | |||
| 1174 | #endif | |||
| 1175 | ,vec_arg.data.vs8 | |||
| 1176 | ); | |||
| 1177 | // write-back | |||
| 1178 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 1179 | goto err; | |||
| 1180 | } | |||
| 1181 | simd_arg_free(&seq_arg); | |||
| 1182 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1183 | err: | |||
| 1184 | simd_arg_free(&seq_arg); | |||
| 1185 | return NULL((void*)0); | |||
| 1186 | } | |||
| 1187 | ||||
| 1188 | #line 178 | |||
| 1189 | static PyObject * | |||
| 1190 | simd__intrin_storen_till_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1191 | { | |||
| 1192 | simd_arg seq_arg = {.dtype = simd_data_qs8}; | |||
| 1193 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1194 | simd_arg vec_arg = {.dtype = simd_data_vs8}; | |||
| 1195 | #if 1 | |||
| 1196 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1197 | #endif | |||
| 1198 | if (!PyArg_ParseTuple( | |||
| 1199 | args, "O&O&O&O&:storen_s8", | |||
| 1200 | simd_arg_converter, &seq_arg, | |||
| 1201 | simd_arg_converter, &stride_arg | |||
| 1202 | #if 1 | |||
| 1203 | ,simd_arg_converter, &nlane_arg | |||
| 1204 | #endif | |||
| 1205 | ,simd_arg_converter, &vec_arg | |||
| 1206 | )) { | |||
| 1207 | return NULL((void*)0); | |||
| 1208 | } | |||
| 1209 | npyv_lanetype_s8 *seq_ptr = seq_arg.data.qs8; | |||
| 1210 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1211 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1212 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s816; | |||
| 1213 | if (stride < 0) { | |||
| 1214 | seq_ptr += cur_seq_len -1; | |||
| 1215 | min_seq_len = -min_seq_len; | |||
| 1216 | } | |||
| 1217 | // overflow guard | |||
| 1218 | if (cur_seq_len < min_seq_len) { | |||
| 1219 | PyErr_Format(PyExc_ValueError, | |||
| 1220 | "storen_till_s8(), according to provided stride %d, the" | |||
| 1221 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1222 | stride, min_seq_len, cur_seq_len | |||
| 1223 | ); | |||
| 1224 | goto err; | |||
| 1225 | } | |||
| 1226 | npyv_storen_till_s8( | |||
| 1227 | seq_ptr, stride | |||
| 1228 | #if 1 | |||
| 1229 | ,nlane_arg.data.u32 | |||
| 1230 | #endif | |||
| 1231 | ,vec_arg.data.vs8 | |||
| 1232 | ); | |||
| 1233 | // write-back | |||
| 1234 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs8, simd_data_qs8)) { | |||
| 1235 | goto err; | |||
| 1236 | } | |||
| 1237 | simd_arg_free(&seq_arg); | |||
| 1238 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1239 | err: | |||
| 1240 | simd_arg_free(&seq_arg); | |||
| 1241 | return NULL((void*)0); | |||
| 1242 | } | |||
| 1243 | ||||
| 1244 | #endif // 0 | |||
| 1245 | ||||
| 1246 | /*************************** | |||
| 1247 | * Misc | |||
| 1248 | ***************************/ | |||
| 1249 | SIMD_IMPL_INTRIN_0(zero_s8, vs8)static PyObject *simd__intrin_zero_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_s8") ) return ((void*)0); simd_arg a = { .dtype = simd_data_vs8, .data = {.vs8 = _mm_setzero_si128()}, }; return simd_arg_to_obj(&a); } | |||
| 1250 | SIMD_IMPL_INTRIN_1(setall_s8, vs8, s8)static PyObject *simd__intrin_setall_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s8}; if (!PyArg_ParseTuple( args, "O&:" "setall_s8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs8 = _mm_set1_epi8((char)(arg.data.s8) )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1251 | SIMD_IMPL_INTRIN_3(select_s8, vs8, vb8, vs8, vs8)static PyObject *simd__intrin_select_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_s8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_select_u8 ( arg1.data.vb8, arg2.data.vs8, arg3.data.vs8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1252 | ||||
| 1253 | #line 246 | |||
| 1254 | #if 1 | |||
| 1255 | SIMD_IMPL_INTRIN_1(reinterpret_u8_s8, vu8, vs8)static PyObject *simd__intrin_reinterpret_u8_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "reinterpret_u8_s8", simd_arg_converter, &arg )) return ( (void*)0); simd_data data = {.vu8 = arg.data.vs8}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 1256 | #endif // simd_sup2 | |||
| 1257 | ||||
| 1258 | #line 246 | |||
| 1259 | #if 1 | |||
| 1260 | SIMD_IMPL_INTRIN_1(reinterpret_s8_s8, vs8, vs8)static PyObject *simd__intrin_reinterpret_s8_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "reinterpret_s8_s8", simd_arg_converter, &arg )) return ( (void*)0); simd_data data = {.vs8 = arg.data.vs8}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1261 | #endif // simd_sup2 | |||
| 1262 | ||||
| 1263 | #line 246 | |||
| 1264 | #if 1 | |||
| 1265 | SIMD_IMPL_INTRIN_1(reinterpret_u16_s8, vu16, vs8)static PyObject *simd__intrin_reinterpret_u16_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1266 | #endif // simd_sup2 | |||
| 1267 | ||||
| 1268 | #line 246 | |||
| 1269 | #if 1 | |||
| 1270 | SIMD_IMPL_INTRIN_1(reinterpret_s16_s8, vs16, vs8)static PyObject *simd__intrin_reinterpret_s16_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 1271 | #endif // simd_sup2 | |||
| 1272 | ||||
| 1273 | #line 246 | |||
| 1274 | #if 1 | |||
| 1275 | SIMD_IMPL_INTRIN_1(reinterpret_u32_s8, vu32, vs8)static PyObject *simd__intrin_reinterpret_u32_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 1276 | #endif // simd_sup2 | |||
| 1277 | ||||
| 1278 | #line 246 | |||
| 1279 | #if 1 | |||
| 1280 | SIMD_IMPL_INTRIN_1(reinterpret_s32_s8, vs32, vs8)static PyObject *simd__intrin_reinterpret_s32_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 1281 | #endif // simd_sup2 | |||
| 1282 | ||||
| 1283 | #line 246 | |||
| 1284 | #if 1 | |||
| 1285 | SIMD_IMPL_INTRIN_1(reinterpret_u64_s8, vu64, vs8)static PyObject *simd__intrin_reinterpret_u64_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 1286 | #endif // simd_sup2 | |||
| 1287 | ||||
| 1288 | #line 246 | |||
| 1289 | #if 1 | |||
| 1290 | SIMD_IMPL_INTRIN_1(reinterpret_s64_s8, vs64, vs8)static PyObject *simd__intrin_reinterpret_s64_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vs8}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 1291 | #endif // simd_sup2 | |||
| 1292 | ||||
| 1293 | #line 246 | |||
| 1294 | #if 1 | |||
| 1295 | SIMD_IMPL_INTRIN_1(reinterpret_f32_s8, vf32, vs8)static PyObject *simd__intrin_reinterpret_f32_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vs8 )}; simd_arg_free(&arg); simd_arg ret = { . data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 1296 | #endif // simd_sup2 | |||
| 1297 | ||||
| 1298 | #line 246 | |||
| 1299 | #if NPY_SIMD_F641 | |||
| 1300 | SIMD_IMPL_INTRIN_1(reinterpret_f64_s8, vf64, vs8)static PyObject *simd__intrin_reinterpret_f64_s8 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_s8", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vs8 )}; simd_arg_free(&arg); simd_arg ret = { . data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 1301 | #endif // simd_sup2 | |||
| 1302 | ||||
| 1303 | ||||
| 1304 | /** | |||
| 1305 | * special definition due to the nature of intrinsics | |||
| 1306 | * npyv_setf_s8 and npy_set_s8. | |||
| 1307 | */ | |||
| 1308 | #line 258 | |||
| 1309 | static PyObject * | |||
| 1310 | simd__intrin_setf_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1311 | { | |||
| 1312 | npyv_lanetype_s8 *data = simd_sequence_from_iterable(args, simd_data_qs8, npyv_nlanes_s816); | |||
| 1313 | if (data == NULL((void*)0)) { | |||
| 1314 | return NULL((void*)0); | |||
| 1315 | } | |||
| 1316 | simd_data r = {.vs8 = npyv_setf_s8(npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1317 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1318 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1319 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1320 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1321 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1322 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1323 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1324 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1325 | data[64] // for setfnpyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16])) | |||
| 1326 | )npyv__setr_epi8((char)(data[1]), (char)(data[2]), (char)(data [3]), (char)(data[4]), (char)(data[5]), (char)(data[6]), (char )(data[7]), (char)(data[8]), (char)(data[9]), (char)(data[10] ), (char)(data[11]), (char)(data[12]), (char)(data[13]), (char )(data[14]), (char)(data[15]), (char)(data[16]))}; | |||
| 1327 | simd_sequence_free(data); | |||
| 1328 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs8); | |||
| 1329 | } | |||
| 1330 | ||||
| 1331 | #line 258 | |||
| 1332 | static PyObject * | |||
| 1333 | simd__intrin_set_s8(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1334 | { | |||
| 1335 | npyv_lanetype_s8 *data = simd_sequence_from_iterable(args, simd_data_qs8, npyv_nlanes_s816); | |||
| 1336 | if (data == NULL((void*)0)) { | |||
| 1337 | return NULL((void*)0); | |||
| 1338 | } | |||
| 1339 | simd_data r = {.vs8 = npyv_set_s8(npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1340 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1341 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1342 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1343 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1344 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1345 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1346 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1347 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1348 | data[64] // for setfnpyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15])) | |||
| 1349 | )npyv__setr_epi8((char)(data[0]), (char)(data[1]), (char)(data [2]), (char)(data[3]), (char)(data[4]), (char)(data[5]), (char )(data[6]), (char)(data[7]), (char)(data[8]), (char)(data[9]) , (char)(data[10]), (char)(data[11]), (char)(data[12]), (char )(data[13]), (char)(data[14]), (char)(data[15]))}; | |||
| 1350 | simd_sequence_free(data); | |||
| 1351 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs8); | |||
| 1352 | } | |||
| 1353 | ||||
| 1354 | ||||
| 1355 | /*************************** | |||
| 1356 | * Reorder | |||
| 1357 | ***************************/ | |||
| 1358 | #line 287 | |||
| 1359 | SIMD_IMPL_INTRIN_2(combinel_s8, vs8, vs8, vs8)static PyObject *simd__intrin_combinel_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_s8", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs8 = _mm_unpacklo_epi64 ( arg1.data.vs8, arg2.data.vs8 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1360 | ||||
| 1361 | #line 287 | |||
| 1362 | SIMD_IMPL_INTRIN_2(combineh_s8, vs8, vs8, vs8)static PyObject *simd__intrin_combineh_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_s8", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs8 = _mm_unpackhi_epi64 ( arg1.data.vs8, arg2.data.vs8 )}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1363 | ||||
| 1364 | ||||
| 1365 | #line 293 | |||
| 1366 | SIMD_IMPL_INTRIN_2(combine_s8, vs8x2, vs8, vs8)static PyObject *simd__intrin_combine_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8x2 = npyv__combine( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8x2 }; return simd_arg_to_obj(&ret); } | |||
| 1367 | ||||
| 1368 | #line 293 | |||
| 1369 | SIMD_IMPL_INTRIN_2(zip_s8, vs8x2, vs8, vs8)static PyObject *simd__intrin_zip_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8x2 = npyv_zip_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8x2 }; return simd_arg_to_obj(&ret); } | |||
| 1370 | ||||
| 1371 | ||||
| 1372 | #if 1 | |||
| 1373 | SIMD_IMPL_INTRIN_1(rev64_s8, vs8, vs8)static PyObject *simd__intrin_rev64_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "rev64_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs8 = npyv_rev64_u8( arg.data.vs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1374 | #endif | |||
| 1375 | ||||
| 1376 | /*************************** | |||
| 1377 | * Operators | |||
| 1378 | ***************************/ | |||
| 1379 | #if 0 > 0 | |||
| 1380 | SIMD_IMPL_INTRIN_2(shl_s8, vs8, vs8, u8)static PyObject *simd__intrin_shl_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_shl_s8( arg1.data.vs8, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1381 | SIMD_IMPL_INTRIN_2(shr_s8, vs8, vs8, u8)static PyObject *simd__intrin_shr_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_shr_s8( arg1.data.vs8, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1382 | // immediate constant | |||
| 1383 | SIMD_IMPL_INTRIN_2IMM(shli_s8, vs8, vs8, 0)static PyObject *simd__intrin_shli_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs8 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shli_s8, vs8 ) data.vs8; simd_arg_free(& arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs8 } ; return simd_arg_to_obj(&ret); } | |||
| 1384 | SIMD_IMPL_INTRIN_2IMM(shri_s8, vs8, vs8, 0)static PyObject *simd__intrin_shri_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs8 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shri_s8, vs8 ) data.vs8; simd_arg_free(& arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs8 } ; return simd_arg_to_obj(&ret); } | |||
| 1385 | #endif // shl_imm | |||
| 1386 | ||||
| 1387 | #line 314 | |||
| 1388 | SIMD_IMPL_INTRIN_2(and_s8, vs8, vs8, vs8)static PyObject *simd__intrin_and_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""and_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_and_si128( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1389 | ||||
| 1390 | #line 314 | |||
| 1391 | SIMD_IMPL_INTRIN_2(or_s8, vs8, vs8, vs8)static PyObject *simd__intrin_or_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""or_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_or_si128( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1392 | ||||
| 1393 | #line 314 | |||
| 1394 | SIMD_IMPL_INTRIN_2(xor_s8, vs8, vs8, vs8)static PyObject *simd__intrin_xor_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_xor_si128( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1395 | ||||
| 1396 | ||||
| 1397 | SIMD_IMPL_INTRIN_1(not_s8, vs8, vs8)static PyObject *simd__intrin_not_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "not_s8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs8 = _mm_xor_si128(arg.data.vs8, _mm_set1_epi32(-1 ))}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1398 | ||||
| 1399 | #line 322 | |||
| 1400 | SIMD_IMPL_INTRIN_2(cmpeq_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmpeq_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_cmpeq_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 1401 | ||||
| 1402 | #line 322 | |||
| 1403 | SIMD_IMPL_INTRIN_2(cmpneq_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmpneq_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_xor_si128(_mm_cmpeq_epi8(arg1 .data.vs8, arg2.data.vs8), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(& ret); } | |||
| 1404 | ||||
| 1405 | #line 322 | |||
| 1406 | SIMD_IMPL_INTRIN_2(cmpgt_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmpgt_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_cmpgt_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 1407 | ||||
| 1408 | #line 322 | |||
| 1409 | SIMD_IMPL_INTRIN_2(cmpge_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmpge_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_xor_si128(_mm_cmpgt_epi8(arg2 .data.vs8, arg1.data.vs8), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(& ret); } | |||
| 1410 | ||||
| 1411 | #line 322 | |||
| 1412 | SIMD_IMPL_INTRIN_2(cmplt_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmplt_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_cmpgt_epi8(arg2.data.vs8, arg1 .data.vs8)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 1413 | ||||
| 1414 | #line 322 | |||
| 1415 | SIMD_IMPL_INTRIN_2(cmple_s8, vb8, vs8, vs8)static PyObject *simd__intrin_cmple_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_xor_si128(_mm_cmpgt_epi8(arg1 .data.vs8, arg2.data.vs8), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(& ret); } | |||
| 1416 | ||||
| 1417 | ||||
| 1418 | /*************************** | |||
| 1419 | * Conversion | |||
| 1420 | ***************************/ | |||
| 1421 | SIMD_IMPL_INTRIN_1(cvt_s8_b8, vs8, vb8)static PyObject *simd__intrin_cvt_s8_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb8}; if (!PyArg_ParseTuple( args, "O&:" "cvt_s8_b8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs8 = arg.data.vb8}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 } ; return simd_arg_to_obj(&ret); } | |||
| 1422 | SIMD_IMPL_INTRIN_1(cvt_b8_s8, vb8, vs8)static PyObject *simd__intrin_cvt_b8_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b8_s8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vb8 = arg.data.vs8}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vb8 } ; return simd_arg_to_obj(&ret); } | |||
| 1423 | #if 0 | |||
| 1424 | SIMD_IMPL_INTRIN_1(expand_s8_s8, vs8x2, vs8)static PyObject *simd__intrin_expand_s8_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "expand_s8_s8", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vs8x2 = npyv_expand_s8_s8( arg.data. vs8 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs8x2 }; return simd_arg_to_obj(&ret ); } | |||
| 1425 | #endif // expand_sup | |||
| 1426 | /*************************** | |||
| 1427 | * Arithmetic | |||
| 1428 | ***************************/ | |||
| 1429 | #line 339 | |||
| 1430 | SIMD_IMPL_INTRIN_2(add_s8, vs8, vs8, vs8)static PyObject *simd__intrin_add_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""add_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_add_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1431 | ||||
| 1432 | #line 339 | |||
| 1433 | SIMD_IMPL_INTRIN_2(sub_s8, vs8, vs8, vs8)static PyObject *simd__intrin_sub_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_sub_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1434 | ||||
| 1435 | ||||
| 1436 | #if 1 | |||
| 1437 | #line 346 | |||
| 1438 | SIMD_IMPL_INTRIN_2(adds_s8, vs8, vs8, vs8)static PyObject *simd__intrin_adds_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_adds_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1439 | ||||
| 1440 | #line 346 | |||
| 1441 | SIMD_IMPL_INTRIN_2(subs_s8, vs8, vs8, vs8)static PyObject *simd__intrin_subs_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = _mm_subs_epi8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1442 | ||||
| 1443 | #endif // sat_sup | |||
| 1444 | ||||
| 1445 | #if 1 | |||
| 1446 | SIMD_IMPL_INTRIN_2(mul_s8, vs8, vs8, vs8)static PyObject *simd__intrin_mul_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_mul_u8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1447 | #endif // mul_sup | |||
| 1448 | ||||
| 1449 | #if 0 | |||
| 1450 | SIMD_IMPL_INTRIN_2(div_s8, vs8, vs8, vs8)static PyObject *simd__intrin_div_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""div_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_div_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1451 | #endif // div_sup | |||
| 1452 | ||||
| 1453 | #if 1 | |||
| 1454 | SIMD_IMPL_INTRIN_1(divisor_s8, vs8x3, s8)static PyObject *simd__intrin_divisor_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s8}; if (!PyArg_ParseTuple( args, "O&:" "divisor_s8", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs8x3 = npyv_divisor_s8( arg.data.s8 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8x3 }; return simd_arg_to_obj(&ret); } | |||
| 1455 | SIMD_IMPL_INTRIN_2(divc_s8, vs8, vs8, vs8x3)static PyObject *simd__intrin_divc_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_divc_s8( arg1.data.vs8, arg2 .data.vs8x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs8 } ; return simd_arg_to_obj(&ret); } | |||
| 1456 | #endif // intdiv_sup | |||
| 1457 | ||||
| 1458 | #if 0 | |||
| 1459 | #line 367 | |||
| 1460 | SIMD_IMPL_INTRIN_3(muladd_s8, vs8, vs8, vs8, vs8)static PyObject *simd__intrin_muladd_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_s8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_muladd_s8 ( arg1.data.vs8, arg2.data.vs8, arg3.data.vs8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1461 | ||||
| 1462 | #line 367 | |||
| 1463 | SIMD_IMPL_INTRIN_3(mulsub_s8, vs8, vs8, vs8, vs8)static PyObject *simd__intrin_mulsub_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_s8", simd_arg_converter, &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_mulsub_s8 ( arg1.data.vs8, arg2.data.vs8, arg3.data.vs8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1464 | ||||
| 1465 | #line 367 | |||
| 1466 | SIMD_IMPL_INTRIN_3(nmuladd_s8, vs8, vs8, vs8, vs8)static PyObject *simd__intrin_nmuladd_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_nmuladd_s8 ( arg1.data.vs8, arg2.data.vs8, arg3.data.vs8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1467 | ||||
| 1468 | #line 367 | |||
| 1469 | SIMD_IMPL_INTRIN_3(nmulsub_s8, vs8, vs8, vs8, vs8)static PyObject *simd__intrin_nmulsub_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs8 = npyv_nmulsub_s8 ( arg1.data.vs8, arg2.data.vs8, arg3.data.vs8 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1470 | ||||
| 1471 | #endif // fused_sup | |||
| 1472 | ||||
| 1473 | #if 0 | |||
| 1474 | SIMD_IMPL_INTRIN_1(sum_s8, s8, vs8)static PyObject *simd__intrin_sum_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "sum_s8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.s8 = npyv_sum_s8( arg.data.vs8 )}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_s8 }; return simd_arg_to_obj(&ret); } | |||
| 1475 | #endif // sum_sup | |||
| 1476 | ||||
| 1477 | #if 0 | |||
| 1478 | SIMD_IMPL_INTRIN_1(sumup_s8, s8, vs8)static PyObject *simd__intrin_sumup_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "sumup_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.s8 = npyv_sumup_s8( arg.data.vs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_s8 }; return simd_arg_to_obj(&ret); } | |||
| 1479 | #endif // sumup_sup | |||
| 1480 | ||||
| 1481 | /*************************** | |||
| 1482 | * Math | |||
| 1483 | ***************************/ | |||
| 1484 | #if 0 | |||
| 1485 | #line 386 | |||
| 1486 | SIMD_IMPL_INTRIN_1(sqrt_s8, vs8, vs8)static PyObject *simd__intrin_sqrt_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_s8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs8 = npyv_sqrt_s8( arg.data.vs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1487 | ||||
| 1488 | #line 386 | |||
| 1489 | SIMD_IMPL_INTRIN_1(recip_s8, vs8, vs8)static PyObject *simd__intrin_recip_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "recip_s8", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs8 = npyv_recip_s8( arg.data.vs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1490 | ||||
| 1491 | #line 386 | |||
| 1492 | SIMD_IMPL_INTRIN_1(abs_s8, vs8, vs8)static PyObject *simd__intrin_abs_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "abs_s8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs8 = npyv_abs_s8( arg.data.vs8 )}; simd_arg_free(& arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 } ; return simd_arg_to_obj(&ret); } | |||
| 1493 | ||||
| 1494 | #line 386 | |||
| 1495 | SIMD_IMPL_INTRIN_1(square_s8, vs8, vs8)static PyObject *simd__intrin_square_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&:" "square_s8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs8 = npyv_square_s8( arg.data.vs8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1496 | ||||
| 1497 | #endif | |||
| 1498 | ||||
| 1499 | #line 393 | |||
| 1500 | SIMD_IMPL_INTRIN_2(max_s8, vs8, vs8, vs8)static PyObject *simd__intrin_max_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""max_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_max_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1501 | ||||
| 1502 | #line 393 | |||
| 1503 | SIMD_IMPL_INTRIN_2(min_s8, vs8, vs8, vs8)static PyObject *simd__intrin_min_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""min_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_min_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1504 | ||||
| 1505 | ||||
| 1506 | #if 0 | |||
| 1507 | #line 400 | |||
| 1508 | SIMD_IMPL_INTRIN_2(maxp_s8, vs8, vs8, vs8)static PyObject *simd__intrin_maxp_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_maxp_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1509 | ||||
| 1510 | #line 400 | |||
| 1511 | SIMD_IMPL_INTRIN_2(minp_s8, vs8, vs8, vs8)static PyObject *simd__intrin_minp_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_s8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs8 = npyv_minp_s8( arg1.data.vs8, arg2 .data.vs8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 1512 | ||||
| 1513 | #endif | |||
| 1514 | ||||
| 1515 | /*************************** | |||
| 1516 | * Mask operations | |||
| 1517 | ***************************/ | |||
| 1518 | #line 410 | |||
| 1519 | SIMD_IMPL_INTRIN_4(ifadd_s8, vs8, vb8, vs8, vs8, vs8)static PyObject *simd__intrin_ifadd_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; simd_arg arg4 = { .dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_s8", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs8 = npyv_ifadd_s8 ( arg1.data.vb8, arg2.data.vs8, arg3.data.vs8, arg4.data.vs8 ) }; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(& ret); } | |||
| 1520 | ||||
| 1521 | #line 410 | |||
| 1522 | SIMD_IMPL_INTRIN_4(ifsub_s8, vs8, vb8, vs8, vs8, vs8)static PyObject *simd__intrin_ifsub_s8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vs8 }; simd_arg arg3 = {.dtype = simd_data_vs8}; simd_arg arg4 = { .dtype = simd_data_vs8}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_s8", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs8 = npyv_ifsub_s8 ( arg1.data.vb8, arg2.data.vs8, arg3.data.vs8, arg4.data.vs8 ) }; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(& ret); } | |||
| 1523 | ||||
| 1524 | ||||
| 1525 | #endif // simd_sup | |||
| 1526 | ||||
| 1527 | #line 34 | |||
| 1528 | #if 1 | |||
| 1529 | /*************************** | |||
| 1530 | * Memory | |||
| 1531 | ***************************/ | |||
| 1532 | #line 41 | |||
| 1533 | SIMD_IMPL_INTRIN_1(load_u16, vu16, qu16)static PyObject *simd__intrin_load_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu16}; if (!PyArg_ParseTuple( args, "O&:" "load_u16", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu16 = npyv_load_u16( arg.data.qu16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1534 | ||||
| 1535 | #line 41 | |||
| 1536 | SIMD_IMPL_INTRIN_1(loada_u16, vu16, qu16)static PyObject *simd__intrin_loada_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu16}; if (!PyArg_ParseTuple( args, "O&:" "loada_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu16 = npyv_loada_u16( arg.data.qu16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1537 | ||||
| 1538 | #line 41 | |||
| 1539 | SIMD_IMPL_INTRIN_1(loads_u16, vu16, qu16)static PyObject *simd__intrin_loads_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu16}; if (!PyArg_ParseTuple( args, "O&:" "loads_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu16 = npyv_loads_u16( arg.data.qu16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1540 | ||||
| 1541 | #line 41 | |||
| 1542 | SIMD_IMPL_INTRIN_1(loadl_u16, vu16, qu16)static PyObject *simd__intrin_loadl_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu16}; if (!PyArg_ParseTuple( args, "O&:" "loadl_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu16 = npyv_loadl_u16( arg.data.qu16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1543 | ||||
| 1544 | #line 46 | |||
| 1545 | // special definition due to the nature of store | |||
| 1546 | static PyObject * | |||
| 1547 | simd__intrin_store_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1548 | { | |||
| 1549 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1550 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1551 | if (!PyArg_ParseTuple( | |||
| 1552 | args, "O&O&:store_u16", | |||
| 1553 | simd_arg_converter, &seq_arg, | |||
| 1554 | simd_arg_converter, &vec_arg | |||
| 1555 | )) { | |||
| 1556 | return NULL((void*)0); | |||
| 1557 | } | |||
| 1558 | npyv_store_u16(seq_arg.data.qu16, vec_arg.data.vu16); | |||
| 1559 | // write-back | |||
| 1560 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1561 | simd_arg_free(&seq_arg); | |||
| 1562 | return NULL((void*)0); | |||
| 1563 | } | |||
| 1564 | simd_arg_free(&seq_arg); | |||
| 1565 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1566 | } | |||
| 1567 | ||||
| 1568 | #line 46 | |||
| 1569 | // special definition due to the nature of storea | |||
| 1570 | static PyObject * | |||
| 1571 | simd__intrin_storea_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1572 | { | |||
| 1573 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1574 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1575 | if (!PyArg_ParseTuple( | |||
| 1576 | args, "O&O&:storea_u16", | |||
| 1577 | simd_arg_converter, &seq_arg, | |||
| 1578 | simd_arg_converter, &vec_arg | |||
| 1579 | )) { | |||
| 1580 | return NULL((void*)0); | |||
| 1581 | } | |||
| 1582 | npyv_storea_u16(seq_arg.data.qu16, vec_arg.data.vu16); | |||
| 1583 | // write-back | |||
| 1584 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1585 | simd_arg_free(&seq_arg); | |||
| 1586 | return NULL((void*)0); | |||
| 1587 | } | |||
| 1588 | simd_arg_free(&seq_arg); | |||
| 1589 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1590 | } | |||
| 1591 | ||||
| 1592 | #line 46 | |||
| 1593 | // special definition due to the nature of stores | |||
| 1594 | static PyObject * | |||
| 1595 | simd__intrin_stores_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1596 | { | |||
| 1597 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1598 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1599 | if (!PyArg_ParseTuple( | |||
| 1600 | args, "O&O&:stores_u16", | |||
| 1601 | simd_arg_converter, &seq_arg, | |||
| 1602 | simd_arg_converter, &vec_arg | |||
| 1603 | )) { | |||
| 1604 | return NULL((void*)0); | |||
| 1605 | } | |||
| 1606 | npyv_stores_u16(seq_arg.data.qu16, vec_arg.data.vu16); | |||
| 1607 | // write-back | |||
| 1608 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1609 | simd_arg_free(&seq_arg); | |||
| 1610 | return NULL((void*)0); | |||
| 1611 | } | |||
| 1612 | simd_arg_free(&seq_arg); | |||
| 1613 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1614 | } | |||
| 1615 | ||||
| 1616 | #line 46 | |||
| 1617 | // special definition due to the nature of storel | |||
| 1618 | static PyObject * | |||
| 1619 | simd__intrin_storel_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1620 | { | |||
| 1621 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1622 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1623 | if (!PyArg_ParseTuple( | |||
| 1624 | args, "O&O&:storel_u16", | |||
| 1625 | simd_arg_converter, &seq_arg, | |||
| 1626 | simd_arg_converter, &vec_arg | |||
| 1627 | )) { | |||
| 1628 | return NULL((void*)0); | |||
| 1629 | } | |||
| 1630 | npyv_storel_u16(seq_arg.data.qu16, vec_arg.data.vu16); | |||
| 1631 | // write-back | |||
| 1632 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1633 | simd_arg_free(&seq_arg); | |||
| 1634 | return NULL((void*)0); | |||
| 1635 | } | |||
| 1636 | simd_arg_free(&seq_arg); | |||
| 1637 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1638 | } | |||
| 1639 | ||||
| 1640 | #line 46 | |||
| 1641 | // special definition due to the nature of storeh | |||
| 1642 | static PyObject * | |||
| 1643 | simd__intrin_storeh_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1644 | { | |||
| 1645 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1646 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1647 | if (!PyArg_ParseTuple( | |||
| 1648 | args, "O&O&:storeh_u16", | |||
| 1649 | simd_arg_converter, &seq_arg, | |||
| 1650 | simd_arg_converter, &vec_arg | |||
| 1651 | )) { | |||
| 1652 | return NULL((void*)0); | |||
| 1653 | } | |||
| 1654 | npyv_storeh_u16(seq_arg.data.qu16, vec_arg.data.vu16); | |||
| 1655 | // write-back | |||
| 1656 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1657 | simd_arg_free(&seq_arg); | |||
| 1658 | return NULL((void*)0); | |||
| 1659 | } | |||
| 1660 | simd_arg_free(&seq_arg); | |||
| 1661 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1662 | } | |||
| 1663 | ||||
| 1664 | ||||
| 1665 | /**************************************** | |||
| 1666 | * Non-contiguous/Partial Memory access | |||
| 1667 | ****************************************/ | |||
| 1668 | #if 0 | |||
| 1669 | // Partial Load | |||
| 1670 | SIMD_IMPL_INTRIN_3(load_till_u16, vu16, qu16, u32, u16)static PyObject *simd__intrin_load_till_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu16}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_u16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_load_till_u16 ( arg1.data.qu16, arg2.data.u32, arg3.data.u16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1671 | SIMD_IMPL_INTRIN_2(load_tillz_u16, vu16, qu16, u32)static PyObject *simd__intrin_load_tillz_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu16}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_u16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu16 = npyv_load_tillz_u16 ( arg1.data.qu16, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 1672 | ||||
| 1673 | // Partial Store | |||
| 1674 | static PyObject * | |||
| 1675 | simd__intrin_store_till_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1676 | { | |||
| 1677 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1678 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1679 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1680 | if (!PyArg_ParseTuple( | |||
| 1681 | args, "O&O&O&:store_till_u16", | |||
| 1682 | simd_arg_converter, &seq_arg, | |||
| 1683 | simd_arg_converter, &nlane_arg, | |||
| 1684 | simd_arg_converter, &vec_arg | |||
| 1685 | )) { | |||
| 1686 | return NULL((void*)0); | |||
| 1687 | } | |||
| 1688 | npyv_store_till_u16( | |||
| 1689 | seq_arg.data.qu16, nlane_arg.data.u32, vec_arg.data.vu16 | |||
| 1690 | ); | |||
| 1691 | // write-back | |||
| 1692 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1693 | simd_arg_free(&seq_arg); | |||
| 1694 | return NULL((void*)0); | |||
| 1695 | } | |||
| 1696 | simd_arg_free(&seq_arg); | |||
| 1697 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1698 | } | |||
| 1699 | ||||
| 1700 | // Non-contiguous Load | |||
| 1701 | #line 112 | |||
| 1702 | static PyObject * | |||
| 1703 | simd__intrin_loadn_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1704 | { | |||
| 1705 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1706 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1707 | #if 0 | |||
| 1708 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1709 | #endif // till | |||
| 1710 | #if 0 | |||
| 1711 | simd_arg fill_arg = {.dtype = simd_data_u16}; | |||
| 1712 | #endif | |||
| 1713 | if (!PyArg_ParseTuple( | |||
| 1714 | args, "O&O&:loadn_u16", | |||
| 1715 | simd_arg_converter, &seq_arg, | |||
| 1716 | simd_arg_converter, &stride_arg | |||
| 1717 | #if 0 | |||
| 1718 | ,simd_arg_converter, &nlane_arg | |||
| 1719 | #endif | |||
| 1720 | #if 0 | |||
| 1721 | ,simd_arg_converter, &fill_arg | |||
| 1722 | #endif | |||
| 1723 | )) { | |||
| 1724 | return NULL((void*)0); | |||
| 1725 | } | |||
| 1726 | npyv_lanetype_u16 *seq_ptr = seq_arg.data.qu16; | |||
| 1727 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1728 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1729 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u168; | |||
| 1730 | if (stride < 0) { | |||
| 1731 | seq_ptr += cur_seq_len -1; | |||
| 1732 | min_seq_len = -min_seq_len; | |||
| 1733 | } | |||
| 1734 | if (cur_seq_len < min_seq_len) { | |||
| 1735 | PyErr_Format(PyExc_ValueError, | |||
| 1736 | "loadn_u16(), according to provided stride %d, the " | |||
| 1737 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1738 | stride, min_seq_len, cur_seq_len | |||
| 1739 | ); | |||
| 1740 | goto err; | |||
| 1741 | } | |||
| 1742 | npyv_u16 rvec = npyv_loadn_u16( | |||
| 1743 | seq_ptr, stride | |||
| 1744 | #if 0 | |||
| 1745 | , nlane_arg.data.u32 | |||
| 1746 | #endif | |||
| 1747 | #if 0 | |||
| 1748 | , fill_arg.data.u16 | |||
| 1749 | #endif | |||
| 1750 | ); | |||
| 1751 | simd_arg ret = { | |||
| 1752 | .dtype = simd_data_vu16, .data = {.vu16=rvec} | |||
| 1753 | }; | |||
| 1754 | simd_arg_free(&seq_arg); | |||
| 1755 | return simd_arg_to_obj(&ret); | |||
| 1756 | err: | |||
| 1757 | simd_arg_free(&seq_arg); | |||
| 1758 | return NULL((void*)0); | |||
| 1759 | } | |||
| 1760 | ||||
| 1761 | #line 112 | |||
| 1762 | static PyObject * | |||
| 1763 | simd__intrin_loadn_till_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1764 | { | |||
| 1765 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1766 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1767 | #if 1 | |||
| 1768 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1769 | #endif // till | |||
| 1770 | #if 1 | |||
| 1771 | simd_arg fill_arg = {.dtype = simd_data_u16}; | |||
| 1772 | #endif | |||
| 1773 | if (!PyArg_ParseTuple( | |||
| 1774 | args, "O&O&O&O&:loadn_till_u16", | |||
| 1775 | simd_arg_converter, &seq_arg, | |||
| 1776 | simd_arg_converter, &stride_arg | |||
| 1777 | #if 1 | |||
| 1778 | ,simd_arg_converter, &nlane_arg | |||
| 1779 | #endif | |||
| 1780 | #if 1 | |||
| 1781 | ,simd_arg_converter, &fill_arg | |||
| 1782 | #endif | |||
| 1783 | )) { | |||
| 1784 | return NULL((void*)0); | |||
| 1785 | } | |||
| 1786 | npyv_lanetype_u16 *seq_ptr = seq_arg.data.qu16; | |||
| 1787 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1788 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1789 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u168; | |||
| 1790 | if (stride < 0) { | |||
| 1791 | seq_ptr += cur_seq_len -1; | |||
| 1792 | min_seq_len = -min_seq_len; | |||
| 1793 | } | |||
| 1794 | if (cur_seq_len < min_seq_len) { | |||
| 1795 | PyErr_Format(PyExc_ValueError, | |||
| 1796 | "loadn_till_u16(), according to provided stride %d, the " | |||
| 1797 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1798 | stride, min_seq_len, cur_seq_len | |||
| 1799 | ); | |||
| 1800 | goto err; | |||
| 1801 | } | |||
| 1802 | npyv_u16 rvec = npyv_loadn_till_u16( | |||
| 1803 | seq_ptr, stride | |||
| 1804 | #if 1 | |||
| 1805 | , nlane_arg.data.u32 | |||
| 1806 | #endif | |||
| 1807 | #if 1 | |||
| 1808 | , fill_arg.data.u16 | |||
| 1809 | #endif | |||
| 1810 | ); | |||
| 1811 | simd_arg ret = { | |||
| 1812 | .dtype = simd_data_vu16, .data = {.vu16=rvec} | |||
| 1813 | }; | |||
| 1814 | simd_arg_free(&seq_arg); | |||
| 1815 | return simd_arg_to_obj(&ret); | |||
| 1816 | err: | |||
| 1817 | simd_arg_free(&seq_arg); | |||
| 1818 | return NULL((void*)0); | |||
| 1819 | } | |||
| 1820 | ||||
| 1821 | #line 112 | |||
| 1822 | static PyObject * | |||
| 1823 | simd__intrin_loadn_tillz_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1824 | { | |||
| 1825 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1826 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1827 | #if 1 | |||
| 1828 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1829 | #endif // till | |||
| 1830 | #if 0 | |||
| 1831 | simd_arg fill_arg = {.dtype = simd_data_u16}; | |||
| 1832 | #endif | |||
| 1833 | if (!PyArg_ParseTuple( | |||
| 1834 | args, "O&O&O&:loadn_tillz_u16", | |||
| 1835 | simd_arg_converter, &seq_arg, | |||
| 1836 | simd_arg_converter, &stride_arg | |||
| 1837 | #if 1 | |||
| 1838 | ,simd_arg_converter, &nlane_arg | |||
| 1839 | #endif | |||
| 1840 | #if 0 | |||
| 1841 | ,simd_arg_converter, &fill_arg | |||
| 1842 | #endif | |||
| 1843 | )) { | |||
| 1844 | return NULL((void*)0); | |||
| 1845 | } | |||
| 1846 | npyv_lanetype_u16 *seq_ptr = seq_arg.data.qu16; | |||
| 1847 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1848 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1849 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u168; | |||
| 1850 | if (stride < 0) { | |||
| 1851 | seq_ptr += cur_seq_len -1; | |||
| 1852 | min_seq_len = -min_seq_len; | |||
| 1853 | } | |||
| 1854 | if (cur_seq_len < min_seq_len) { | |||
| 1855 | PyErr_Format(PyExc_ValueError, | |||
| 1856 | "loadn_tillz_u16(), according to provided stride %d, the " | |||
| 1857 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1858 | stride, min_seq_len, cur_seq_len | |||
| 1859 | ); | |||
| 1860 | goto err; | |||
| 1861 | } | |||
| 1862 | npyv_u16 rvec = npyv_loadn_tillz_u16( | |||
| 1863 | seq_ptr, stride | |||
| 1864 | #if 1 | |||
| 1865 | , nlane_arg.data.u32 | |||
| 1866 | #endif | |||
| 1867 | #if 0 | |||
| 1868 | , fill_arg.data.u16 | |||
| 1869 | #endif | |||
| 1870 | ); | |||
| 1871 | simd_arg ret = { | |||
| 1872 | .dtype = simd_data_vu16, .data = {.vu16=rvec} | |||
| 1873 | }; | |||
| 1874 | simd_arg_free(&seq_arg); | |||
| 1875 | return simd_arg_to_obj(&ret); | |||
| 1876 | err: | |||
| 1877 | simd_arg_free(&seq_arg); | |||
| 1878 | return NULL((void*)0); | |||
| 1879 | } | |||
| 1880 | ||||
| 1881 | ||||
| 1882 | // Non-contiguous Store | |||
| 1883 | #line 178 | |||
| 1884 | static PyObject * | |||
| 1885 | simd__intrin_storen_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1886 | { | |||
| 1887 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1888 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1889 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1890 | #if 0 | |||
| 1891 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1892 | #endif | |||
| 1893 | if (!PyArg_ParseTuple( | |||
| 1894 | args, "O&O&O&:storen_u16", | |||
| 1895 | simd_arg_converter, &seq_arg, | |||
| 1896 | simd_arg_converter, &stride_arg | |||
| 1897 | #if 0 | |||
| 1898 | ,simd_arg_converter, &nlane_arg | |||
| 1899 | #endif | |||
| 1900 | ,simd_arg_converter, &vec_arg | |||
| 1901 | )) { | |||
| 1902 | return NULL((void*)0); | |||
| 1903 | } | |||
| 1904 | npyv_lanetype_u16 *seq_ptr = seq_arg.data.qu16; | |||
| 1905 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1906 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1907 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u168; | |||
| 1908 | if (stride < 0) { | |||
| 1909 | seq_ptr += cur_seq_len -1; | |||
| 1910 | min_seq_len = -min_seq_len; | |||
| 1911 | } | |||
| 1912 | // overflow guard | |||
| 1913 | if (cur_seq_len < min_seq_len) { | |||
| 1914 | PyErr_Format(PyExc_ValueError, | |||
| 1915 | "storen_u16(), according to provided stride %d, the" | |||
| 1916 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1917 | stride, min_seq_len, cur_seq_len | |||
| 1918 | ); | |||
| 1919 | goto err; | |||
| 1920 | } | |||
| 1921 | npyv_storen_u16( | |||
| 1922 | seq_ptr, stride | |||
| 1923 | #if 0 | |||
| 1924 | ,nlane_arg.data.u32 | |||
| 1925 | #endif | |||
| 1926 | ,vec_arg.data.vu16 | |||
| 1927 | ); | |||
| 1928 | // write-back | |||
| 1929 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1930 | goto err; | |||
| 1931 | } | |||
| 1932 | simd_arg_free(&seq_arg); | |||
| 1933 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1934 | err: | |||
| 1935 | simd_arg_free(&seq_arg); | |||
| 1936 | return NULL((void*)0); | |||
| 1937 | } | |||
| 1938 | ||||
| 1939 | #line 178 | |||
| 1940 | static PyObject * | |||
| 1941 | simd__intrin_storen_till_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 1942 | { | |||
| 1943 | simd_arg seq_arg = {.dtype = simd_data_qu16}; | |||
| 1944 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 1945 | simd_arg vec_arg = {.dtype = simd_data_vu16}; | |||
| 1946 | #if 1 | |||
| 1947 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 1948 | #endif | |||
| 1949 | if (!PyArg_ParseTuple( | |||
| 1950 | args, "O&O&O&O&:storen_u16", | |||
| 1951 | simd_arg_converter, &seq_arg, | |||
| 1952 | simd_arg_converter, &stride_arg | |||
| 1953 | #if 1 | |||
| 1954 | ,simd_arg_converter, &nlane_arg | |||
| 1955 | #endif | |||
| 1956 | ,simd_arg_converter, &vec_arg | |||
| 1957 | )) { | |||
| 1958 | return NULL((void*)0); | |||
| 1959 | } | |||
| 1960 | npyv_lanetype_u16 *seq_ptr = seq_arg.data.qu16; | |||
| 1961 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 1962 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 1963 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u168; | |||
| 1964 | if (stride < 0) { | |||
| 1965 | seq_ptr += cur_seq_len -1; | |||
| 1966 | min_seq_len = -min_seq_len; | |||
| 1967 | } | |||
| 1968 | // overflow guard | |||
| 1969 | if (cur_seq_len < min_seq_len) { | |||
| 1970 | PyErr_Format(PyExc_ValueError, | |||
| 1971 | "storen_till_u16(), according to provided stride %d, the" | |||
| 1972 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 1973 | stride, min_seq_len, cur_seq_len | |||
| 1974 | ); | |||
| 1975 | goto err; | |||
| 1976 | } | |||
| 1977 | npyv_storen_till_u16( | |||
| 1978 | seq_ptr, stride | |||
| 1979 | #if 1 | |||
| 1980 | ,nlane_arg.data.u32 | |||
| 1981 | #endif | |||
| 1982 | ,vec_arg.data.vu16 | |||
| 1983 | ); | |||
| 1984 | // write-back | |||
| 1985 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu16, simd_data_qu16)) { | |||
| 1986 | goto err; | |||
| 1987 | } | |||
| 1988 | simd_arg_free(&seq_arg); | |||
| 1989 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 1990 | err: | |||
| 1991 | simd_arg_free(&seq_arg); | |||
| 1992 | return NULL((void*)0); | |||
| 1993 | } | |||
| 1994 | ||||
| 1995 | #endif // 0 | |||
| 1996 | ||||
| 1997 | /*************************** | |||
| 1998 | * Misc | |||
| 1999 | ***************************/ | |||
| 2000 | SIMD_IMPL_INTRIN_0(zero_u16, vu16)static PyObject *simd__intrin_zero_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_u16") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vu16, .data = {.vu16 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 2001 | SIMD_IMPL_INTRIN_1(setall_u16, vu16, u16)static PyObject *simd__intrin_setall_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u16}; if (!PyArg_ParseTuple( args, "O&:" "setall_u16", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu16 = _mm_set1_epi16((short)(arg.data .u16))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2002 | SIMD_IMPL_INTRIN_3(select_u16, vu16, vb16, vu16, vu16)static PyObject *simd__intrin_select_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_select_u8 ( arg1.data.vb16, arg2.data.vu16, arg3.data.vu16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2003 | ||||
| 2004 | #line 246 | |||
| 2005 | #if 1 | |||
| 2006 | SIMD_IMPL_INTRIN_1(reinterpret_u8_u16, vu8, vu16)static PyObject *simd__intrin_reinterpret_u8_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vu16 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 2007 | #endif // simd_sup2 | |||
| 2008 | ||||
| 2009 | #line 246 | |||
| 2010 | #if 1 | |||
| 2011 | SIMD_IMPL_INTRIN_1(reinterpret_s8_u16, vs8, vu16)static PyObject *simd__intrin_reinterpret_s8_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vu16 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 2012 | #endif // simd_sup2 | |||
| 2013 | ||||
| 2014 | #line 246 | |||
| 2015 | #if 1 | |||
| 2016 | SIMD_IMPL_INTRIN_1(reinterpret_u16_u16, vu16, vu16)static PyObject *simd__intrin_reinterpret_u16_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2017 | #endif // simd_sup2 | |||
| 2018 | ||||
| 2019 | #line 246 | |||
| 2020 | #if 1 | |||
| 2021 | SIMD_IMPL_INTRIN_1(reinterpret_s16_u16, vs16, vu16)static PyObject *simd__intrin_reinterpret_s16_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2022 | #endif // simd_sup2 | |||
| 2023 | ||||
| 2024 | #line 246 | |||
| 2025 | #if 1 | |||
| 2026 | SIMD_IMPL_INTRIN_1(reinterpret_u32_u16, vu32, vu16)static PyObject *simd__intrin_reinterpret_u32_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 2027 | #endif // simd_sup2 | |||
| 2028 | ||||
| 2029 | #line 246 | |||
| 2030 | #if 1 | |||
| 2031 | SIMD_IMPL_INTRIN_1(reinterpret_s32_u16, vs32, vu16)static PyObject *simd__intrin_reinterpret_s32_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 2032 | #endif // simd_sup2 | |||
| 2033 | ||||
| 2034 | #line 246 | |||
| 2035 | #if 1 | |||
| 2036 | SIMD_IMPL_INTRIN_1(reinterpret_u64_u16, vu64, vu16)static PyObject *simd__intrin_reinterpret_u64_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 2037 | #endif // simd_sup2 | |||
| 2038 | ||||
| 2039 | #line 246 | |||
| 2040 | #if 1 | |||
| 2041 | SIMD_IMPL_INTRIN_1(reinterpret_s64_u16, vs64, vu16)static PyObject *simd__intrin_reinterpret_s64_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vu16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 2042 | #endif // simd_sup2 | |||
| 2043 | ||||
| 2044 | #line 246 | |||
| 2045 | #if 1 | |||
| 2046 | SIMD_IMPL_INTRIN_1(reinterpret_f32_u16, vf32, vu16)static PyObject *simd__intrin_reinterpret_f32_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vu16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 2047 | #endif // simd_sup2 | |||
| 2048 | ||||
| 2049 | #line 246 | |||
| 2050 | #if NPY_SIMD_F641 | |||
| 2051 | SIMD_IMPL_INTRIN_1(reinterpret_f64_u16, vf64, vu16)static PyObject *simd__intrin_reinterpret_f64_u16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_u16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vu16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 2052 | #endif // simd_sup2 | |||
| 2053 | ||||
| 2054 | ||||
| 2055 | /** | |||
| 2056 | * special definition due to the nature of intrinsics | |||
| 2057 | * npyv_setf_u16 and npy_set_u16. | |||
| 2058 | */ | |||
| 2059 | #line 258 | |||
| 2060 | static PyObject * | |||
| 2061 | simd__intrin_setf_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2062 | { | |||
| 2063 | npyv_lanetype_u16 *data = simd_sequence_from_iterable(args, simd_data_qu16, npyv_nlanes_u168); | |||
| 2064 | if (data == NULL((void*)0)) { | |||
| 2065 | return NULL((void*)0); | |||
| 2066 | } | |||
| 2067 | simd_data r = {.vu16 = npyv_setf_u16(npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2068 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2069 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2070 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2071 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2072 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2073 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2074 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2075 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2076 | data[64] // for setfnpyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2077 | )npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8]))}; | |||
| 2078 | simd_sequence_free(data); | |||
| 2079 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu16); | |||
| 2080 | } | |||
| 2081 | ||||
| 2082 | #line 258 | |||
| 2083 | static PyObject * | |||
| 2084 | simd__intrin_set_u16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2085 | { | |||
| 2086 | npyv_lanetype_u16 *data = simd_sequence_from_iterable(args, simd_data_qu16, npyv_nlanes_u168); | |||
| 2087 | if (data == NULL((void*)0)) { | |||
| 2088 | return NULL((void*)0); | |||
| 2089 | } | |||
| 2090 | simd_data r = {.vu16 = npyv_set_u16(npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2091 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2092 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2093 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2094 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2095 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2096 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2097 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2098 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2099 | data[64] // for setfnpyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2100 | )npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7]))}; | |||
| 2101 | simd_sequence_free(data); | |||
| 2102 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu16); | |||
| 2103 | } | |||
| 2104 | ||||
| 2105 | ||||
| 2106 | /*************************** | |||
| 2107 | * Reorder | |||
| 2108 | ***************************/ | |||
| 2109 | #line 287 | |||
| 2110 | SIMD_IMPL_INTRIN_2(combinel_u16, vu16, vu16, vu16)static PyObject *simd__intrin_combinel_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_u16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu16 = _mm_unpacklo_epi64 ( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2111 | ||||
| 2112 | #line 287 | |||
| 2113 | SIMD_IMPL_INTRIN_2(combineh_u16, vu16, vu16, vu16)static PyObject *simd__intrin_combineh_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_u16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu16 = _mm_unpackhi_epi64 ( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2114 | ||||
| 2115 | ||||
| 2116 | #line 293 | |||
| 2117 | SIMD_IMPL_INTRIN_2(combine_u16, vu16x2, vu16, vu16)static PyObject *simd__intrin_combine_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_u16", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu16x2 = npyv__combine ( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16x2 }; return simd_arg_to_obj(&ret); } | |||
| 2118 | ||||
| 2119 | #line 293 | |||
| 2120 | SIMD_IMPL_INTRIN_2(zip_u16, vu16x2, vu16, vu16)static PyObject *simd__intrin_zip_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16x2 = npyv_zip_u16( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16x2 }; return simd_arg_to_obj(&ret); } | |||
| 2121 | ||||
| 2122 | ||||
| 2123 | #if 1 | |||
| 2124 | SIMD_IMPL_INTRIN_1(rev64_u16, vu16, vu16)static PyObject *simd__intrin_rev64_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "rev64_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu16 = npyv_rev64_u16( arg.data.vu16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2125 | #endif | |||
| 2126 | ||||
| 2127 | /*************************** | |||
| 2128 | * Operators | |||
| 2129 | ***************************/ | |||
| 2130 | #if 15 > 0 | |||
| 2131 | SIMD_IMPL_INTRIN_2(shl_u16, vu16, vu16, u8)static PyObject *simd__intrin_shl_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_sll_epi16(arg1.data.vu16, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2132 | SIMD_IMPL_INTRIN_2(shr_u16, vu16, vu16, u8)static PyObject *simd__intrin_shr_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_srl_epi16(arg1.data.vu16, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2133 | // immediate constant | |||
| 2134 | SIMD_IMPL_INTRIN_2IMM(shli_u16, vu16, vu16, 15)static PyObject *simd__intrin_shli_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu16 = 0 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 0) : 1 == arg2.data.u8 ? _mm_slli_epi16 (arg1.data.vu16, 1) : 2 == arg2.data.u8 ? _mm_slli_epi16(arg1 .data.vu16, 2) : 3 == arg2.data.u8 ? _mm_slli_epi16(arg1.data .vu16, 3) : 4 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 5) : 6 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 6) : 7 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi16(arg1.data.vu16, 8) : 9 == arg2.data .u8 ? _mm_slli_epi16(arg1.data.vu16, 9) : 10 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 10) : 11 == arg2.data.u8 ? _mm_slli_epi16 (arg1.data.vu16, 11) : 12 == arg2.data.u8 ? _mm_slli_epi16(arg1 .data.vu16, 12) : 13 == arg2.data.u8 ? _mm_slli_epi16(arg1.data .vu16, 13) : 14 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vu16, 15 ) : data.vu16; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(& ret); } | |||
| 2135 | SIMD_IMPL_INTRIN_2IMM(shri_u16, vu16, vu16, 16)static PyObject *simd__intrin_shri_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu16 = 1 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 1) : 2 == arg2.data.u8 ? _mm_srli_epi16 (arg1.data.vu16, 2) : 3 == arg2.data.u8 ? _mm_srli_epi16(arg1 .data.vu16, 3) : 4 == arg2.data.u8 ? _mm_srli_epi16(arg1.data .vu16, 4) : 5 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16 , 5) : 6 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 6) : 7 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 7) : 8 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 8) : 9 == arg2 .data.u8 ? _mm_srli_epi16(arg1.data.vu16, 9) : 10 == arg2.data .u8 ? _mm_srli_epi16(arg1.data.vu16, 10) : 11 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 11) : 12 == arg2.data.u8 ? _mm_srli_epi16 (arg1.data.vu16, 12) : 13 == arg2.data.u8 ? _mm_srli_epi16(arg1 .data.vu16, 13) : 14 == arg2.data.u8 ? _mm_srli_epi16(arg1.data .vu16, 14) : 15 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16 , 15) : 16 == arg2.data.u8 ? _mm_srli_epi16(arg1.data.vu16, 16 ) : data.vu16; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(& ret); } | |||
| 2136 | #endif // shl_imm | |||
| 2137 | ||||
| 2138 | #line 314 | |||
| 2139 | SIMD_IMPL_INTRIN_2(and_u16, vu16, vu16, vu16)static PyObject *simd__intrin_and_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""and_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_and_si128( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2140 | ||||
| 2141 | #line 314 | |||
| 2142 | SIMD_IMPL_INTRIN_2(or_u16, vu16, vu16, vu16)static PyObject *simd__intrin_or_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""or_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_or_si128( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2143 | ||||
| 2144 | #line 314 | |||
| 2145 | SIMD_IMPL_INTRIN_2(xor_u16, vu16, vu16, vu16)static PyObject *simd__intrin_xor_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_xor_si128( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2146 | ||||
| 2147 | ||||
| 2148 | SIMD_IMPL_INTRIN_1(not_u16, vu16, vu16)static PyObject *simd__intrin_not_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "not_u16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu16 = _mm_xor_si128(arg.data.vu16, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2149 | ||||
| 2150 | #line 322 | |||
| 2151 | SIMD_IMPL_INTRIN_2(cmpeq_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmpeq_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpeq_epi16( arg1.data.vu16 , arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2152 | ||||
| 2153 | #line 322 | |||
| 2154 | SIMD_IMPL_INTRIN_2(cmpneq_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmpneq_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_xor_si128(_mm_cmpeq_epi16(arg1 .data.vu16, arg2.data.vu16), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2155 | ||||
| 2156 | #line 322 | |||
| 2157 | SIMD_IMPL_INTRIN_2(cmpgt_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmpgt_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = npyv_cmpgt_u16( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2158 | ||||
| 2159 | #line 322 | |||
| 2160 | SIMD_IMPL_INTRIN_2(cmpge_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmpge_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpeq_epi16(_mm_subs_epu16( arg2.data.vu16, arg1.data.vu16), _mm_setzero_si128())}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2161 | ||||
| 2162 | #line 322 | |||
| 2163 | SIMD_IMPL_INTRIN_2(cmplt_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmplt_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = npyv_cmpgt_u16(arg2.data.vu16, arg1 .data.vu16)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2164 | ||||
| 2165 | #line 322 | |||
| 2166 | SIMD_IMPL_INTRIN_2(cmple_u16, vb16, vu16, vu16)static PyObject *simd__intrin_cmple_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpeq_epi16(_mm_subs_epu16( arg1.data.vu16, arg2.data.vu16), _mm_setzero_si128())}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2167 | ||||
| 2168 | ||||
| 2169 | /*************************** | |||
| 2170 | * Conversion | |||
| 2171 | ***************************/ | |||
| 2172 | SIMD_IMPL_INTRIN_1(cvt_u16_b16, vu16, vb16)static PyObject *simd__intrin_cvt_u16_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb16}; if (!PyArg_ParseTuple( args, "O&:" "cvt_u16_b16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu16 = arg.data.vb16}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2173 | SIMD_IMPL_INTRIN_1(cvt_b16_u16, vb16, vu16)static PyObject *simd__intrin_cvt_b16_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b16_u16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb16 = arg.data.vu16}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2174 | #if 1 | |||
| 2175 | SIMD_IMPL_INTRIN_1(expand_u32_u16, vu32x2, vu16)static PyObject *simd__intrin_expand_u32_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "expand_u32_u16", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vu32x2 = npyv_expand_u32_u16( arg.data .vu16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32x2 }; return simd_arg_to_obj(&ret ); } | |||
| 2176 | #endif // expand_sup | |||
| 2177 | /*************************** | |||
| 2178 | * Arithmetic | |||
| 2179 | ***************************/ | |||
| 2180 | #line 339 | |||
| 2181 | SIMD_IMPL_INTRIN_2(add_u16, vu16, vu16, vu16)static PyObject *simd__intrin_add_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""add_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_add_epi16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2182 | ||||
| 2183 | #line 339 | |||
| 2184 | SIMD_IMPL_INTRIN_2(sub_u16, vu16, vu16, vu16)static PyObject *simd__intrin_sub_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_sub_epi16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2185 | ||||
| 2186 | ||||
| 2187 | #if 1 | |||
| 2188 | #line 346 | |||
| 2189 | SIMD_IMPL_INTRIN_2(adds_u16, vu16, vu16, vu16)static PyObject *simd__intrin_adds_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_adds_epu16( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2190 | ||||
| 2191 | #line 346 | |||
| 2192 | SIMD_IMPL_INTRIN_2(subs_u16, vu16, vu16, vu16)static PyObject *simd__intrin_subs_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_subs_epu16( arg1.data.vu16, arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2193 | ||||
| 2194 | #endif // sat_sup | |||
| 2195 | ||||
| 2196 | #if 1 | |||
| 2197 | SIMD_IMPL_INTRIN_2(mul_u16, vu16, vu16, vu16)static PyObject *simd__intrin_mul_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = _mm_mullo_epi16( arg1.data.vu16 , arg2.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2198 | #endif // mul_sup | |||
| 2199 | ||||
| 2200 | #if 0 | |||
| 2201 | SIMD_IMPL_INTRIN_2(div_u16, vu16, vu16, vu16)static PyObject *simd__intrin_div_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""div_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_div_u16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2202 | #endif // div_sup | |||
| 2203 | ||||
| 2204 | #if 1 | |||
| 2205 | SIMD_IMPL_INTRIN_1(divisor_u16, vu16x3, u16)static PyObject *simd__intrin_divisor_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u16}; if (!PyArg_ParseTuple( args, "O&:" "divisor_u16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu16x3 = npyv_divisor_u16( arg.data.u16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu16x3 }; return simd_arg_to_obj(&ret); } | |||
| 2206 | SIMD_IMPL_INTRIN_2(divc_u16, vu16, vu16, vu16x3)static PyObject *simd__intrin_divc_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_divc_u16( arg1.data.vu16, arg2 .data.vu16x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2207 | #endif // intdiv_sup | |||
| 2208 | ||||
| 2209 | #if 0 | |||
| 2210 | #line 367 | |||
| 2211 | SIMD_IMPL_INTRIN_3(muladd_u16, vu16, vu16, vu16, vu16)static PyObject *simd__intrin_muladd_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_muladd_u16 ( arg1.data.vu16, arg2.data.vu16, arg3.data.vu16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2212 | ||||
| 2213 | #line 367 | |||
| 2214 | SIMD_IMPL_INTRIN_3(mulsub_u16, vu16, vu16, vu16, vu16)static PyObject *simd__intrin_mulsub_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_mulsub_u16 ( arg1.data.vu16, arg2.data.vu16, arg3.data.vu16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2215 | ||||
| 2216 | #line 367 | |||
| 2217 | SIMD_IMPL_INTRIN_3(nmuladd_u16, vu16, vu16, vu16, vu16)static PyObject *simd__intrin_nmuladd_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_nmuladd_u16 ( arg1.data.vu16, arg2.data.vu16, arg3.data.vu16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2218 | ||||
| 2219 | #line 367 | |||
| 2220 | SIMD_IMPL_INTRIN_3(nmulsub_u16, vu16, vu16, vu16, vu16)static PyObject *simd__intrin_nmulsub_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu16 = npyv_nmulsub_u16 ( arg1.data.vu16, arg2.data.vu16, arg3.data.vu16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2221 | ||||
| 2222 | #endif // fused_sup | |||
| 2223 | ||||
| 2224 | #if 0 | |||
| 2225 | SIMD_IMPL_INTRIN_1(sum_u16, u16, vu16)static PyObject *simd__intrin_sum_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "sum_u16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.u16 = npyv_sum_u16( arg.data.vu16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_u16 }; return simd_arg_to_obj(&ret); } | |||
| 2226 | #endif // sum_sup | |||
| 2227 | ||||
| 2228 | #if 1 | |||
| 2229 | SIMD_IMPL_INTRIN_1(sumup_u16, u32, vu16)static PyObject *simd__intrin_sumup_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "sumup_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.u32 = npyv_sumup_u16( arg.data.vu16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u32 }; return simd_arg_to_obj(&ret); } | |||
| 2230 | #endif // sumup_sup | |||
| 2231 | ||||
| 2232 | /*************************** | |||
| 2233 | * Math | |||
| 2234 | ***************************/ | |||
| 2235 | #if 0 | |||
| 2236 | #line 386 | |||
| 2237 | SIMD_IMPL_INTRIN_1(sqrt_u16, vu16, vu16)static PyObject *simd__intrin_sqrt_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_u16", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu16 = npyv_sqrt_u16( arg.data.vu16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2238 | ||||
| 2239 | #line 386 | |||
| 2240 | SIMD_IMPL_INTRIN_1(recip_u16, vu16, vu16)static PyObject *simd__intrin_recip_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "recip_u16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu16 = npyv_recip_u16( arg.data.vu16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2241 | ||||
| 2242 | #line 386 | |||
| 2243 | SIMD_IMPL_INTRIN_1(abs_u16, vu16, vu16)static PyObject *simd__intrin_abs_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "abs_u16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu16 = npyv_abs_u16( arg.data.vu16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2244 | ||||
| 2245 | #line 386 | |||
| 2246 | SIMD_IMPL_INTRIN_1(square_u16, vu16, vu16)static PyObject *simd__intrin_square_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&:" "square_u16", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu16 = npyv_square_u16( arg.data.vu16 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2247 | ||||
| 2248 | #endif | |||
| 2249 | ||||
| 2250 | #line 393 | |||
| 2251 | SIMD_IMPL_INTRIN_2(max_u16, vu16, vu16, vu16)static PyObject *simd__intrin_max_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""max_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_max_u16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2252 | ||||
| 2253 | #line 393 | |||
| 2254 | SIMD_IMPL_INTRIN_2(min_u16, vu16, vu16, vu16)static PyObject *simd__intrin_min_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""min_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_min_u16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2255 | ||||
| 2256 | ||||
| 2257 | #if 0 | |||
| 2258 | #line 400 | |||
| 2259 | SIMD_IMPL_INTRIN_2(maxp_u16, vu16, vu16, vu16)static PyObject *simd__intrin_maxp_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_maxp_u16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2260 | ||||
| 2261 | #line 400 | |||
| 2262 | SIMD_IMPL_INTRIN_2(minp_u16, vu16, vu16, vu16)static PyObject *simd__intrin_minp_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_u16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu16 = npyv_minp_u16( arg1.data.vu16, arg2 .data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret); } | |||
| 2263 | ||||
| 2264 | #endif | |||
| 2265 | ||||
| 2266 | /*************************** | |||
| 2267 | * Mask operations | |||
| 2268 | ***************************/ | |||
| 2269 | #line 410 | |||
| 2270 | SIMD_IMPL_INTRIN_4(ifadd_u16, vu16, vb16, vu16, vu16, vu16)static PyObject *simd__intrin_ifadd_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; simd_arg arg4 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_u16", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu16 = npyv_ifadd_u16 ( arg1.data.vb16, arg2.data.vu16, arg3.data.vu16, arg4.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(& ret); } | |||
| 2271 | ||||
| 2272 | #line 410 | |||
| 2273 | SIMD_IMPL_INTRIN_4(ifsub_u16, vu16, vb16, vu16, vu16, vu16)static PyObject *simd__intrin_ifsub_u16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vu16 }; simd_arg arg3 = {.dtype = simd_data_vu16}; simd_arg arg4 = {.dtype = simd_data_vu16}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_u16", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu16 = npyv_ifsub_u16 ( arg1.data.vb16, arg2.data.vu16, arg3.data.vu16, arg4.data.vu16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj(& ret); } | |||
| 2274 | ||||
| 2275 | ||||
| 2276 | #endif // simd_sup | |||
| 2277 | ||||
| 2278 | #line 34 | |||
| 2279 | #if 1 | |||
| 2280 | /*************************** | |||
| 2281 | * Memory | |||
| 2282 | ***************************/ | |||
| 2283 | #line 41 | |||
| 2284 | SIMD_IMPL_INTRIN_1(load_s16, vs16, qs16)static PyObject *simd__intrin_load_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs16}; if (!PyArg_ParseTuple( args, "O&:" "load_s16", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs16 = npyv_load_s16( arg.data.qs16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2285 | ||||
| 2286 | #line 41 | |||
| 2287 | SIMD_IMPL_INTRIN_1(loada_s16, vs16, qs16)static PyObject *simd__intrin_loada_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs16}; if (!PyArg_ParseTuple( args, "O&:" "loada_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs16 = npyv_loada_s16( arg.data.qs16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2288 | ||||
| 2289 | #line 41 | |||
| 2290 | SIMD_IMPL_INTRIN_1(loads_s16, vs16, qs16)static PyObject *simd__intrin_loads_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs16}; if (!PyArg_ParseTuple( args, "O&:" "loads_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs16 = npyv_loads_s16( arg.data.qs16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2291 | ||||
| 2292 | #line 41 | |||
| 2293 | SIMD_IMPL_INTRIN_1(loadl_s16, vs16, qs16)static PyObject *simd__intrin_loadl_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs16}; if (!PyArg_ParseTuple( args, "O&:" "loadl_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs16 = npyv_loadl_s16( arg.data.qs16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2294 | ||||
| 2295 | #line 46 | |||
| 2296 | // special definition due to the nature of store | |||
| 2297 | static PyObject * | |||
| 2298 | simd__intrin_store_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2299 | { | |||
| 2300 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2301 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2302 | if (!PyArg_ParseTuple( | |||
| 2303 | args, "O&O&:store_s16", | |||
| 2304 | simd_arg_converter, &seq_arg, | |||
| 2305 | simd_arg_converter, &vec_arg | |||
| 2306 | )) { | |||
| 2307 | return NULL((void*)0); | |||
| 2308 | } | |||
| 2309 | npyv_store_s16(seq_arg.data.qs16, vec_arg.data.vs16); | |||
| 2310 | // write-back | |||
| 2311 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2312 | simd_arg_free(&seq_arg); | |||
| 2313 | return NULL((void*)0); | |||
| 2314 | } | |||
| 2315 | simd_arg_free(&seq_arg); | |||
| 2316 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2317 | } | |||
| 2318 | ||||
| 2319 | #line 46 | |||
| 2320 | // special definition due to the nature of storea | |||
| 2321 | static PyObject * | |||
| 2322 | simd__intrin_storea_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2323 | { | |||
| 2324 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2325 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2326 | if (!PyArg_ParseTuple( | |||
| 2327 | args, "O&O&:storea_s16", | |||
| 2328 | simd_arg_converter, &seq_arg, | |||
| 2329 | simd_arg_converter, &vec_arg | |||
| 2330 | )) { | |||
| 2331 | return NULL((void*)0); | |||
| 2332 | } | |||
| 2333 | npyv_storea_s16(seq_arg.data.qs16, vec_arg.data.vs16); | |||
| 2334 | // write-back | |||
| 2335 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2336 | simd_arg_free(&seq_arg); | |||
| 2337 | return NULL((void*)0); | |||
| 2338 | } | |||
| 2339 | simd_arg_free(&seq_arg); | |||
| 2340 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2341 | } | |||
| 2342 | ||||
| 2343 | #line 46 | |||
| 2344 | // special definition due to the nature of stores | |||
| 2345 | static PyObject * | |||
| 2346 | simd__intrin_stores_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2347 | { | |||
| 2348 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2349 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2350 | if (!PyArg_ParseTuple( | |||
| 2351 | args, "O&O&:stores_s16", | |||
| 2352 | simd_arg_converter, &seq_arg, | |||
| 2353 | simd_arg_converter, &vec_arg | |||
| 2354 | )) { | |||
| 2355 | return NULL((void*)0); | |||
| 2356 | } | |||
| 2357 | npyv_stores_s16(seq_arg.data.qs16, vec_arg.data.vs16); | |||
| 2358 | // write-back | |||
| 2359 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2360 | simd_arg_free(&seq_arg); | |||
| 2361 | return NULL((void*)0); | |||
| 2362 | } | |||
| 2363 | simd_arg_free(&seq_arg); | |||
| 2364 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2365 | } | |||
| 2366 | ||||
| 2367 | #line 46 | |||
| 2368 | // special definition due to the nature of storel | |||
| 2369 | static PyObject * | |||
| 2370 | simd__intrin_storel_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2371 | { | |||
| 2372 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2373 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2374 | if (!PyArg_ParseTuple( | |||
| 2375 | args, "O&O&:storel_s16", | |||
| 2376 | simd_arg_converter, &seq_arg, | |||
| 2377 | simd_arg_converter, &vec_arg | |||
| 2378 | )) { | |||
| 2379 | return NULL((void*)0); | |||
| 2380 | } | |||
| 2381 | npyv_storel_s16(seq_arg.data.qs16, vec_arg.data.vs16); | |||
| 2382 | // write-back | |||
| 2383 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2384 | simd_arg_free(&seq_arg); | |||
| 2385 | return NULL((void*)0); | |||
| 2386 | } | |||
| 2387 | simd_arg_free(&seq_arg); | |||
| 2388 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2389 | } | |||
| 2390 | ||||
| 2391 | #line 46 | |||
| 2392 | // special definition due to the nature of storeh | |||
| 2393 | static PyObject * | |||
| 2394 | simd__intrin_storeh_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2395 | { | |||
| 2396 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2397 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2398 | if (!PyArg_ParseTuple( | |||
| 2399 | args, "O&O&:storeh_s16", | |||
| 2400 | simd_arg_converter, &seq_arg, | |||
| 2401 | simd_arg_converter, &vec_arg | |||
| 2402 | )) { | |||
| 2403 | return NULL((void*)0); | |||
| 2404 | } | |||
| 2405 | npyv_storeh_s16(seq_arg.data.qs16, vec_arg.data.vs16); | |||
| 2406 | // write-back | |||
| 2407 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2408 | simd_arg_free(&seq_arg); | |||
| 2409 | return NULL((void*)0); | |||
| 2410 | } | |||
| 2411 | simd_arg_free(&seq_arg); | |||
| 2412 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2413 | } | |||
| 2414 | ||||
| 2415 | ||||
| 2416 | /**************************************** | |||
| 2417 | * Non-contiguous/Partial Memory access | |||
| 2418 | ****************************************/ | |||
| 2419 | #if 0 | |||
| 2420 | // Partial Load | |||
| 2421 | SIMD_IMPL_INTRIN_3(load_till_s16, vs16, qs16, u32, s16)static PyObject *simd__intrin_load_till_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs16}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_s16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_load_till_s16 ( arg1.data.qs16, arg2.data.u32, arg3.data.s16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2422 | SIMD_IMPL_INTRIN_2(load_tillz_s16, vs16, qs16, u32)static PyObject *simd__intrin_load_tillz_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs16}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_s16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs16 = npyv_load_tillz_s16 ( arg1.data.qs16, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2423 | ||||
| 2424 | // Partial Store | |||
| 2425 | static PyObject * | |||
| 2426 | simd__intrin_store_till_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2427 | { | |||
| 2428 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2429 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2430 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2431 | if (!PyArg_ParseTuple( | |||
| 2432 | args, "O&O&O&:store_till_s16", | |||
| 2433 | simd_arg_converter, &seq_arg, | |||
| 2434 | simd_arg_converter, &nlane_arg, | |||
| 2435 | simd_arg_converter, &vec_arg | |||
| 2436 | )) { | |||
| 2437 | return NULL((void*)0); | |||
| 2438 | } | |||
| 2439 | npyv_store_till_s16( | |||
| 2440 | seq_arg.data.qs16, nlane_arg.data.u32, vec_arg.data.vs16 | |||
| 2441 | ); | |||
| 2442 | // write-back | |||
| 2443 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2444 | simd_arg_free(&seq_arg); | |||
| 2445 | return NULL((void*)0); | |||
| 2446 | } | |||
| 2447 | simd_arg_free(&seq_arg); | |||
| 2448 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2449 | } | |||
| 2450 | ||||
| 2451 | // Non-contiguous Load | |||
| 2452 | #line 112 | |||
| 2453 | static PyObject * | |||
| 2454 | simd__intrin_loadn_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2455 | { | |||
| 2456 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2457 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 2458 | #if 0 | |||
| 2459 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2460 | #endif // till | |||
| 2461 | #if 0 | |||
| 2462 | simd_arg fill_arg = {.dtype = simd_data_s16}; | |||
| 2463 | #endif | |||
| 2464 | if (!PyArg_ParseTuple( | |||
| 2465 | args, "O&O&:loadn_s16", | |||
| 2466 | simd_arg_converter, &seq_arg, | |||
| 2467 | simd_arg_converter, &stride_arg | |||
| 2468 | #if 0 | |||
| 2469 | ,simd_arg_converter, &nlane_arg | |||
| 2470 | #endif | |||
| 2471 | #if 0 | |||
| 2472 | ,simd_arg_converter, &fill_arg | |||
| 2473 | #endif | |||
| 2474 | )) { | |||
| 2475 | return NULL((void*)0); | |||
| 2476 | } | |||
| 2477 | npyv_lanetype_s16 *seq_ptr = seq_arg.data.qs16; | |||
| 2478 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 2479 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 2480 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s168; | |||
| 2481 | if (stride < 0) { | |||
| 2482 | seq_ptr += cur_seq_len -1; | |||
| 2483 | min_seq_len = -min_seq_len; | |||
| 2484 | } | |||
| 2485 | if (cur_seq_len < min_seq_len) { | |||
| 2486 | PyErr_Format(PyExc_ValueError, | |||
| 2487 | "loadn_s16(), according to provided stride %d, the " | |||
| 2488 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 2489 | stride, min_seq_len, cur_seq_len | |||
| 2490 | ); | |||
| 2491 | goto err; | |||
| 2492 | } | |||
| 2493 | npyv_s16 rvec = npyv_loadn_s16( | |||
| 2494 | seq_ptr, stride | |||
| 2495 | #if 0 | |||
| 2496 | , nlane_arg.data.u32 | |||
| 2497 | #endif | |||
| 2498 | #if 0 | |||
| 2499 | , fill_arg.data.s16 | |||
| 2500 | #endif | |||
| 2501 | ); | |||
| 2502 | simd_arg ret = { | |||
| 2503 | .dtype = simd_data_vs16, .data = {.vs16=rvec} | |||
| 2504 | }; | |||
| 2505 | simd_arg_free(&seq_arg); | |||
| 2506 | return simd_arg_to_obj(&ret); | |||
| 2507 | err: | |||
| 2508 | simd_arg_free(&seq_arg); | |||
| 2509 | return NULL((void*)0); | |||
| 2510 | } | |||
| 2511 | ||||
| 2512 | #line 112 | |||
| 2513 | static PyObject * | |||
| 2514 | simd__intrin_loadn_till_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2515 | { | |||
| 2516 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2517 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 2518 | #if 1 | |||
| 2519 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2520 | #endif // till | |||
| 2521 | #if 1 | |||
| 2522 | simd_arg fill_arg = {.dtype = simd_data_s16}; | |||
| 2523 | #endif | |||
| 2524 | if (!PyArg_ParseTuple( | |||
| 2525 | args, "O&O&O&O&:loadn_till_s16", | |||
| 2526 | simd_arg_converter, &seq_arg, | |||
| 2527 | simd_arg_converter, &stride_arg | |||
| 2528 | #if 1 | |||
| 2529 | ,simd_arg_converter, &nlane_arg | |||
| 2530 | #endif | |||
| 2531 | #if 1 | |||
| 2532 | ,simd_arg_converter, &fill_arg | |||
| 2533 | #endif | |||
| 2534 | )) { | |||
| 2535 | return NULL((void*)0); | |||
| 2536 | } | |||
| 2537 | npyv_lanetype_s16 *seq_ptr = seq_arg.data.qs16; | |||
| 2538 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 2539 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 2540 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s168; | |||
| 2541 | if (stride < 0) { | |||
| 2542 | seq_ptr += cur_seq_len -1; | |||
| 2543 | min_seq_len = -min_seq_len; | |||
| 2544 | } | |||
| 2545 | if (cur_seq_len < min_seq_len) { | |||
| 2546 | PyErr_Format(PyExc_ValueError, | |||
| 2547 | "loadn_till_s16(), according to provided stride %d, the " | |||
| 2548 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 2549 | stride, min_seq_len, cur_seq_len | |||
| 2550 | ); | |||
| 2551 | goto err; | |||
| 2552 | } | |||
| 2553 | npyv_s16 rvec = npyv_loadn_till_s16( | |||
| 2554 | seq_ptr, stride | |||
| 2555 | #if 1 | |||
| 2556 | , nlane_arg.data.u32 | |||
| 2557 | #endif | |||
| 2558 | #if 1 | |||
| 2559 | , fill_arg.data.s16 | |||
| 2560 | #endif | |||
| 2561 | ); | |||
| 2562 | simd_arg ret = { | |||
| 2563 | .dtype = simd_data_vs16, .data = {.vs16=rvec} | |||
| 2564 | }; | |||
| 2565 | simd_arg_free(&seq_arg); | |||
| 2566 | return simd_arg_to_obj(&ret); | |||
| 2567 | err: | |||
| 2568 | simd_arg_free(&seq_arg); | |||
| 2569 | return NULL((void*)0); | |||
| 2570 | } | |||
| 2571 | ||||
| 2572 | #line 112 | |||
| 2573 | static PyObject * | |||
| 2574 | simd__intrin_loadn_tillz_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2575 | { | |||
| 2576 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2577 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 2578 | #if 1 | |||
| 2579 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2580 | #endif // till | |||
| 2581 | #if 0 | |||
| 2582 | simd_arg fill_arg = {.dtype = simd_data_s16}; | |||
| 2583 | #endif | |||
| 2584 | if (!PyArg_ParseTuple( | |||
| 2585 | args, "O&O&O&:loadn_tillz_s16", | |||
| 2586 | simd_arg_converter, &seq_arg, | |||
| 2587 | simd_arg_converter, &stride_arg | |||
| 2588 | #if 1 | |||
| 2589 | ,simd_arg_converter, &nlane_arg | |||
| 2590 | #endif | |||
| 2591 | #if 0 | |||
| 2592 | ,simd_arg_converter, &fill_arg | |||
| 2593 | #endif | |||
| 2594 | )) { | |||
| 2595 | return NULL((void*)0); | |||
| 2596 | } | |||
| 2597 | npyv_lanetype_s16 *seq_ptr = seq_arg.data.qs16; | |||
| 2598 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 2599 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 2600 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s168; | |||
| 2601 | if (stride < 0) { | |||
| 2602 | seq_ptr += cur_seq_len -1; | |||
| 2603 | min_seq_len = -min_seq_len; | |||
| 2604 | } | |||
| 2605 | if (cur_seq_len < min_seq_len) { | |||
| 2606 | PyErr_Format(PyExc_ValueError, | |||
| 2607 | "loadn_tillz_s16(), according to provided stride %d, the " | |||
| 2608 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 2609 | stride, min_seq_len, cur_seq_len | |||
| 2610 | ); | |||
| 2611 | goto err; | |||
| 2612 | } | |||
| 2613 | npyv_s16 rvec = npyv_loadn_tillz_s16( | |||
| 2614 | seq_ptr, stride | |||
| 2615 | #if 1 | |||
| 2616 | , nlane_arg.data.u32 | |||
| 2617 | #endif | |||
| 2618 | #if 0 | |||
| 2619 | , fill_arg.data.s16 | |||
| 2620 | #endif | |||
| 2621 | ); | |||
| 2622 | simd_arg ret = { | |||
| 2623 | .dtype = simd_data_vs16, .data = {.vs16=rvec} | |||
| 2624 | }; | |||
| 2625 | simd_arg_free(&seq_arg); | |||
| 2626 | return simd_arg_to_obj(&ret); | |||
| 2627 | err: | |||
| 2628 | simd_arg_free(&seq_arg); | |||
| 2629 | return NULL((void*)0); | |||
| 2630 | } | |||
| 2631 | ||||
| 2632 | ||||
| 2633 | // Non-contiguous Store | |||
| 2634 | #line 178 | |||
| 2635 | static PyObject * | |||
| 2636 | simd__intrin_storen_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2637 | { | |||
| 2638 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2639 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 2640 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2641 | #if 0 | |||
| 2642 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2643 | #endif | |||
| 2644 | if (!PyArg_ParseTuple( | |||
| 2645 | args, "O&O&O&:storen_s16", | |||
| 2646 | simd_arg_converter, &seq_arg, | |||
| 2647 | simd_arg_converter, &stride_arg | |||
| 2648 | #if 0 | |||
| 2649 | ,simd_arg_converter, &nlane_arg | |||
| 2650 | #endif | |||
| 2651 | ,simd_arg_converter, &vec_arg | |||
| 2652 | )) { | |||
| 2653 | return NULL((void*)0); | |||
| 2654 | } | |||
| 2655 | npyv_lanetype_s16 *seq_ptr = seq_arg.data.qs16; | |||
| 2656 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 2657 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 2658 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s168; | |||
| 2659 | if (stride < 0) { | |||
| 2660 | seq_ptr += cur_seq_len -1; | |||
| 2661 | min_seq_len = -min_seq_len; | |||
| 2662 | } | |||
| 2663 | // overflow guard | |||
| 2664 | if (cur_seq_len < min_seq_len) { | |||
| 2665 | PyErr_Format(PyExc_ValueError, | |||
| 2666 | "storen_s16(), according to provided stride %d, the" | |||
| 2667 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 2668 | stride, min_seq_len, cur_seq_len | |||
| 2669 | ); | |||
| 2670 | goto err; | |||
| 2671 | } | |||
| 2672 | npyv_storen_s16( | |||
| 2673 | seq_ptr, stride | |||
| 2674 | #if 0 | |||
| 2675 | ,nlane_arg.data.u32 | |||
| 2676 | #endif | |||
| 2677 | ,vec_arg.data.vs16 | |||
| 2678 | ); | |||
| 2679 | // write-back | |||
| 2680 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2681 | goto err; | |||
| 2682 | } | |||
| 2683 | simd_arg_free(&seq_arg); | |||
| 2684 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2685 | err: | |||
| 2686 | simd_arg_free(&seq_arg); | |||
| 2687 | return NULL((void*)0); | |||
| 2688 | } | |||
| 2689 | ||||
| 2690 | #line 178 | |||
| 2691 | static PyObject * | |||
| 2692 | simd__intrin_storen_till_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2693 | { | |||
| 2694 | simd_arg seq_arg = {.dtype = simd_data_qs16}; | |||
| 2695 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 2696 | simd_arg vec_arg = {.dtype = simd_data_vs16}; | |||
| 2697 | #if 1 | |||
| 2698 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 2699 | #endif | |||
| 2700 | if (!PyArg_ParseTuple( | |||
| 2701 | args, "O&O&O&O&:storen_s16", | |||
| 2702 | simd_arg_converter, &seq_arg, | |||
| 2703 | simd_arg_converter, &stride_arg | |||
| 2704 | #if 1 | |||
| 2705 | ,simd_arg_converter, &nlane_arg | |||
| 2706 | #endif | |||
| 2707 | ,simd_arg_converter, &vec_arg | |||
| 2708 | )) { | |||
| 2709 | return NULL((void*)0); | |||
| 2710 | } | |||
| 2711 | npyv_lanetype_s16 *seq_ptr = seq_arg.data.qs16; | |||
| 2712 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 2713 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 2714 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s168; | |||
| 2715 | if (stride < 0) { | |||
| 2716 | seq_ptr += cur_seq_len -1; | |||
| 2717 | min_seq_len = -min_seq_len; | |||
| 2718 | } | |||
| 2719 | // overflow guard | |||
| 2720 | if (cur_seq_len < min_seq_len) { | |||
| 2721 | PyErr_Format(PyExc_ValueError, | |||
| 2722 | "storen_till_s16(), according to provided stride %d, the" | |||
| 2723 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 2724 | stride, min_seq_len, cur_seq_len | |||
| 2725 | ); | |||
| 2726 | goto err; | |||
| 2727 | } | |||
| 2728 | npyv_storen_till_s16( | |||
| 2729 | seq_ptr, stride | |||
| 2730 | #if 1 | |||
| 2731 | ,nlane_arg.data.u32 | |||
| 2732 | #endif | |||
| 2733 | ,vec_arg.data.vs16 | |||
| 2734 | ); | |||
| 2735 | // write-back | |||
| 2736 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs16, simd_data_qs16)) { | |||
| 2737 | goto err; | |||
| 2738 | } | |||
| 2739 | simd_arg_free(&seq_arg); | |||
| 2740 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 2741 | err: | |||
| 2742 | simd_arg_free(&seq_arg); | |||
| 2743 | return NULL((void*)0); | |||
| 2744 | } | |||
| 2745 | ||||
| 2746 | #endif // 0 | |||
| 2747 | ||||
| 2748 | /*************************** | |||
| 2749 | * Misc | |||
| 2750 | ***************************/ | |||
| 2751 | SIMD_IMPL_INTRIN_0(zero_s16, vs16)static PyObject *simd__intrin_zero_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_s16") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vs16, .data = {.vs16 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 2752 | SIMD_IMPL_INTRIN_1(setall_s16, vs16, s16)static PyObject *simd__intrin_setall_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s16}; if (!PyArg_ParseTuple( args, "O&:" "setall_s16", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs16 = _mm_set1_epi16((short)(arg.data .s16))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2753 | SIMD_IMPL_INTRIN_3(select_s16, vs16, vb16, vs16, vs16)static PyObject *simd__intrin_select_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_select_u8 ( arg1.data.vb16, arg2.data.vs16, arg3.data.vs16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2754 | ||||
| 2755 | #line 246 | |||
| 2756 | #if 1 | |||
| 2757 | SIMD_IMPL_INTRIN_1(reinterpret_u8_s16, vu8, vs16)static PyObject *simd__intrin_reinterpret_u8_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vs16 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 2758 | #endif // simd_sup2 | |||
| 2759 | ||||
| 2760 | #line 246 | |||
| 2761 | #if 1 | |||
| 2762 | SIMD_IMPL_INTRIN_1(reinterpret_s8_s16, vs8, vs16)static PyObject *simd__intrin_reinterpret_s8_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vs16 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 2763 | #endif // simd_sup2 | |||
| 2764 | ||||
| 2765 | #line 246 | |||
| 2766 | #if 1 | |||
| 2767 | SIMD_IMPL_INTRIN_1(reinterpret_u16_s16, vu16, vs16)static PyObject *simd__intrin_reinterpret_u16_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2768 | #endif // simd_sup2 | |||
| 2769 | ||||
| 2770 | #line 246 | |||
| 2771 | #if 1 | |||
| 2772 | SIMD_IMPL_INTRIN_1(reinterpret_s16_s16, vs16, vs16)static PyObject *simd__intrin_reinterpret_s16_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2773 | #endif // simd_sup2 | |||
| 2774 | ||||
| 2775 | #line 246 | |||
| 2776 | #if 1 | |||
| 2777 | SIMD_IMPL_INTRIN_1(reinterpret_u32_s16, vu32, vs16)static PyObject *simd__intrin_reinterpret_u32_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 2778 | #endif // simd_sup2 | |||
| 2779 | ||||
| 2780 | #line 246 | |||
| 2781 | #if 1 | |||
| 2782 | SIMD_IMPL_INTRIN_1(reinterpret_s32_s16, vs32, vs16)static PyObject *simd__intrin_reinterpret_s32_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 2783 | #endif // simd_sup2 | |||
| 2784 | ||||
| 2785 | #line 246 | |||
| 2786 | #if 1 | |||
| 2787 | SIMD_IMPL_INTRIN_1(reinterpret_u64_s16, vu64, vs16)static PyObject *simd__intrin_reinterpret_u64_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 2788 | #endif // simd_sup2 | |||
| 2789 | ||||
| 2790 | #line 246 | |||
| 2791 | #if 1 | |||
| 2792 | SIMD_IMPL_INTRIN_1(reinterpret_s64_s16, vs64, vs16)static PyObject *simd__intrin_reinterpret_s64_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vs16}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 2793 | #endif // simd_sup2 | |||
| 2794 | ||||
| 2795 | #line 246 | |||
| 2796 | #if 1 | |||
| 2797 | SIMD_IMPL_INTRIN_1(reinterpret_f32_s16, vf32, vs16)static PyObject *simd__intrin_reinterpret_f32_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vs16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 2798 | #endif // simd_sup2 | |||
| 2799 | ||||
| 2800 | #line 246 | |||
| 2801 | #if NPY_SIMD_F641 | |||
| 2802 | SIMD_IMPL_INTRIN_1(reinterpret_f64_s16, vf64, vs16)static PyObject *simd__intrin_reinterpret_f64_s16 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_s16", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vs16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 2803 | #endif // simd_sup2 | |||
| 2804 | ||||
| 2805 | ||||
| 2806 | /** | |||
| 2807 | * special definition due to the nature of intrinsics | |||
| 2808 | * npyv_setf_s16 and npy_set_s16. | |||
| 2809 | */ | |||
| 2810 | #line 258 | |||
| 2811 | static PyObject * | |||
| 2812 | simd__intrin_setf_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2813 | { | |||
| 2814 | npyv_lanetype_s16 *data = simd_sequence_from_iterable(args, simd_data_qs16, npyv_nlanes_s168); | |||
| 2815 | if (data == NULL((void*)0)) { | |||
| 2816 | return NULL((void*)0); | |||
| 2817 | } | |||
| 2818 | simd_data r = {.vs16 = npyv_setf_s16(npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2819 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2820 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2821 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2822 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2823 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2824 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2825 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2826 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2827 | data[64] // for setfnpyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8])) | |||
| 2828 | )npyv__setr_epi16((short)(data[1]), (short)(data[2]), (short)( data[3]), (short)(data[4]), (short)(data[5]), (short)(data[6] ), (short)(data[7]), (short)(data[8]))}; | |||
| 2829 | simd_sequence_free(data); | |||
| 2830 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs16); | |||
| 2831 | } | |||
| 2832 | ||||
| 2833 | #line 258 | |||
| 2834 | static PyObject * | |||
| 2835 | simd__intrin_set_s16(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 2836 | { | |||
| 2837 | npyv_lanetype_s16 *data = simd_sequence_from_iterable(args, simd_data_qs16, npyv_nlanes_s168); | |||
| 2838 | if (data == NULL((void*)0)) { | |||
| 2839 | return NULL((void*)0); | |||
| 2840 | } | |||
| 2841 | simd_data r = {.vs16 = npyv_set_s16(npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2842 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2843 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2844 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2845 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2846 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2847 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2848 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2849 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2850 | data[64] // for setfnpyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7])) | |||
| 2851 | )npyv__setr_epi16((short)(data[0]), (short)(data[1]), (short)( data[2]), (short)(data[3]), (short)(data[4]), (short)(data[5] ), (short)(data[6]), (short)(data[7]))}; | |||
| 2852 | simd_sequence_free(data); | |||
| 2853 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs16); | |||
| 2854 | } | |||
| 2855 | ||||
| 2856 | ||||
| 2857 | /*************************** | |||
| 2858 | * Reorder | |||
| 2859 | ***************************/ | |||
| 2860 | #line 287 | |||
| 2861 | SIMD_IMPL_INTRIN_2(combinel_s16, vs16, vs16, vs16)static PyObject *simd__intrin_combinel_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_s16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs16 = _mm_unpacklo_epi64 ( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2862 | ||||
| 2863 | #line 287 | |||
| 2864 | SIMD_IMPL_INTRIN_2(combineh_s16, vs16, vs16, vs16)static PyObject *simd__intrin_combineh_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_s16" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs16 = _mm_unpackhi_epi64 ( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2865 | ||||
| 2866 | ||||
| 2867 | #line 293 | |||
| 2868 | SIMD_IMPL_INTRIN_2(combine_s16, vs16x2, vs16, vs16)static PyObject *simd__intrin_combine_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_s16", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs16x2 = npyv__combine ( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16x2 }; return simd_arg_to_obj(&ret); } | |||
| 2869 | ||||
| 2870 | #line 293 | |||
| 2871 | SIMD_IMPL_INTRIN_2(zip_s16, vs16x2, vs16, vs16)static PyObject *simd__intrin_zip_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16x2 = npyv_zip_s16( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16x2 }; return simd_arg_to_obj(&ret); } | |||
| 2872 | ||||
| 2873 | ||||
| 2874 | #if 1 | |||
| 2875 | SIMD_IMPL_INTRIN_1(rev64_s16, vs16, vs16)static PyObject *simd__intrin_rev64_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "rev64_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs16 = npyv_rev64_u16( arg.data.vs16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2876 | #endif | |||
| 2877 | ||||
| 2878 | /*************************** | |||
| 2879 | * Operators | |||
| 2880 | ***************************/ | |||
| 2881 | #if 15 > 0 | |||
| 2882 | SIMD_IMPL_INTRIN_2(shl_s16, vs16, vs16, u8)static PyObject *simd__intrin_shl_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_sll_epi16(arg1.data.vs16, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2883 | SIMD_IMPL_INTRIN_2(shr_s16, vs16, vs16, u8)static PyObject *simd__intrin_shr_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_sra_epi16(arg1.data.vs16, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2884 | // immediate constant | |||
| 2885 | SIMD_IMPL_INTRIN_2IMM(shli_s16, vs16, vs16, 15)static PyObject *simd__intrin_shli_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs16 = 0 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 0) : 1 == arg2.data.u8 ? _mm_slli_epi16 (arg1.data.vs16, 1) : 2 == arg2.data.u8 ? _mm_slli_epi16(arg1 .data.vs16, 2) : 3 == arg2.data.u8 ? _mm_slli_epi16(arg1.data .vs16, 3) : 4 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 5) : 6 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 6) : 7 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi16(arg1.data.vs16, 8) : 9 == arg2.data .u8 ? _mm_slli_epi16(arg1.data.vs16, 9) : 10 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 10) : 11 == arg2.data.u8 ? _mm_slli_epi16 (arg1.data.vs16, 11) : 12 == arg2.data.u8 ? _mm_slli_epi16(arg1 .data.vs16, 12) : 13 == arg2.data.u8 ? _mm_slli_epi16(arg1.data .vs16, 13) : 14 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi16(arg1.data.vs16, 15 ) : data.vs16; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(& ret); } | |||
| 2886 | SIMD_IMPL_INTRIN_2IMM(shri_s16, vs16, vs16, 16)static PyObject *simd__intrin_shri_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs16 = 1 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 1) : 2 == arg2.data.u8 ? _mm_srai_epi16 (arg1.data.vs16, 2) : 3 == arg2.data.u8 ? _mm_srai_epi16(arg1 .data.vs16, 3) : 4 == arg2.data.u8 ? _mm_srai_epi16(arg1.data .vs16, 4) : 5 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16 , 5) : 6 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 6) : 7 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 7) : 8 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 8) : 9 == arg2 .data.u8 ? _mm_srai_epi16(arg1.data.vs16, 9) : 10 == arg2.data .u8 ? _mm_srai_epi16(arg1.data.vs16, 10) : 11 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 11) : 12 == arg2.data.u8 ? _mm_srai_epi16 (arg1.data.vs16, 12) : 13 == arg2.data.u8 ? _mm_srai_epi16(arg1 .data.vs16, 13) : 14 == arg2.data.u8 ? _mm_srai_epi16(arg1.data .vs16, 14) : 15 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16 , 15) : 16 == arg2.data.u8 ? _mm_srai_epi16(arg1.data.vs16, 16 ) : data.vs16; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(& ret); } | |||
| 2887 | #endif // shl_imm | |||
| 2888 | ||||
| 2889 | #line 314 | |||
| 2890 | SIMD_IMPL_INTRIN_2(and_s16, vs16, vs16, vs16)static PyObject *simd__intrin_and_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""and_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_and_si128( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2891 | ||||
| 2892 | #line 314 | |||
| 2893 | SIMD_IMPL_INTRIN_2(or_s16, vs16, vs16, vs16)static PyObject *simd__intrin_or_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""or_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_or_si128( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2894 | ||||
| 2895 | #line 314 | |||
| 2896 | SIMD_IMPL_INTRIN_2(xor_s16, vs16, vs16, vs16)static PyObject *simd__intrin_xor_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_xor_si128( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2897 | ||||
| 2898 | ||||
| 2899 | SIMD_IMPL_INTRIN_1(not_s16, vs16, vs16)static PyObject *simd__intrin_not_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "not_s16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs16 = _mm_xor_si128(arg.data.vs16, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 2900 | ||||
| 2901 | #line 322 | |||
| 2902 | SIMD_IMPL_INTRIN_2(cmpeq_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmpeq_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpeq_epi16( arg1.data.vs16 , arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2903 | ||||
| 2904 | #line 322 | |||
| 2905 | SIMD_IMPL_INTRIN_2(cmpneq_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmpneq_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_xor_si128(_mm_cmpeq_epi16(arg1 .data.vs16, arg2.data.vs16), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2906 | ||||
| 2907 | #line 322 | |||
| 2908 | SIMD_IMPL_INTRIN_2(cmpgt_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmpgt_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpgt_epi16( arg1.data.vs16 , arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2909 | ||||
| 2910 | #line 322 | |||
| 2911 | SIMD_IMPL_INTRIN_2(cmpge_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmpge_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_xor_si128(_mm_cmpgt_epi16(arg2 .data.vs16, arg1.data.vs16), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2912 | ||||
| 2913 | #line 322 | |||
| 2914 | SIMD_IMPL_INTRIN_2(cmplt_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmplt_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_cmpgt_epi16(arg2.data.vs16, arg1.data.vs16)}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2915 | ||||
| 2916 | #line 322 | |||
| 2917 | SIMD_IMPL_INTRIN_2(cmple_s16, vb16, vs16, vs16)static PyObject *simd__intrin_cmple_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_xor_si128(_mm_cmpgt_epi16(arg1 .data.vs16, arg2.data.vs16), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(& ret); } | |||
| 2918 | ||||
| 2919 | ||||
| 2920 | /*************************** | |||
| 2921 | * Conversion | |||
| 2922 | ***************************/ | |||
| 2923 | SIMD_IMPL_INTRIN_1(cvt_s16_b16, vs16, vb16)static PyObject *simd__intrin_cvt_s16_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb16}; if (!PyArg_ParseTuple( args, "O&:" "cvt_s16_b16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs16 = arg.data.vb16}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2924 | SIMD_IMPL_INTRIN_1(cvt_b16_s16, vb16, vs16)static PyObject *simd__intrin_cvt_b16_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b16_s16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb16 = arg.data.vs16}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 2925 | #if 0 | |||
| 2926 | SIMD_IMPL_INTRIN_1(expand_s16_s16, vs16x2, vs16)static PyObject *simd__intrin_expand_s16_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "expand_s16_s16", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vs16x2 = npyv_expand_s16_s16( arg.data .vs16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16x2 }; return simd_arg_to_obj(&ret ); } | |||
| 2927 | #endif // expand_sup | |||
| 2928 | /*************************** | |||
| 2929 | * Arithmetic | |||
| 2930 | ***************************/ | |||
| 2931 | #line 339 | |||
| 2932 | SIMD_IMPL_INTRIN_2(add_s16, vs16, vs16, vs16)static PyObject *simd__intrin_add_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""add_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_add_epi16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2933 | ||||
| 2934 | #line 339 | |||
| 2935 | SIMD_IMPL_INTRIN_2(sub_s16, vs16, vs16, vs16)static PyObject *simd__intrin_sub_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_sub_epi16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2936 | ||||
| 2937 | ||||
| 2938 | #if 1 | |||
| 2939 | #line 346 | |||
| 2940 | SIMD_IMPL_INTRIN_2(adds_s16, vs16, vs16, vs16)static PyObject *simd__intrin_adds_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_adds_epi16( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2941 | ||||
| 2942 | #line 346 | |||
| 2943 | SIMD_IMPL_INTRIN_2(subs_s16, vs16, vs16, vs16)static PyObject *simd__intrin_subs_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_subs_epi16( arg1.data.vs16, arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2944 | ||||
| 2945 | #endif // sat_sup | |||
| 2946 | ||||
| 2947 | #if 1 | |||
| 2948 | SIMD_IMPL_INTRIN_2(mul_s16, vs16, vs16, vs16)static PyObject *simd__intrin_mul_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_mullo_epi16( arg1.data.vs16 , arg2.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2949 | #endif // mul_sup | |||
| 2950 | ||||
| 2951 | #if 0 | |||
| 2952 | SIMD_IMPL_INTRIN_2(div_s16, vs16, vs16, vs16)static PyObject *simd__intrin_div_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""div_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = npyv_div_s16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2953 | #endif // div_sup | |||
| 2954 | ||||
| 2955 | #if 1 | |||
| 2956 | SIMD_IMPL_INTRIN_1(divisor_s16, vs16x3, s16)static PyObject *simd__intrin_divisor_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s16}; if (!PyArg_ParseTuple( args, "O&:" "divisor_s16", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs16x3 = npyv_divisor_s16( arg.data.s16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs16x3 }; return simd_arg_to_obj(&ret); } | |||
| 2957 | SIMD_IMPL_INTRIN_2(divc_s16, vs16, vs16, vs16x3)static PyObject *simd__intrin_divc_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = npyv_divc_s16( arg1.data.vs16, arg2 .data.vs16x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2958 | #endif // intdiv_sup | |||
| 2959 | ||||
| 2960 | #if 0 | |||
| 2961 | #line 367 | |||
| 2962 | SIMD_IMPL_INTRIN_3(muladd_s16, vs16, vs16, vs16, vs16)static PyObject *simd__intrin_muladd_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_muladd_s16 ( arg1.data.vs16, arg2.data.vs16, arg3.data.vs16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2963 | ||||
| 2964 | #line 367 | |||
| 2965 | SIMD_IMPL_INTRIN_3(mulsub_s16, vs16, vs16, vs16, vs16)static PyObject *simd__intrin_mulsub_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_mulsub_s16 ( arg1.data.vs16, arg2.data.vs16, arg3.data.vs16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2966 | ||||
| 2967 | #line 367 | |||
| 2968 | SIMD_IMPL_INTRIN_3(nmuladd_s16, vs16, vs16, vs16, vs16)static PyObject *simd__intrin_nmuladd_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_nmuladd_s16 ( arg1.data.vs16, arg2.data.vs16, arg3.data.vs16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2969 | ||||
| 2970 | #line 367 | |||
| 2971 | SIMD_IMPL_INTRIN_3(nmulsub_s16, vs16, vs16, vs16, vs16)static PyObject *simd__intrin_nmulsub_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs16 = npyv_nmulsub_s16 ( arg1.data.vs16, arg2.data.vs16, arg3.data.vs16 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2972 | ||||
| 2973 | #endif // fused_sup | |||
| 2974 | ||||
| 2975 | #if 0 | |||
| 2976 | SIMD_IMPL_INTRIN_1(sum_s16, s16, vs16)static PyObject *simd__intrin_sum_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "sum_s16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.s16 = npyv_sum_s16( arg.data.vs16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_s16 }; return simd_arg_to_obj(&ret); } | |||
| 2977 | #endif // sum_sup | |||
| 2978 | ||||
| 2979 | #if 0 | |||
| 2980 | SIMD_IMPL_INTRIN_1(sumup_s16, s16, vs16)static PyObject *simd__intrin_sumup_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "sumup_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.s16 = npyv_sumup_s16( arg.data.vs16 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_s16 }; return simd_arg_to_obj(&ret); } | |||
| 2981 | #endif // sumup_sup | |||
| 2982 | ||||
| 2983 | /*************************** | |||
| 2984 | * Math | |||
| 2985 | ***************************/ | |||
| 2986 | #if 0 | |||
| 2987 | #line 386 | |||
| 2988 | SIMD_IMPL_INTRIN_1(sqrt_s16, vs16, vs16)static PyObject *simd__intrin_sqrt_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_s16", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs16 = npyv_sqrt_s16( arg.data.vs16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2989 | ||||
| 2990 | #line 386 | |||
| 2991 | SIMD_IMPL_INTRIN_1(recip_s16, vs16, vs16)static PyObject *simd__intrin_recip_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "recip_s16", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs16 = npyv_recip_s16( arg.data.vs16 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2992 | ||||
| 2993 | #line 386 | |||
| 2994 | SIMD_IMPL_INTRIN_1(abs_s16, vs16, vs16)static PyObject *simd__intrin_abs_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "abs_s16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs16 = npyv_abs_s16( arg.data.vs16 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2995 | ||||
| 2996 | #line 386 | |||
| 2997 | SIMD_IMPL_INTRIN_1(square_s16, vs16, vs16)static PyObject *simd__intrin_square_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&:" "square_s16", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs16 = npyv_square_s16( arg.data.vs16 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 2998 | ||||
| 2999 | #endif | |||
| 3000 | ||||
| 3001 | #line 393 | |||
| 3002 | SIMD_IMPL_INTRIN_2(max_s16, vs16, vs16, vs16)static PyObject *simd__intrin_max_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""max_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_max_epi16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 3003 | ||||
| 3004 | #line 393 | |||
| 3005 | SIMD_IMPL_INTRIN_2(min_s16, vs16, vs16, vs16)static PyObject *simd__intrin_min_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""min_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = _mm_min_epi16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 3006 | ||||
| 3007 | ||||
| 3008 | #if 0 | |||
| 3009 | #line 400 | |||
| 3010 | SIMD_IMPL_INTRIN_2(maxp_s16, vs16, vs16, vs16)static PyObject *simd__intrin_maxp_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = npyv_maxp_s16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 3011 | ||||
| 3012 | #line 400 | |||
| 3013 | SIMD_IMPL_INTRIN_2(minp_s16, vs16, vs16, vs16)static PyObject *simd__intrin_minp_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_s16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs16 = npyv_minp_s16( arg1.data.vs16, arg2 .data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret); } | |||
| 3014 | ||||
| 3015 | #endif | |||
| 3016 | ||||
| 3017 | /*************************** | |||
| 3018 | * Mask operations | |||
| 3019 | ***************************/ | |||
| 3020 | #line 410 | |||
| 3021 | SIMD_IMPL_INTRIN_4(ifadd_s16, vs16, vb16, vs16, vs16, vs16)static PyObject *simd__intrin_ifadd_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; simd_arg arg4 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_s16", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs16 = npyv_ifadd_s16 ( arg1.data.vb16, arg2.data.vs16, arg3.data.vs16, arg4.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(& ret); } | |||
| 3022 | ||||
| 3023 | #line 410 | |||
| 3024 | SIMD_IMPL_INTRIN_4(ifsub_s16, vs16, vb16, vs16, vs16, vs16)static PyObject *simd__intrin_ifsub_s16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vs16 }; simd_arg arg3 = {.dtype = simd_data_vs16}; simd_arg arg4 = {.dtype = simd_data_vs16}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_s16", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs16 = npyv_ifsub_s16 ( arg1.data.vb16, arg2.data.vs16, arg3.data.vs16, arg4.data.vs16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj(& ret); } | |||
| 3025 | ||||
| 3026 | ||||
| 3027 | #endif // simd_sup | |||
| 3028 | ||||
| 3029 | #line 34 | |||
| 3030 | #if 1 | |||
| 3031 | /*************************** | |||
| 3032 | * Memory | |||
| 3033 | ***************************/ | |||
| 3034 | #line 41 | |||
| 3035 | SIMD_IMPL_INTRIN_1(load_u32, vu32, qu32)static PyObject *simd__intrin_load_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu32}; if (!PyArg_ParseTuple( args, "O&:" "load_u32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu32 = npyv_load_u32( arg.data.qu32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3036 | ||||
| 3037 | #line 41 | |||
| 3038 | SIMD_IMPL_INTRIN_1(loada_u32, vu32, qu32)static PyObject *simd__intrin_loada_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu32}; if (!PyArg_ParseTuple( args, "O&:" "loada_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu32 = npyv_loada_u32( arg.data.qu32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3039 | ||||
| 3040 | #line 41 | |||
| 3041 | SIMD_IMPL_INTRIN_1(loads_u32, vu32, qu32)static PyObject *simd__intrin_loads_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu32}; if (!PyArg_ParseTuple( args, "O&:" "loads_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu32 = npyv_loads_u32( arg.data.qu32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3042 | ||||
| 3043 | #line 41 | |||
| 3044 | SIMD_IMPL_INTRIN_1(loadl_u32, vu32, qu32)static PyObject *simd__intrin_loadl_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu32}; if (!PyArg_ParseTuple( args, "O&:" "loadl_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu32 = npyv_loadl_u32( arg.data.qu32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3045 | ||||
| 3046 | #line 46 | |||
| 3047 | // special definition due to the nature of store | |||
| 3048 | static PyObject * | |||
| 3049 | simd__intrin_store_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3050 | { | |||
| 3051 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3052 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3053 | if (!PyArg_ParseTuple( | |||
| 3054 | args, "O&O&:store_u32", | |||
| 3055 | simd_arg_converter, &seq_arg, | |||
| 3056 | simd_arg_converter, &vec_arg | |||
| 3057 | )) { | |||
| 3058 | return NULL((void*)0); | |||
| 3059 | } | |||
| 3060 | npyv_store_u32(seq_arg.data.qu32, vec_arg.data.vu32); | |||
| 3061 | // write-back | |||
| 3062 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3063 | simd_arg_free(&seq_arg); | |||
| 3064 | return NULL((void*)0); | |||
| 3065 | } | |||
| 3066 | simd_arg_free(&seq_arg); | |||
| 3067 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3068 | } | |||
| 3069 | ||||
| 3070 | #line 46 | |||
| 3071 | // special definition due to the nature of storea | |||
| 3072 | static PyObject * | |||
| 3073 | simd__intrin_storea_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3074 | { | |||
| 3075 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3076 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3077 | if (!PyArg_ParseTuple( | |||
| 3078 | args, "O&O&:storea_u32", | |||
| 3079 | simd_arg_converter, &seq_arg, | |||
| 3080 | simd_arg_converter, &vec_arg | |||
| 3081 | )) { | |||
| 3082 | return NULL((void*)0); | |||
| 3083 | } | |||
| 3084 | npyv_storea_u32(seq_arg.data.qu32, vec_arg.data.vu32); | |||
| 3085 | // write-back | |||
| 3086 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3087 | simd_arg_free(&seq_arg); | |||
| 3088 | return NULL((void*)0); | |||
| 3089 | } | |||
| 3090 | simd_arg_free(&seq_arg); | |||
| 3091 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3092 | } | |||
| 3093 | ||||
| 3094 | #line 46 | |||
| 3095 | // special definition due to the nature of stores | |||
| 3096 | static PyObject * | |||
| 3097 | simd__intrin_stores_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3098 | { | |||
| 3099 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3100 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3101 | if (!PyArg_ParseTuple( | |||
| 3102 | args, "O&O&:stores_u32", | |||
| 3103 | simd_arg_converter, &seq_arg, | |||
| 3104 | simd_arg_converter, &vec_arg | |||
| 3105 | )) { | |||
| 3106 | return NULL((void*)0); | |||
| 3107 | } | |||
| 3108 | npyv_stores_u32(seq_arg.data.qu32, vec_arg.data.vu32); | |||
| 3109 | // write-back | |||
| 3110 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3111 | simd_arg_free(&seq_arg); | |||
| 3112 | return NULL((void*)0); | |||
| 3113 | } | |||
| 3114 | simd_arg_free(&seq_arg); | |||
| 3115 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3116 | } | |||
| 3117 | ||||
| 3118 | #line 46 | |||
| 3119 | // special definition due to the nature of storel | |||
| 3120 | static PyObject * | |||
| 3121 | simd__intrin_storel_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3122 | { | |||
| 3123 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3124 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3125 | if (!PyArg_ParseTuple( | |||
| 3126 | args, "O&O&:storel_u32", | |||
| 3127 | simd_arg_converter, &seq_arg, | |||
| 3128 | simd_arg_converter, &vec_arg | |||
| 3129 | )) { | |||
| 3130 | return NULL((void*)0); | |||
| 3131 | } | |||
| 3132 | npyv_storel_u32(seq_arg.data.qu32, vec_arg.data.vu32); | |||
| 3133 | // write-back | |||
| 3134 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3135 | simd_arg_free(&seq_arg); | |||
| 3136 | return NULL((void*)0); | |||
| 3137 | } | |||
| 3138 | simd_arg_free(&seq_arg); | |||
| 3139 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3140 | } | |||
| 3141 | ||||
| 3142 | #line 46 | |||
| 3143 | // special definition due to the nature of storeh | |||
| 3144 | static PyObject * | |||
| 3145 | simd__intrin_storeh_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3146 | { | |||
| 3147 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3148 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3149 | if (!PyArg_ParseTuple( | |||
| 3150 | args, "O&O&:storeh_u32", | |||
| 3151 | simd_arg_converter, &seq_arg, | |||
| 3152 | simd_arg_converter, &vec_arg | |||
| 3153 | )) { | |||
| 3154 | return NULL((void*)0); | |||
| 3155 | } | |||
| 3156 | npyv_storeh_u32(seq_arg.data.qu32, vec_arg.data.vu32); | |||
| 3157 | // write-back | |||
| 3158 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3159 | simd_arg_free(&seq_arg); | |||
| 3160 | return NULL((void*)0); | |||
| 3161 | } | |||
| 3162 | simd_arg_free(&seq_arg); | |||
| 3163 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3164 | } | |||
| 3165 | ||||
| 3166 | ||||
| 3167 | /**************************************** | |||
| 3168 | * Non-contiguous/Partial Memory access | |||
| 3169 | ****************************************/ | |||
| 3170 | #if 1 | |||
| 3171 | // Partial Load | |||
| 3172 | SIMD_IMPL_INTRIN_3(load_till_u32, vu32, qu32, u32, u32)static PyObject *simd__intrin_load_till_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu32}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_u32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_load_till_u32 ( arg1.data.qu32, arg2.data.u32, arg3.data.u32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3173 | SIMD_IMPL_INTRIN_2(load_tillz_u32, vu32, qu32, u32)static PyObject *simd__intrin_load_tillz_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu32}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_u32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu32 = npyv_load_tillz_u32 ( arg1.data.qu32, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3174 | ||||
| 3175 | // Partial Store | |||
| 3176 | static PyObject * | |||
| 3177 | simd__intrin_store_till_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3178 | { | |||
| 3179 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3180 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3181 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3182 | if (!PyArg_ParseTuple( | |||
| 3183 | args, "O&O&O&:store_till_u32", | |||
| 3184 | simd_arg_converter, &seq_arg, | |||
| 3185 | simd_arg_converter, &nlane_arg, | |||
| 3186 | simd_arg_converter, &vec_arg | |||
| 3187 | )) { | |||
| 3188 | return NULL((void*)0); | |||
| 3189 | } | |||
| 3190 | npyv_store_till_u32( | |||
| 3191 | seq_arg.data.qu32, nlane_arg.data.u32, vec_arg.data.vu32 | |||
| 3192 | ); | |||
| 3193 | // write-back | |||
| 3194 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3195 | simd_arg_free(&seq_arg); | |||
| 3196 | return NULL((void*)0); | |||
| 3197 | } | |||
| 3198 | simd_arg_free(&seq_arg); | |||
| 3199 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3200 | } | |||
| 3201 | ||||
| 3202 | // Non-contiguous Load | |||
| 3203 | #line 112 | |||
| 3204 | static PyObject * | |||
| 3205 | simd__intrin_loadn_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3206 | { | |||
| 3207 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3208 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3209 | #if 0 | |||
| 3210 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3211 | #endif // till | |||
| 3212 | #if 0 | |||
| 3213 | simd_arg fill_arg = {.dtype = simd_data_u32}; | |||
| 3214 | #endif | |||
| 3215 | if (!PyArg_ParseTuple( | |||
| 3216 | args, "O&O&:loadn_u32", | |||
| 3217 | simd_arg_converter, &seq_arg, | |||
| 3218 | simd_arg_converter, &stride_arg | |||
| 3219 | #if 0 | |||
| 3220 | ,simd_arg_converter, &nlane_arg | |||
| 3221 | #endif | |||
| 3222 | #if 0 | |||
| 3223 | ,simd_arg_converter, &fill_arg | |||
| 3224 | #endif | |||
| 3225 | )) { | |||
| 3226 | return NULL((void*)0); | |||
| 3227 | } | |||
| 3228 | npyv_lanetype_u32 *seq_ptr = seq_arg.data.qu32; | |||
| 3229 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3230 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3231 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u324; | |||
| 3232 | if (stride < 0) { | |||
| 3233 | seq_ptr += cur_seq_len -1; | |||
| 3234 | min_seq_len = -min_seq_len; | |||
| 3235 | } | |||
| 3236 | if (cur_seq_len < min_seq_len) { | |||
| 3237 | PyErr_Format(PyExc_ValueError, | |||
| 3238 | "loadn_u32(), according to provided stride %d, the " | |||
| 3239 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3240 | stride, min_seq_len, cur_seq_len | |||
| 3241 | ); | |||
| 3242 | goto err; | |||
| 3243 | } | |||
| 3244 | npyv_u32 rvec = npyv_loadn_u32( | |||
| 3245 | seq_ptr, stride | |||
| 3246 | #if 0 | |||
| 3247 | , nlane_arg.data.u32 | |||
| 3248 | #endif | |||
| 3249 | #if 0 | |||
| 3250 | , fill_arg.data.u32 | |||
| 3251 | #endif | |||
| 3252 | ); | |||
| 3253 | simd_arg ret = { | |||
| 3254 | .dtype = simd_data_vu32, .data = {.vu32=rvec} | |||
| 3255 | }; | |||
| 3256 | simd_arg_free(&seq_arg); | |||
| 3257 | return simd_arg_to_obj(&ret); | |||
| 3258 | err: | |||
| 3259 | simd_arg_free(&seq_arg); | |||
| 3260 | return NULL((void*)0); | |||
| 3261 | } | |||
| 3262 | ||||
| 3263 | #line 112 | |||
| 3264 | static PyObject * | |||
| 3265 | simd__intrin_loadn_till_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3266 | { | |||
| 3267 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3268 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3269 | #if 1 | |||
| 3270 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3271 | #endif // till | |||
| 3272 | #if 1 | |||
| 3273 | simd_arg fill_arg = {.dtype = simd_data_u32}; | |||
| 3274 | #endif | |||
| 3275 | if (!PyArg_ParseTuple( | |||
| 3276 | args, "O&O&O&O&:loadn_till_u32", | |||
| 3277 | simd_arg_converter, &seq_arg, | |||
| 3278 | simd_arg_converter, &stride_arg | |||
| 3279 | #if 1 | |||
| 3280 | ,simd_arg_converter, &nlane_arg | |||
| 3281 | #endif | |||
| 3282 | #if 1 | |||
| 3283 | ,simd_arg_converter, &fill_arg | |||
| 3284 | #endif | |||
| 3285 | )) { | |||
| 3286 | return NULL((void*)0); | |||
| 3287 | } | |||
| 3288 | npyv_lanetype_u32 *seq_ptr = seq_arg.data.qu32; | |||
| 3289 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3290 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3291 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u324; | |||
| 3292 | if (stride < 0) { | |||
| 3293 | seq_ptr += cur_seq_len -1; | |||
| 3294 | min_seq_len = -min_seq_len; | |||
| 3295 | } | |||
| 3296 | if (cur_seq_len < min_seq_len) { | |||
| 3297 | PyErr_Format(PyExc_ValueError, | |||
| 3298 | "loadn_till_u32(), according to provided stride %d, the " | |||
| 3299 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3300 | stride, min_seq_len, cur_seq_len | |||
| 3301 | ); | |||
| 3302 | goto err; | |||
| 3303 | } | |||
| 3304 | npyv_u32 rvec = npyv_loadn_till_u32( | |||
| 3305 | seq_ptr, stride | |||
| 3306 | #if 1 | |||
| 3307 | , nlane_arg.data.u32 | |||
| 3308 | #endif | |||
| 3309 | #if 1 | |||
| 3310 | , fill_arg.data.u32 | |||
| 3311 | #endif | |||
| 3312 | ); | |||
| 3313 | simd_arg ret = { | |||
| 3314 | .dtype = simd_data_vu32, .data = {.vu32=rvec} | |||
| 3315 | }; | |||
| 3316 | simd_arg_free(&seq_arg); | |||
| 3317 | return simd_arg_to_obj(&ret); | |||
| 3318 | err: | |||
| 3319 | simd_arg_free(&seq_arg); | |||
| 3320 | return NULL((void*)0); | |||
| 3321 | } | |||
| 3322 | ||||
| 3323 | #line 112 | |||
| 3324 | static PyObject * | |||
| 3325 | simd__intrin_loadn_tillz_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3326 | { | |||
| 3327 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3328 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3329 | #if 1 | |||
| 3330 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3331 | #endif // till | |||
| 3332 | #if 0 | |||
| 3333 | simd_arg fill_arg = {.dtype = simd_data_u32}; | |||
| 3334 | #endif | |||
| 3335 | if (!PyArg_ParseTuple( | |||
| 3336 | args, "O&O&O&:loadn_tillz_u32", | |||
| 3337 | simd_arg_converter, &seq_arg, | |||
| 3338 | simd_arg_converter, &stride_arg | |||
| 3339 | #if 1 | |||
| 3340 | ,simd_arg_converter, &nlane_arg | |||
| 3341 | #endif | |||
| 3342 | #if 0 | |||
| 3343 | ,simd_arg_converter, &fill_arg | |||
| 3344 | #endif | |||
| 3345 | )) { | |||
| 3346 | return NULL((void*)0); | |||
| 3347 | } | |||
| 3348 | npyv_lanetype_u32 *seq_ptr = seq_arg.data.qu32; | |||
| 3349 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3350 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3351 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u324; | |||
| 3352 | if (stride < 0) { | |||
| 3353 | seq_ptr += cur_seq_len -1; | |||
| 3354 | min_seq_len = -min_seq_len; | |||
| 3355 | } | |||
| 3356 | if (cur_seq_len < min_seq_len) { | |||
| 3357 | PyErr_Format(PyExc_ValueError, | |||
| 3358 | "loadn_tillz_u32(), according to provided stride %d, the " | |||
| 3359 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3360 | stride, min_seq_len, cur_seq_len | |||
| 3361 | ); | |||
| 3362 | goto err; | |||
| 3363 | } | |||
| 3364 | npyv_u32 rvec = npyv_loadn_tillz_u32( | |||
| 3365 | seq_ptr, stride | |||
| 3366 | #if 1 | |||
| 3367 | , nlane_arg.data.u32 | |||
| 3368 | #endif | |||
| 3369 | #if 0 | |||
| 3370 | , fill_arg.data.u32 | |||
| 3371 | #endif | |||
| 3372 | ); | |||
| 3373 | simd_arg ret = { | |||
| 3374 | .dtype = simd_data_vu32, .data = {.vu32=rvec} | |||
| 3375 | }; | |||
| 3376 | simd_arg_free(&seq_arg); | |||
| 3377 | return simd_arg_to_obj(&ret); | |||
| 3378 | err: | |||
| 3379 | simd_arg_free(&seq_arg); | |||
| 3380 | return NULL((void*)0); | |||
| 3381 | } | |||
| 3382 | ||||
| 3383 | ||||
| 3384 | // Non-contiguous Store | |||
| 3385 | #line 178 | |||
| 3386 | static PyObject * | |||
| 3387 | simd__intrin_storen_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3388 | { | |||
| 3389 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3390 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3391 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3392 | #if 0 | |||
| 3393 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3394 | #endif | |||
| 3395 | if (!PyArg_ParseTuple( | |||
| 3396 | args, "O&O&O&:storen_u32", | |||
| 3397 | simd_arg_converter, &seq_arg, | |||
| 3398 | simd_arg_converter, &stride_arg | |||
| 3399 | #if 0 | |||
| 3400 | ,simd_arg_converter, &nlane_arg | |||
| 3401 | #endif | |||
| 3402 | ,simd_arg_converter, &vec_arg | |||
| 3403 | )) { | |||
| 3404 | return NULL((void*)0); | |||
| 3405 | } | |||
| 3406 | npyv_lanetype_u32 *seq_ptr = seq_arg.data.qu32; | |||
| 3407 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3408 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3409 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u324; | |||
| 3410 | if (stride < 0) { | |||
| 3411 | seq_ptr += cur_seq_len -1; | |||
| 3412 | min_seq_len = -min_seq_len; | |||
| 3413 | } | |||
| 3414 | // overflow guard | |||
| 3415 | if (cur_seq_len < min_seq_len) { | |||
| 3416 | PyErr_Format(PyExc_ValueError, | |||
| 3417 | "storen_u32(), according to provided stride %d, the" | |||
| 3418 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3419 | stride, min_seq_len, cur_seq_len | |||
| 3420 | ); | |||
| 3421 | goto err; | |||
| 3422 | } | |||
| 3423 | npyv_storen_u32( | |||
| 3424 | seq_ptr, stride | |||
| 3425 | #if 0 | |||
| 3426 | ,nlane_arg.data.u32 | |||
| 3427 | #endif | |||
| 3428 | ,vec_arg.data.vu32 | |||
| 3429 | ); | |||
| 3430 | // write-back | |||
| 3431 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3432 | goto err; | |||
| 3433 | } | |||
| 3434 | simd_arg_free(&seq_arg); | |||
| 3435 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3436 | err: | |||
| 3437 | simd_arg_free(&seq_arg); | |||
| 3438 | return NULL((void*)0); | |||
| 3439 | } | |||
| 3440 | ||||
| 3441 | #line 178 | |||
| 3442 | static PyObject * | |||
| 3443 | simd__intrin_storen_till_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3444 | { | |||
| 3445 | simd_arg seq_arg = {.dtype = simd_data_qu32}; | |||
| 3446 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3447 | simd_arg vec_arg = {.dtype = simd_data_vu32}; | |||
| 3448 | #if 1 | |||
| 3449 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3450 | #endif | |||
| 3451 | if (!PyArg_ParseTuple( | |||
| 3452 | args, "O&O&O&O&:storen_u32", | |||
| 3453 | simd_arg_converter, &seq_arg, | |||
| 3454 | simd_arg_converter, &stride_arg | |||
| 3455 | #if 1 | |||
| 3456 | ,simd_arg_converter, &nlane_arg | |||
| 3457 | #endif | |||
| 3458 | ,simd_arg_converter, &vec_arg | |||
| 3459 | )) { | |||
| 3460 | return NULL((void*)0); | |||
| 3461 | } | |||
| 3462 | npyv_lanetype_u32 *seq_ptr = seq_arg.data.qu32; | |||
| 3463 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3464 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3465 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u324; | |||
| 3466 | if (stride < 0) { | |||
| 3467 | seq_ptr += cur_seq_len -1; | |||
| 3468 | min_seq_len = -min_seq_len; | |||
| 3469 | } | |||
| 3470 | // overflow guard | |||
| 3471 | if (cur_seq_len < min_seq_len) { | |||
| 3472 | PyErr_Format(PyExc_ValueError, | |||
| 3473 | "storen_till_u32(), according to provided stride %d, the" | |||
| 3474 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3475 | stride, min_seq_len, cur_seq_len | |||
| 3476 | ); | |||
| 3477 | goto err; | |||
| 3478 | } | |||
| 3479 | npyv_storen_till_u32( | |||
| 3480 | seq_ptr, stride | |||
| 3481 | #if 1 | |||
| 3482 | ,nlane_arg.data.u32 | |||
| 3483 | #endif | |||
| 3484 | ,vec_arg.data.vu32 | |||
| 3485 | ); | |||
| 3486 | // write-back | |||
| 3487 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu32, simd_data_qu32)) { | |||
| 3488 | goto err; | |||
| 3489 | } | |||
| 3490 | simd_arg_free(&seq_arg); | |||
| 3491 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3492 | err: | |||
| 3493 | simd_arg_free(&seq_arg); | |||
| 3494 | return NULL((void*)0); | |||
| 3495 | } | |||
| 3496 | ||||
| 3497 | #endif // 1 | |||
| 3498 | ||||
| 3499 | /*************************** | |||
| 3500 | * Misc | |||
| 3501 | ***************************/ | |||
| 3502 | SIMD_IMPL_INTRIN_0(zero_u32, vu32)static PyObject *simd__intrin_zero_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_u32") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vu32, .data = {.vu32 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 3503 | SIMD_IMPL_INTRIN_1(setall_u32, vu32, u32)static PyObject *simd__intrin_setall_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u32}; if (!PyArg_ParseTuple( args, "O&:" "setall_u32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu32 = _mm_set1_epi32((int)(arg.data.u32 ))}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3504 | SIMD_IMPL_INTRIN_3(select_u32, vu32, vb32, vu32, vu32)static PyObject *simd__intrin_select_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_select_u8 ( arg1.data.vb32, arg2.data.vu32, arg3.data.vu32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3505 | ||||
| 3506 | #line 246 | |||
| 3507 | #if 1 | |||
| 3508 | SIMD_IMPL_INTRIN_1(reinterpret_u8_u32, vu8, vu32)static PyObject *simd__intrin_reinterpret_u8_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vu32 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 3509 | #endif // simd_sup2 | |||
| 3510 | ||||
| 3511 | #line 246 | |||
| 3512 | #if 1 | |||
| 3513 | SIMD_IMPL_INTRIN_1(reinterpret_s8_u32, vs8, vu32)static PyObject *simd__intrin_reinterpret_s8_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vu32 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 3514 | #endif // simd_sup2 | |||
| 3515 | ||||
| 3516 | #line 246 | |||
| 3517 | #if 1 | |||
| 3518 | SIMD_IMPL_INTRIN_1(reinterpret_u16_u32, vu16, vu32)static PyObject *simd__intrin_reinterpret_u16_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 3519 | #endif // simd_sup2 | |||
| 3520 | ||||
| 3521 | #line 246 | |||
| 3522 | #if 1 | |||
| 3523 | SIMD_IMPL_INTRIN_1(reinterpret_s16_u32, vs16, vu32)static PyObject *simd__intrin_reinterpret_s16_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 3524 | #endif // simd_sup2 | |||
| 3525 | ||||
| 3526 | #line 246 | |||
| 3527 | #if 1 | |||
| 3528 | SIMD_IMPL_INTRIN_1(reinterpret_u32_u32, vu32, vu32)static PyObject *simd__intrin_reinterpret_u32_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 3529 | #endif // simd_sup2 | |||
| 3530 | ||||
| 3531 | #line 246 | |||
| 3532 | #if 1 | |||
| 3533 | SIMD_IMPL_INTRIN_1(reinterpret_s32_u32, vs32, vu32)static PyObject *simd__intrin_reinterpret_s32_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 3534 | #endif // simd_sup2 | |||
| 3535 | ||||
| 3536 | #line 246 | |||
| 3537 | #if 1 | |||
| 3538 | SIMD_IMPL_INTRIN_1(reinterpret_u64_u32, vu64, vu32)static PyObject *simd__intrin_reinterpret_u64_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 3539 | #endif // simd_sup2 | |||
| 3540 | ||||
| 3541 | #line 246 | |||
| 3542 | #if 1 | |||
| 3543 | SIMD_IMPL_INTRIN_1(reinterpret_s64_u32, vs64, vu32)static PyObject *simd__intrin_reinterpret_s64_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vu32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 3544 | #endif // simd_sup2 | |||
| 3545 | ||||
| 3546 | #line 246 | |||
| 3547 | #if 1 | |||
| 3548 | SIMD_IMPL_INTRIN_1(reinterpret_f32_u32, vf32, vu32)static PyObject *simd__intrin_reinterpret_f32_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vu32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 3549 | #endif // simd_sup2 | |||
| 3550 | ||||
| 3551 | #line 246 | |||
| 3552 | #if NPY_SIMD_F641 | |||
| 3553 | SIMD_IMPL_INTRIN_1(reinterpret_f64_u32, vf64, vu32)static PyObject *simd__intrin_reinterpret_f64_u32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_u32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vu32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 3554 | #endif // simd_sup2 | |||
| 3555 | ||||
| 3556 | ||||
| 3557 | /** | |||
| 3558 | * special definition due to the nature of intrinsics | |||
| 3559 | * npyv_setf_u32 and npy_set_u32. | |||
| 3560 | */ | |||
| 3561 | #line 258 | |||
| 3562 | static PyObject * | |||
| 3563 | simd__intrin_setf_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3564 | { | |||
| 3565 | npyv_lanetype_u32 *data = simd_sequence_from_iterable(args, simd_data_qu32, npyv_nlanes_u324); | |||
| 3566 | if (data == NULL((void*)0)) { | |||
| 3567 | return NULL((void*)0); | |||
| 3568 | } | |||
| 3569 | simd_data r = {.vu32 = npyv_setf_u32(npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3570 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3571 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3572 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3573 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3574 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3575 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3576 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3577 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3578 | data[64] // for setfnpyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 3579 | )npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4]))}; | |||
| 3580 | simd_sequence_free(data); | |||
| 3581 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu32); | |||
| 3582 | } | |||
| 3583 | ||||
| 3584 | #line 258 | |||
| 3585 | static PyObject * | |||
| 3586 | simd__intrin_set_u32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3587 | { | |||
| 3588 | npyv_lanetype_u32 *data = simd_sequence_from_iterable(args, simd_data_qu32, npyv_nlanes_u324); | |||
| 3589 | if (data == NULL((void*)0)) { | |||
| 3590 | return NULL((void*)0); | |||
| 3591 | } | |||
| 3592 | simd_data r = {.vu32 = npyv_set_u32(npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3593 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3594 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3595 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3596 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3597 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3598 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3599 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3600 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3601 | data[64] // for setfnpyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 3602 | )npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3]))}; | |||
| 3603 | simd_sequence_free(data); | |||
| 3604 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu32); | |||
| 3605 | } | |||
| 3606 | ||||
| 3607 | ||||
| 3608 | /*************************** | |||
| 3609 | * Reorder | |||
| 3610 | ***************************/ | |||
| 3611 | #line 287 | |||
| 3612 | SIMD_IMPL_INTRIN_2(combinel_u32, vu32, vu32, vu32)static PyObject *simd__intrin_combinel_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_u32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu32 = _mm_unpacklo_epi64 ( arg1.data.vu32, arg2.data.vu32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3613 | ||||
| 3614 | #line 287 | |||
| 3615 | SIMD_IMPL_INTRIN_2(combineh_u32, vu32, vu32, vu32)static PyObject *simd__intrin_combineh_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_u32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu32 = _mm_unpackhi_epi64 ( arg1.data.vu32, arg2.data.vu32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3616 | ||||
| 3617 | ||||
| 3618 | #line 293 | |||
| 3619 | SIMD_IMPL_INTRIN_2(combine_u32, vu32x2, vu32, vu32)static PyObject *simd__intrin_combine_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_u32", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu32x2 = npyv__combine ( arg1.data.vu32, arg2.data.vu32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32x2 }; return simd_arg_to_obj(&ret); } | |||
| 3620 | ||||
| 3621 | #line 293 | |||
| 3622 | SIMD_IMPL_INTRIN_2(zip_u32, vu32x2, vu32, vu32)static PyObject *simd__intrin_zip_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32x2 = npyv_zip_u32( arg1.data.vu32, arg2.data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32x2 }; return simd_arg_to_obj(&ret); } | |||
| 3623 | ||||
| 3624 | ||||
| 3625 | #if 1 | |||
| 3626 | SIMD_IMPL_INTRIN_1(rev64_u32, vu32, vu32)static PyObject *simd__intrin_rev64_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "rev64_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu32 = npyv_rev64_u32( arg.data.vu32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3627 | #endif | |||
| 3628 | ||||
| 3629 | /*************************** | |||
| 3630 | * Operators | |||
| 3631 | ***************************/ | |||
| 3632 | #if 31 > 0 | |||
| 3633 | SIMD_IMPL_INTRIN_2(shl_u32, vu32, vu32, u8)static PyObject *simd__intrin_shl_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_sll_epi32(arg1.data.vu32, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3634 | SIMD_IMPL_INTRIN_2(shr_u32, vu32, vu32, u8)static PyObject *simd__intrin_shr_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_srl_epi32(arg1.data.vu32, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3635 | // immediate constant | |||
| 3636 | SIMD_IMPL_INTRIN_2IMM(shli_u32, vu32, vu32, 31)static PyObject *simd__intrin_shli_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu32 = 0 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 0) : 1 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vu32, 1) : 2 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vu32, 2) : 3 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vu32, 3) : 4 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 5) : 6 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 6) : 7 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi32(arg1.data.vu32, 8) : 9 == arg2.data .u8 ? _mm_slli_epi32(arg1.data.vu32, 9) : 10 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 10) : 11 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vu32, 11) : 12 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vu32, 12) : 13 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vu32, 13) : 14 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 15 ) : 16 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 16) : 17 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 17) : 18 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 18) : 19 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 19) : 20 == arg2 .data.u8 ? _mm_slli_epi32(arg1.data.vu32, 20) : 21 == arg2.data .u8 ? _mm_slli_epi32(arg1.data.vu32, 21) : 22 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 22) : 23 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vu32, 23) : 24 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vu32, 24) : 25 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vu32, 25) : 26 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32 , 26) : 27 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 27 ) : 28 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 28) : 29 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 29) : 30 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 30) : 31 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vu32, 31) : data.vu32 ; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3637 | SIMD_IMPL_INTRIN_2IMM(shri_u32, vu32, vu32, 32)static PyObject *simd__intrin_shri_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu32 = 1 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 1) : 2 == arg2.data.u8 ? _mm_srli_epi32 (arg1.data.vu32, 2) : 3 == arg2.data.u8 ? _mm_srli_epi32(arg1 .data.vu32, 3) : 4 == arg2.data.u8 ? _mm_srli_epi32(arg1.data .vu32, 4) : 5 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32 , 5) : 6 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 6) : 7 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 7) : 8 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 8) : 9 == arg2 .data.u8 ? _mm_srli_epi32(arg1.data.vu32, 9) : 10 == arg2.data .u8 ? _mm_srli_epi32(arg1.data.vu32, 10) : 11 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 11) : 12 == arg2.data.u8 ? _mm_srli_epi32 (arg1.data.vu32, 12) : 13 == arg2.data.u8 ? _mm_srli_epi32(arg1 .data.vu32, 13) : 14 == arg2.data.u8 ? _mm_srli_epi32(arg1.data .vu32, 14) : 15 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32 , 15) : 16 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 16 ) : 17 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 17) : 18 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 18) : 19 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 19) : 20 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 20) : 21 == arg2 .data.u8 ? _mm_srli_epi32(arg1.data.vu32, 21) : 22 == arg2.data .u8 ? _mm_srli_epi32(arg1.data.vu32, 22) : 23 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 23) : 24 == arg2.data.u8 ? _mm_srli_epi32 (arg1.data.vu32, 24) : 25 == arg2.data.u8 ? _mm_srli_epi32(arg1 .data.vu32, 25) : 26 == arg2.data.u8 ? _mm_srli_epi32(arg1.data .vu32, 26) : 27 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32 , 27) : 28 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 28 ) : 29 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 29) : 30 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 30) : 31 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 31) : 32 == arg2.data.u8 ? _mm_srli_epi32(arg1.data.vu32, 32) : data.vu32 ; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3638 | #endif // shl_imm | |||
| 3639 | ||||
| 3640 | #line 314 | |||
| 3641 | SIMD_IMPL_INTRIN_2(and_u32, vu32, vu32, vu32)static PyObject *simd__intrin_and_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""and_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_and_si128( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3642 | ||||
| 3643 | #line 314 | |||
| 3644 | SIMD_IMPL_INTRIN_2(or_u32, vu32, vu32, vu32)static PyObject *simd__intrin_or_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""or_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_or_si128( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3645 | ||||
| 3646 | #line 314 | |||
| 3647 | SIMD_IMPL_INTRIN_2(xor_u32, vu32, vu32, vu32)static PyObject *simd__intrin_xor_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_xor_si128( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3648 | ||||
| 3649 | ||||
| 3650 | SIMD_IMPL_INTRIN_1(not_u32, vu32, vu32)static PyObject *simd__intrin_not_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "not_u32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu32 = _mm_xor_si128(arg.data.vu32, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 3651 | ||||
| 3652 | #line 322 | |||
| 3653 | SIMD_IMPL_INTRIN_2(cmpeq_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmpeq_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_cmpeq_epi32( arg1.data.vu32 , arg2.data.vu32 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 3654 | ||||
| 3655 | #line 322 | |||
| 3656 | SIMD_IMPL_INTRIN_2(cmpneq_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmpneq_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(_mm_cmpeq_epi32(arg1 .data.vu32, arg2.data.vu32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 3657 | ||||
| 3658 | #line 322 | |||
| 3659 | SIMD_IMPL_INTRIN_2(cmpgt_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmpgt_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = npyv_cmpgt_u32( arg1.data.vu32, arg2.data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 3660 | ||||
| 3661 | #line 322 | |||
| 3662 | SIMD_IMPL_INTRIN_2(cmpge_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmpge_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(npyv_cmpgt_u32(arg2 .data.vu32, arg1.data.vu32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 3663 | ||||
| 3664 | #line 322 | |||
| 3665 | SIMD_IMPL_INTRIN_2(cmplt_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmplt_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = npyv_cmpgt_u32(arg2.data.vu32, arg1 .data.vu32)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 3666 | ||||
| 3667 | #line 322 | |||
| 3668 | SIMD_IMPL_INTRIN_2(cmple_u32, vb32, vu32, vu32)static PyObject *simd__intrin_cmple_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(npyv_cmpgt_u32(arg1 .data.vu32, arg2.data.vu32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 3669 | ||||
| 3670 | ||||
| 3671 | /*************************** | |||
| 3672 | * Conversion | |||
| 3673 | ***************************/ | |||
| 3674 | SIMD_IMPL_INTRIN_1(cvt_u32_b32, vu32, vb32)static PyObject *simd__intrin_cvt_u32_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_u32_b32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu32 = arg.data.vb32}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3675 | SIMD_IMPL_INTRIN_1(cvt_b32_u32, vb32, vu32)static PyObject *simd__intrin_cvt_b32_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b32_u32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb32 = arg.data.vu32}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 3676 | #if 0 | |||
| 3677 | SIMD_IMPL_INTRIN_1(expand_u32_u32, vu32x2, vu32)static PyObject *simd__intrin_expand_u32_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "expand_u32_u32", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vu32x2 = npyv_expand_u32_u32( arg.data .vu32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32x2 }; return simd_arg_to_obj(&ret ); } | |||
| 3678 | #endif // expand_sup | |||
| 3679 | /*************************** | |||
| 3680 | * Arithmetic | |||
| 3681 | ***************************/ | |||
| 3682 | #line 339 | |||
| 3683 | SIMD_IMPL_INTRIN_2(add_u32, vu32, vu32, vu32)static PyObject *simd__intrin_add_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""add_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_add_epi32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3684 | ||||
| 3685 | #line 339 | |||
| 3686 | SIMD_IMPL_INTRIN_2(sub_u32, vu32, vu32, vu32)static PyObject *simd__intrin_sub_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = _mm_sub_epi32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3687 | ||||
| 3688 | ||||
| 3689 | #if 0 | |||
| 3690 | #line 346 | |||
| 3691 | SIMD_IMPL_INTRIN_2(adds_u32, vu32, vu32, vu32)static PyObject *simd__intrin_adds_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_adds_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3692 | ||||
| 3693 | #line 346 | |||
| 3694 | SIMD_IMPL_INTRIN_2(subs_u32, vu32, vu32, vu32)static PyObject *simd__intrin_subs_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_subs_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3695 | ||||
| 3696 | #endif // sat_sup | |||
| 3697 | ||||
| 3698 | #if 1 | |||
| 3699 | SIMD_IMPL_INTRIN_2(mul_u32, vu32, vu32, vu32)static PyObject *simd__intrin_mul_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_mul_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3700 | #endif // mul_sup | |||
| 3701 | ||||
| 3702 | #if 0 | |||
| 3703 | SIMD_IMPL_INTRIN_2(div_u32, vu32, vu32, vu32)static PyObject *simd__intrin_div_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""div_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_div_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3704 | #endif // div_sup | |||
| 3705 | ||||
| 3706 | #if 1 | |||
| 3707 | SIMD_IMPL_INTRIN_1(divisor_u32, vu32x3, u32)static PyObject *simd__intrin_divisor_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u32}; if (!PyArg_ParseTuple( args, "O&:" "divisor_u32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu32x3 = npyv_divisor_u32( arg.data.u32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu32x3 }; return simd_arg_to_obj(&ret); } | |||
| 3708 | SIMD_IMPL_INTRIN_2(divc_u32, vu32, vu32, vu32x3)static PyObject *simd__intrin_divc_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_divc_u32( arg1.data.vu32, arg2 .data.vu32x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3709 | #endif // intdiv_sup | |||
| 3710 | ||||
| 3711 | #if 0 | |||
| 3712 | #line 367 | |||
| 3713 | SIMD_IMPL_INTRIN_3(muladd_u32, vu32, vu32, vu32, vu32)static PyObject *simd__intrin_muladd_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_muladd_u32 ( arg1.data.vu32, arg2.data.vu32, arg3.data.vu32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3714 | ||||
| 3715 | #line 367 | |||
| 3716 | SIMD_IMPL_INTRIN_3(mulsub_u32, vu32, vu32, vu32, vu32)static PyObject *simd__intrin_mulsub_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_mulsub_u32 ( arg1.data.vu32, arg2.data.vu32, arg3.data.vu32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3717 | ||||
| 3718 | #line 367 | |||
| 3719 | SIMD_IMPL_INTRIN_3(nmuladd_u32, vu32, vu32, vu32, vu32)static PyObject *simd__intrin_nmuladd_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_nmuladd_u32 ( arg1.data.vu32, arg2.data.vu32, arg3.data.vu32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3720 | ||||
| 3721 | #line 367 | |||
| 3722 | SIMD_IMPL_INTRIN_3(nmulsub_u32, vu32, vu32, vu32, vu32)static PyObject *simd__intrin_nmulsub_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu32 = npyv_nmulsub_u32 ( arg1.data.vu32, arg2.data.vu32, arg3.data.vu32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3723 | ||||
| 3724 | #endif // fused_sup | |||
| 3725 | ||||
| 3726 | #if 1 | |||
| 3727 | SIMD_IMPL_INTRIN_1(sum_u32, u32, vu32)static PyObject *simd__intrin_sum_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "sum_u32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.u32 = npyv_sum_u32( arg.data.vu32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_u32 }; return simd_arg_to_obj(&ret); } | |||
| 3728 | #endif // sum_sup | |||
| 3729 | ||||
| 3730 | #if 0 | |||
| 3731 | SIMD_IMPL_INTRIN_1(sumup_u32, u32, vu32)static PyObject *simd__intrin_sumup_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "sumup_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.u32 = npyv_sumup_u32( arg.data.vu32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u32 }; return simd_arg_to_obj(&ret); } | |||
| 3732 | #endif // sumup_sup | |||
| 3733 | ||||
| 3734 | /*************************** | |||
| 3735 | * Math | |||
| 3736 | ***************************/ | |||
| 3737 | #if 0 | |||
| 3738 | #line 386 | |||
| 3739 | SIMD_IMPL_INTRIN_1(sqrt_u32, vu32, vu32)static PyObject *simd__intrin_sqrt_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_u32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu32 = npyv_sqrt_u32( arg.data.vu32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3740 | ||||
| 3741 | #line 386 | |||
| 3742 | SIMD_IMPL_INTRIN_1(recip_u32, vu32, vu32)static PyObject *simd__intrin_recip_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "recip_u32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu32 = npyv_recip_u32( arg.data.vu32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3743 | ||||
| 3744 | #line 386 | |||
| 3745 | SIMD_IMPL_INTRIN_1(abs_u32, vu32, vu32)static PyObject *simd__intrin_abs_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "abs_u32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu32 = npyv_abs_u32( arg.data.vu32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3746 | ||||
| 3747 | #line 386 | |||
| 3748 | SIMD_IMPL_INTRIN_1(square_u32, vu32, vu32)static PyObject *simd__intrin_square_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&:" "square_u32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu32 = npyv_square_u32( arg.data.vu32 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3749 | ||||
| 3750 | #endif | |||
| 3751 | ||||
| 3752 | #line 393 | |||
| 3753 | SIMD_IMPL_INTRIN_2(max_u32, vu32, vu32, vu32)static PyObject *simd__intrin_max_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""max_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_max_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3754 | ||||
| 3755 | #line 393 | |||
| 3756 | SIMD_IMPL_INTRIN_2(min_u32, vu32, vu32, vu32)static PyObject *simd__intrin_min_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""min_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_min_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3757 | ||||
| 3758 | ||||
| 3759 | #if 0 | |||
| 3760 | #line 400 | |||
| 3761 | SIMD_IMPL_INTRIN_2(maxp_u32, vu32, vu32, vu32)static PyObject *simd__intrin_maxp_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_maxp_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3762 | ||||
| 3763 | #line 400 | |||
| 3764 | SIMD_IMPL_INTRIN_2(minp_u32, vu32, vu32, vu32)static PyObject *simd__intrin_minp_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_u32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu32 = npyv_minp_u32( arg1.data.vu32, arg2 .data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret); } | |||
| 3765 | ||||
| 3766 | #endif | |||
| 3767 | ||||
| 3768 | /*************************** | |||
| 3769 | * Mask operations | |||
| 3770 | ***************************/ | |||
| 3771 | #line 410 | |||
| 3772 | SIMD_IMPL_INTRIN_4(ifadd_u32, vu32, vb32, vu32, vu32, vu32)static PyObject *simd__intrin_ifadd_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; simd_arg arg4 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_u32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu32 = npyv_ifadd_u32 ( arg1.data.vb32, arg2.data.vu32, arg3.data.vu32, arg4.data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(& ret); } | |||
| 3773 | ||||
| 3774 | #line 410 | |||
| 3775 | SIMD_IMPL_INTRIN_4(ifsub_u32, vu32, vb32, vu32, vu32, vu32)static PyObject *simd__intrin_ifsub_u32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vu32 }; simd_arg arg3 = {.dtype = simd_data_vu32}; simd_arg arg4 = {.dtype = simd_data_vu32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_u32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu32 = npyv_ifsub_u32 ( arg1.data.vb32, arg2.data.vu32, arg3.data.vu32, arg4.data.vu32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj(& ret); } | |||
| 3776 | ||||
| 3777 | ||||
| 3778 | #endif // simd_sup | |||
| 3779 | ||||
| 3780 | #line 34 | |||
| 3781 | #if 1 | |||
| 3782 | /*************************** | |||
| 3783 | * Memory | |||
| 3784 | ***************************/ | |||
| 3785 | #line 41 | |||
| 3786 | SIMD_IMPL_INTRIN_1(load_s32, vs32, qs32)static PyObject *simd__intrin_load_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs32}; if (!PyArg_ParseTuple( args, "O&:" "load_s32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs32 = npyv_load_s32( arg.data.qs32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3787 | ||||
| 3788 | #line 41 | |||
| 3789 | SIMD_IMPL_INTRIN_1(loada_s32, vs32, qs32)static PyObject *simd__intrin_loada_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs32}; if (!PyArg_ParseTuple( args, "O&:" "loada_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs32 = npyv_loada_s32( arg.data.qs32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3790 | ||||
| 3791 | #line 41 | |||
| 3792 | SIMD_IMPL_INTRIN_1(loads_s32, vs32, qs32)static PyObject *simd__intrin_loads_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs32}; if (!PyArg_ParseTuple( args, "O&:" "loads_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs32 = npyv_loads_s32( arg.data.qs32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3793 | ||||
| 3794 | #line 41 | |||
| 3795 | SIMD_IMPL_INTRIN_1(loadl_s32, vs32, qs32)static PyObject *simd__intrin_loadl_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs32}; if (!PyArg_ParseTuple( args, "O&:" "loadl_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs32 = npyv_loadl_s32( arg.data.qs32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3796 | ||||
| 3797 | #line 46 | |||
| 3798 | // special definition due to the nature of store | |||
| 3799 | static PyObject * | |||
| 3800 | simd__intrin_store_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3801 | { | |||
| 3802 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3803 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3804 | if (!PyArg_ParseTuple( | |||
| 3805 | args, "O&O&:store_s32", | |||
| 3806 | simd_arg_converter, &seq_arg, | |||
| 3807 | simd_arg_converter, &vec_arg | |||
| 3808 | )) { | |||
| 3809 | return NULL((void*)0); | |||
| 3810 | } | |||
| 3811 | npyv_store_s32(seq_arg.data.qs32, vec_arg.data.vs32); | |||
| 3812 | // write-back | |||
| 3813 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3814 | simd_arg_free(&seq_arg); | |||
| 3815 | return NULL((void*)0); | |||
| 3816 | } | |||
| 3817 | simd_arg_free(&seq_arg); | |||
| 3818 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3819 | } | |||
| 3820 | ||||
| 3821 | #line 46 | |||
| 3822 | // special definition due to the nature of storea | |||
| 3823 | static PyObject * | |||
| 3824 | simd__intrin_storea_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3825 | { | |||
| 3826 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3827 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3828 | if (!PyArg_ParseTuple( | |||
| 3829 | args, "O&O&:storea_s32", | |||
| 3830 | simd_arg_converter, &seq_arg, | |||
| 3831 | simd_arg_converter, &vec_arg | |||
| 3832 | )) { | |||
| 3833 | return NULL((void*)0); | |||
| 3834 | } | |||
| 3835 | npyv_storea_s32(seq_arg.data.qs32, vec_arg.data.vs32); | |||
| 3836 | // write-back | |||
| 3837 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3838 | simd_arg_free(&seq_arg); | |||
| 3839 | return NULL((void*)0); | |||
| 3840 | } | |||
| 3841 | simd_arg_free(&seq_arg); | |||
| 3842 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3843 | } | |||
| 3844 | ||||
| 3845 | #line 46 | |||
| 3846 | // special definition due to the nature of stores | |||
| 3847 | static PyObject * | |||
| 3848 | simd__intrin_stores_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3849 | { | |||
| 3850 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3851 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3852 | if (!PyArg_ParseTuple( | |||
| 3853 | args, "O&O&:stores_s32", | |||
| 3854 | simd_arg_converter, &seq_arg, | |||
| 3855 | simd_arg_converter, &vec_arg | |||
| 3856 | )) { | |||
| 3857 | return NULL((void*)0); | |||
| 3858 | } | |||
| 3859 | npyv_stores_s32(seq_arg.data.qs32, vec_arg.data.vs32); | |||
| 3860 | // write-back | |||
| 3861 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3862 | simd_arg_free(&seq_arg); | |||
| 3863 | return NULL((void*)0); | |||
| 3864 | } | |||
| 3865 | simd_arg_free(&seq_arg); | |||
| 3866 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3867 | } | |||
| 3868 | ||||
| 3869 | #line 46 | |||
| 3870 | // special definition due to the nature of storel | |||
| 3871 | static PyObject * | |||
| 3872 | simd__intrin_storel_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3873 | { | |||
| 3874 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3875 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3876 | if (!PyArg_ParseTuple( | |||
| 3877 | args, "O&O&:storel_s32", | |||
| 3878 | simd_arg_converter, &seq_arg, | |||
| 3879 | simd_arg_converter, &vec_arg | |||
| 3880 | )) { | |||
| 3881 | return NULL((void*)0); | |||
| 3882 | } | |||
| 3883 | npyv_storel_s32(seq_arg.data.qs32, vec_arg.data.vs32); | |||
| 3884 | // write-back | |||
| 3885 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3886 | simd_arg_free(&seq_arg); | |||
| 3887 | return NULL((void*)0); | |||
| 3888 | } | |||
| 3889 | simd_arg_free(&seq_arg); | |||
| 3890 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3891 | } | |||
| 3892 | ||||
| 3893 | #line 46 | |||
| 3894 | // special definition due to the nature of storeh | |||
| 3895 | static PyObject * | |||
| 3896 | simd__intrin_storeh_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3897 | { | |||
| 3898 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3899 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3900 | if (!PyArg_ParseTuple( | |||
| 3901 | args, "O&O&:storeh_s32", | |||
| 3902 | simd_arg_converter, &seq_arg, | |||
| 3903 | simd_arg_converter, &vec_arg | |||
| 3904 | )) { | |||
| 3905 | return NULL((void*)0); | |||
| 3906 | } | |||
| 3907 | npyv_storeh_s32(seq_arg.data.qs32, vec_arg.data.vs32); | |||
| 3908 | // write-back | |||
| 3909 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3910 | simd_arg_free(&seq_arg); | |||
| 3911 | return NULL((void*)0); | |||
| 3912 | } | |||
| 3913 | simd_arg_free(&seq_arg); | |||
| 3914 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3915 | } | |||
| 3916 | ||||
| 3917 | ||||
| 3918 | /**************************************** | |||
| 3919 | * Non-contiguous/Partial Memory access | |||
| 3920 | ****************************************/ | |||
| 3921 | #if 1 | |||
| 3922 | // Partial Load | |||
| 3923 | SIMD_IMPL_INTRIN_3(load_till_s32, vs32, qs32, u32, s32)static PyObject *simd__intrin_load_till_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs32}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_s32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_load_till_s32 ( arg1.data.qs32, arg2.data.u32, arg3.data.s32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3924 | SIMD_IMPL_INTRIN_2(load_tillz_s32, vs32, qs32, u32)static PyObject *simd__intrin_load_tillz_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs32}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_s32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs32 = npyv_load_tillz_s32 ( arg1.data.qs32, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 3925 | ||||
| 3926 | // Partial Store | |||
| 3927 | static PyObject * | |||
| 3928 | simd__intrin_store_till_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3929 | { | |||
| 3930 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3931 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3932 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 3933 | if (!PyArg_ParseTuple( | |||
| 3934 | args, "O&O&O&:store_till_s32", | |||
| 3935 | simd_arg_converter, &seq_arg, | |||
| 3936 | simd_arg_converter, &nlane_arg, | |||
| 3937 | simd_arg_converter, &vec_arg | |||
| 3938 | )) { | |||
| 3939 | return NULL((void*)0); | |||
| 3940 | } | |||
| 3941 | npyv_store_till_s32( | |||
| 3942 | seq_arg.data.qs32, nlane_arg.data.u32, vec_arg.data.vs32 | |||
| 3943 | ); | |||
| 3944 | // write-back | |||
| 3945 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 3946 | simd_arg_free(&seq_arg); | |||
| 3947 | return NULL((void*)0); | |||
| 3948 | } | |||
| 3949 | simd_arg_free(&seq_arg); | |||
| 3950 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 3951 | } | |||
| 3952 | ||||
| 3953 | // Non-contiguous Load | |||
| 3954 | #line 112 | |||
| 3955 | static PyObject * | |||
| 3956 | simd__intrin_loadn_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 3957 | { | |||
| 3958 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 3959 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 3960 | #if 0 | |||
| 3961 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 3962 | #endif // till | |||
| 3963 | #if 0 | |||
| 3964 | simd_arg fill_arg = {.dtype = simd_data_s32}; | |||
| 3965 | #endif | |||
| 3966 | if (!PyArg_ParseTuple( | |||
| 3967 | args, "O&O&:loadn_s32", | |||
| 3968 | simd_arg_converter, &seq_arg, | |||
| 3969 | simd_arg_converter, &stride_arg | |||
| 3970 | #if 0 | |||
| 3971 | ,simd_arg_converter, &nlane_arg | |||
| 3972 | #endif | |||
| 3973 | #if 0 | |||
| 3974 | ,simd_arg_converter, &fill_arg | |||
| 3975 | #endif | |||
| 3976 | )) { | |||
| 3977 | return NULL((void*)0); | |||
| 3978 | } | |||
| 3979 | npyv_lanetype_s32 *seq_ptr = seq_arg.data.qs32; | |||
| 3980 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 3981 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 3982 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s324; | |||
| 3983 | if (stride < 0) { | |||
| 3984 | seq_ptr += cur_seq_len -1; | |||
| 3985 | min_seq_len = -min_seq_len; | |||
| 3986 | } | |||
| 3987 | if (cur_seq_len < min_seq_len) { | |||
| 3988 | PyErr_Format(PyExc_ValueError, | |||
| 3989 | "loadn_s32(), according to provided stride %d, the " | |||
| 3990 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 3991 | stride, min_seq_len, cur_seq_len | |||
| 3992 | ); | |||
| 3993 | goto err; | |||
| 3994 | } | |||
| 3995 | npyv_s32 rvec = npyv_loadn_s32( | |||
| 3996 | seq_ptr, stride | |||
| 3997 | #if 0 | |||
| 3998 | , nlane_arg.data.u32 | |||
| 3999 | #endif | |||
| 4000 | #if 0 | |||
| 4001 | , fill_arg.data.s32 | |||
| 4002 | #endif | |||
| 4003 | ); | |||
| 4004 | simd_arg ret = { | |||
| 4005 | .dtype = simd_data_vs32, .data = {.vs32=rvec} | |||
| 4006 | }; | |||
| 4007 | simd_arg_free(&seq_arg); | |||
| 4008 | return simd_arg_to_obj(&ret); | |||
| 4009 | err: | |||
| 4010 | simd_arg_free(&seq_arg); | |||
| 4011 | return NULL((void*)0); | |||
| 4012 | } | |||
| 4013 | ||||
| 4014 | #line 112 | |||
| 4015 | static PyObject * | |||
| 4016 | simd__intrin_loadn_till_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4017 | { | |||
| 4018 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 4019 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4020 | #if 1 | |||
| 4021 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4022 | #endif // till | |||
| 4023 | #if 1 | |||
| 4024 | simd_arg fill_arg = {.dtype = simd_data_s32}; | |||
| 4025 | #endif | |||
| 4026 | if (!PyArg_ParseTuple( | |||
| 4027 | args, "O&O&O&O&:loadn_till_s32", | |||
| 4028 | simd_arg_converter, &seq_arg, | |||
| 4029 | simd_arg_converter, &stride_arg | |||
| 4030 | #if 1 | |||
| 4031 | ,simd_arg_converter, &nlane_arg | |||
| 4032 | #endif | |||
| 4033 | #if 1 | |||
| 4034 | ,simd_arg_converter, &fill_arg | |||
| 4035 | #endif | |||
| 4036 | )) { | |||
| 4037 | return NULL((void*)0); | |||
| 4038 | } | |||
| 4039 | npyv_lanetype_s32 *seq_ptr = seq_arg.data.qs32; | |||
| 4040 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4041 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4042 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s324; | |||
| 4043 | if (stride < 0) { | |||
| 4044 | seq_ptr += cur_seq_len -1; | |||
| 4045 | min_seq_len = -min_seq_len; | |||
| 4046 | } | |||
| 4047 | if (cur_seq_len < min_seq_len) { | |||
| 4048 | PyErr_Format(PyExc_ValueError, | |||
| 4049 | "loadn_till_s32(), according to provided stride %d, the " | |||
| 4050 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4051 | stride, min_seq_len, cur_seq_len | |||
| 4052 | ); | |||
| 4053 | goto err; | |||
| 4054 | } | |||
| 4055 | npyv_s32 rvec = npyv_loadn_till_s32( | |||
| 4056 | seq_ptr, stride | |||
| 4057 | #if 1 | |||
| 4058 | , nlane_arg.data.u32 | |||
| 4059 | #endif | |||
| 4060 | #if 1 | |||
| 4061 | , fill_arg.data.s32 | |||
| 4062 | #endif | |||
| 4063 | ); | |||
| 4064 | simd_arg ret = { | |||
| 4065 | .dtype = simd_data_vs32, .data = {.vs32=rvec} | |||
| 4066 | }; | |||
| 4067 | simd_arg_free(&seq_arg); | |||
| 4068 | return simd_arg_to_obj(&ret); | |||
| 4069 | err: | |||
| 4070 | simd_arg_free(&seq_arg); | |||
| 4071 | return NULL((void*)0); | |||
| 4072 | } | |||
| 4073 | ||||
| 4074 | #line 112 | |||
| 4075 | static PyObject * | |||
| 4076 | simd__intrin_loadn_tillz_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4077 | { | |||
| 4078 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 4079 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4080 | #if 1 | |||
| 4081 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4082 | #endif // till | |||
| 4083 | #if 0 | |||
| 4084 | simd_arg fill_arg = {.dtype = simd_data_s32}; | |||
| 4085 | #endif | |||
| 4086 | if (!PyArg_ParseTuple( | |||
| 4087 | args, "O&O&O&:loadn_tillz_s32", | |||
| 4088 | simd_arg_converter, &seq_arg, | |||
| 4089 | simd_arg_converter, &stride_arg | |||
| 4090 | #if 1 | |||
| 4091 | ,simd_arg_converter, &nlane_arg | |||
| 4092 | #endif | |||
| 4093 | #if 0 | |||
| 4094 | ,simd_arg_converter, &fill_arg | |||
| 4095 | #endif | |||
| 4096 | )) { | |||
| 4097 | return NULL((void*)0); | |||
| 4098 | } | |||
| 4099 | npyv_lanetype_s32 *seq_ptr = seq_arg.data.qs32; | |||
| 4100 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4101 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4102 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s324; | |||
| 4103 | if (stride < 0) { | |||
| 4104 | seq_ptr += cur_seq_len -1; | |||
| 4105 | min_seq_len = -min_seq_len; | |||
| 4106 | } | |||
| 4107 | if (cur_seq_len < min_seq_len) { | |||
| 4108 | PyErr_Format(PyExc_ValueError, | |||
| 4109 | "loadn_tillz_s32(), according to provided stride %d, the " | |||
| 4110 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4111 | stride, min_seq_len, cur_seq_len | |||
| 4112 | ); | |||
| 4113 | goto err; | |||
| 4114 | } | |||
| 4115 | npyv_s32 rvec = npyv_loadn_tillz_s32( | |||
| 4116 | seq_ptr, stride | |||
| 4117 | #if 1 | |||
| 4118 | , nlane_arg.data.u32 | |||
| 4119 | #endif | |||
| 4120 | #if 0 | |||
| 4121 | , fill_arg.data.s32 | |||
| 4122 | #endif | |||
| 4123 | ); | |||
| 4124 | simd_arg ret = { | |||
| 4125 | .dtype = simd_data_vs32, .data = {.vs32=rvec} | |||
| 4126 | }; | |||
| 4127 | simd_arg_free(&seq_arg); | |||
| 4128 | return simd_arg_to_obj(&ret); | |||
| 4129 | err: | |||
| 4130 | simd_arg_free(&seq_arg); | |||
| 4131 | return NULL((void*)0); | |||
| 4132 | } | |||
| 4133 | ||||
| 4134 | ||||
| 4135 | // Non-contiguous Store | |||
| 4136 | #line 178 | |||
| 4137 | static PyObject * | |||
| 4138 | simd__intrin_storen_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4139 | { | |||
| 4140 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 4141 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4142 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 4143 | #if 0 | |||
| 4144 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4145 | #endif | |||
| 4146 | if (!PyArg_ParseTuple( | |||
| 4147 | args, "O&O&O&:storen_s32", | |||
| 4148 | simd_arg_converter, &seq_arg, | |||
| 4149 | simd_arg_converter, &stride_arg | |||
| 4150 | #if 0 | |||
| 4151 | ,simd_arg_converter, &nlane_arg | |||
| 4152 | #endif | |||
| 4153 | ,simd_arg_converter, &vec_arg | |||
| 4154 | )) { | |||
| 4155 | return NULL((void*)0); | |||
| 4156 | } | |||
| 4157 | npyv_lanetype_s32 *seq_ptr = seq_arg.data.qs32; | |||
| 4158 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4159 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4160 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s324; | |||
| 4161 | if (stride < 0) { | |||
| 4162 | seq_ptr += cur_seq_len -1; | |||
| 4163 | min_seq_len = -min_seq_len; | |||
| 4164 | } | |||
| 4165 | // overflow guard | |||
| 4166 | if (cur_seq_len < min_seq_len) { | |||
| 4167 | PyErr_Format(PyExc_ValueError, | |||
| 4168 | "storen_s32(), according to provided stride %d, the" | |||
| 4169 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4170 | stride, min_seq_len, cur_seq_len | |||
| 4171 | ); | |||
| 4172 | goto err; | |||
| 4173 | } | |||
| 4174 | npyv_storen_s32( | |||
| 4175 | seq_ptr, stride | |||
| 4176 | #if 0 | |||
| 4177 | ,nlane_arg.data.u32 | |||
| 4178 | #endif | |||
| 4179 | ,vec_arg.data.vs32 | |||
| 4180 | ); | |||
| 4181 | // write-back | |||
| 4182 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 4183 | goto err; | |||
| 4184 | } | |||
| 4185 | simd_arg_free(&seq_arg); | |||
| 4186 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4187 | err: | |||
| 4188 | simd_arg_free(&seq_arg); | |||
| 4189 | return NULL((void*)0); | |||
| 4190 | } | |||
| 4191 | ||||
| 4192 | #line 178 | |||
| 4193 | static PyObject * | |||
| 4194 | simd__intrin_storen_till_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4195 | { | |||
| 4196 | simd_arg seq_arg = {.dtype = simd_data_qs32}; | |||
| 4197 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4198 | simd_arg vec_arg = {.dtype = simd_data_vs32}; | |||
| 4199 | #if 1 | |||
| 4200 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4201 | #endif | |||
| 4202 | if (!PyArg_ParseTuple( | |||
| 4203 | args, "O&O&O&O&:storen_s32", | |||
| 4204 | simd_arg_converter, &seq_arg, | |||
| 4205 | simd_arg_converter, &stride_arg | |||
| 4206 | #if 1 | |||
| 4207 | ,simd_arg_converter, &nlane_arg | |||
| 4208 | #endif | |||
| 4209 | ,simd_arg_converter, &vec_arg | |||
| 4210 | )) { | |||
| 4211 | return NULL((void*)0); | |||
| 4212 | } | |||
| 4213 | npyv_lanetype_s32 *seq_ptr = seq_arg.data.qs32; | |||
| 4214 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4215 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4216 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s324; | |||
| 4217 | if (stride < 0) { | |||
| 4218 | seq_ptr += cur_seq_len -1; | |||
| 4219 | min_seq_len = -min_seq_len; | |||
| 4220 | } | |||
| 4221 | // overflow guard | |||
| 4222 | if (cur_seq_len < min_seq_len) { | |||
| 4223 | PyErr_Format(PyExc_ValueError, | |||
| 4224 | "storen_till_s32(), according to provided stride %d, the" | |||
| 4225 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4226 | stride, min_seq_len, cur_seq_len | |||
| 4227 | ); | |||
| 4228 | goto err; | |||
| 4229 | } | |||
| 4230 | npyv_storen_till_s32( | |||
| 4231 | seq_ptr, stride | |||
| 4232 | #if 1 | |||
| 4233 | ,nlane_arg.data.u32 | |||
| 4234 | #endif | |||
| 4235 | ,vec_arg.data.vs32 | |||
| 4236 | ); | |||
| 4237 | // write-back | |||
| 4238 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs32, simd_data_qs32)) { | |||
| 4239 | goto err; | |||
| 4240 | } | |||
| 4241 | simd_arg_free(&seq_arg); | |||
| 4242 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4243 | err: | |||
| 4244 | simd_arg_free(&seq_arg); | |||
| 4245 | return NULL((void*)0); | |||
| 4246 | } | |||
| 4247 | ||||
| 4248 | #endif // 1 | |||
| 4249 | ||||
| 4250 | /*************************** | |||
| 4251 | * Misc | |||
| 4252 | ***************************/ | |||
| 4253 | SIMD_IMPL_INTRIN_0(zero_s32, vs32)static PyObject *simd__intrin_zero_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_s32") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vs32, .data = {.vs32 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 4254 | SIMD_IMPL_INTRIN_1(setall_s32, vs32, s32)static PyObject *simd__intrin_setall_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s32}; if (!PyArg_ParseTuple( args, "O&:" "setall_s32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs32 = _mm_set1_epi32((int)(arg.data.s32 ))}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4255 | SIMD_IMPL_INTRIN_3(select_s32, vs32, vb32, vs32, vs32)static PyObject *simd__intrin_select_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_select_u8 ( arg1.data.vb32, arg2.data.vs32, arg3.data.vs32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4256 | ||||
| 4257 | #line 246 | |||
| 4258 | #if 1 | |||
| 4259 | SIMD_IMPL_INTRIN_1(reinterpret_u8_s32, vu8, vs32)static PyObject *simd__intrin_reinterpret_u8_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vs32 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 4260 | #endif // simd_sup2 | |||
| 4261 | ||||
| 4262 | #line 246 | |||
| 4263 | #if 1 | |||
| 4264 | SIMD_IMPL_INTRIN_1(reinterpret_s8_s32, vs8, vs32)static PyObject *simd__intrin_reinterpret_s8_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vs32 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 4265 | #endif // simd_sup2 | |||
| 4266 | ||||
| 4267 | #line 246 | |||
| 4268 | #if 1 | |||
| 4269 | SIMD_IMPL_INTRIN_1(reinterpret_u16_s32, vu16, vs32)static PyObject *simd__intrin_reinterpret_u16_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 4270 | #endif // simd_sup2 | |||
| 4271 | ||||
| 4272 | #line 246 | |||
| 4273 | #if 1 | |||
| 4274 | SIMD_IMPL_INTRIN_1(reinterpret_s16_s32, vs16, vs32)static PyObject *simd__intrin_reinterpret_s16_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 4275 | #endif // simd_sup2 | |||
| 4276 | ||||
| 4277 | #line 246 | |||
| 4278 | #if 1 | |||
| 4279 | SIMD_IMPL_INTRIN_1(reinterpret_u32_s32, vu32, vs32)static PyObject *simd__intrin_reinterpret_u32_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 4280 | #endif // simd_sup2 | |||
| 4281 | ||||
| 4282 | #line 246 | |||
| 4283 | #if 1 | |||
| 4284 | SIMD_IMPL_INTRIN_1(reinterpret_s32_s32, vs32, vs32)static PyObject *simd__intrin_reinterpret_s32_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 4285 | #endif // simd_sup2 | |||
| 4286 | ||||
| 4287 | #line 246 | |||
| 4288 | #if 1 | |||
| 4289 | SIMD_IMPL_INTRIN_1(reinterpret_u64_s32, vu64, vs32)static PyObject *simd__intrin_reinterpret_u64_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 4290 | #endif // simd_sup2 | |||
| 4291 | ||||
| 4292 | #line 246 | |||
| 4293 | #if 1 | |||
| 4294 | SIMD_IMPL_INTRIN_1(reinterpret_s64_s32, vs64, vs32)static PyObject *simd__intrin_reinterpret_s64_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vs32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 4295 | #endif // simd_sup2 | |||
| 4296 | ||||
| 4297 | #line 246 | |||
| 4298 | #if 1 | |||
| 4299 | SIMD_IMPL_INTRIN_1(reinterpret_f32_s32, vf32, vs32)static PyObject *simd__intrin_reinterpret_f32_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vs32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 4300 | #endif // simd_sup2 | |||
| 4301 | ||||
| 4302 | #line 246 | |||
| 4303 | #if NPY_SIMD_F641 | |||
| 4304 | SIMD_IMPL_INTRIN_1(reinterpret_f64_s32, vf64, vs32)static PyObject *simd__intrin_reinterpret_f64_s32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_s32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vs32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 4305 | #endif // simd_sup2 | |||
| 4306 | ||||
| 4307 | ||||
| 4308 | /** | |||
| 4309 | * special definition due to the nature of intrinsics | |||
| 4310 | * npyv_setf_s32 and npy_set_s32. | |||
| 4311 | */ | |||
| 4312 | #line 258 | |||
| 4313 | static PyObject * | |||
| 4314 | simd__intrin_setf_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4315 | { | |||
| 4316 | npyv_lanetype_s32 *data = simd_sequence_from_iterable(args, simd_data_qs32, npyv_nlanes_s324); | |||
| 4317 | if (data == NULL((void*)0)) { | |||
| 4318 | return NULL((void*)0); | |||
| 4319 | } | |||
| 4320 | simd_data r = {.vs32 = npyv_setf_s32(npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4321 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4322 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4323 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4324 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4325 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4326 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4327 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4328 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4329 | data[64] // for setfnpyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4])) | |||
| 4330 | )npyv__setr_epi32((int)(data[1]), (int)(data[2]), (int)(data[3 ]), (int)(data[4]))}; | |||
| 4331 | simd_sequence_free(data); | |||
| 4332 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs32); | |||
| 4333 | } | |||
| 4334 | ||||
| 4335 | #line 258 | |||
| 4336 | static PyObject * | |||
| 4337 | simd__intrin_set_s32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4338 | { | |||
| 4339 | npyv_lanetype_s32 *data = simd_sequence_from_iterable(args, simd_data_qs32, npyv_nlanes_s324); | |||
| 4340 | if (data == NULL((void*)0)) { | |||
| 4341 | return NULL((void*)0); | |||
| 4342 | } | |||
| 4343 | simd_data r = {.vs32 = npyv_set_s32(npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4344 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4345 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4346 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4347 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4348 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4349 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4350 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4351 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4352 | data[64] // for setfnpyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3])) | |||
| 4353 | )npyv__setr_epi32((int)(data[0]), (int)(data[1]), (int)(data[2 ]), (int)(data[3]))}; | |||
| 4354 | simd_sequence_free(data); | |||
| 4355 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs32); | |||
| 4356 | } | |||
| 4357 | ||||
| 4358 | ||||
| 4359 | /*************************** | |||
| 4360 | * Reorder | |||
| 4361 | ***************************/ | |||
| 4362 | #line 287 | |||
| 4363 | SIMD_IMPL_INTRIN_2(combinel_s32, vs32, vs32, vs32)static PyObject *simd__intrin_combinel_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_s32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs32 = _mm_unpacklo_epi64 ( arg1.data.vs32, arg2.data.vs32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4364 | ||||
| 4365 | #line 287 | |||
| 4366 | SIMD_IMPL_INTRIN_2(combineh_s32, vs32, vs32, vs32)static PyObject *simd__intrin_combineh_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_s32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs32 = _mm_unpackhi_epi64 ( arg1.data.vs32, arg2.data.vs32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4367 | ||||
| 4368 | ||||
| 4369 | #line 293 | |||
| 4370 | SIMD_IMPL_INTRIN_2(combine_s32, vs32x2, vs32, vs32)static PyObject *simd__intrin_combine_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_s32", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs32x2 = npyv__combine ( arg1.data.vs32, arg2.data.vs32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32x2 }; return simd_arg_to_obj(&ret); } | |||
| 4371 | ||||
| 4372 | #line 293 | |||
| 4373 | SIMD_IMPL_INTRIN_2(zip_s32, vs32x2, vs32, vs32)static PyObject *simd__intrin_zip_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32x2 = npyv_zip_s32( arg1.data.vs32, arg2.data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32x2 }; return simd_arg_to_obj(&ret); } | |||
| 4374 | ||||
| 4375 | ||||
| 4376 | #if 1 | |||
| 4377 | SIMD_IMPL_INTRIN_1(rev64_s32, vs32, vs32)static PyObject *simd__intrin_rev64_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "rev64_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs32 = npyv_rev64_u32( arg.data.vs32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4378 | #endif | |||
| 4379 | ||||
| 4380 | /*************************** | |||
| 4381 | * Operators | |||
| 4382 | ***************************/ | |||
| 4383 | #if 31 > 0 | |||
| 4384 | SIMD_IMPL_INTRIN_2(shl_s32, vs32, vs32, u8)static PyObject *simd__intrin_shl_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_sll_epi32(arg1.data.vs32, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4385 | SIMD_IMPL_INTRIN_2(shr_s32, vs32, vs32, u8)static PyObject *simd__intrin_shr_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_sra_epi32(arg1.data.vs32, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4386 | // immediate constant | |||
| 4387 | SIMD_IMPL_INTRIN_2IMM(shli_s32, vs32, vs32, 31)static PyObject *simd__intrin_shli_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs32 = 0 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 0) : 1 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vs32, 1) : 2 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vs32, 2) : 3 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vs32, 3) : 4 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 5) : 6 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 6) : 7 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi32(arg1.data.vs32, 8) : 9 == arg2.data .u8 ? _mm_slli_epi32(arg1.data.vs32, 9) : 10 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 10) : 11 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vs32, 11) : 12 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vs32, 12) : 13 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vs32, 13) : 14 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 15 ) : 16 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 16) : 17 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 17) : 18 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 18) : 19 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 19) : 20 == arg2 .data.u8 ? _mm_slli_epi32(arg1.data.vs32, 20) : 21 == arg2.data .u8 ? _mm_slli_epi32(arg1.data.vs32, 21) : 22 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 22) : 23 == arg2.data.u8 ? _mm_slli_epi32 (arg1.data.vs32, 23) : 24 == arg2.data.u8 ? _mm_slli_epi32(arg1 .data.vs32, 24) : 25 == arg2.data.u8 ? _mm_slli_epi32(arg1.data .vs32, 25) : 26 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32 , 26) : 27 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 27 ) : 28 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 28) : 29 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 29) : 30 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 30) : 31 == arg2.data.u8 ? _mm_slli_epi32(arg1.data.vs32, 31) : data.vs32 ; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4388 | SIMD_IMPL_INTRIN_2IMM(shri_s32, vs32, vs32, 32)static PyObject *simd__intrin_shri_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs32 = 1 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 1) : 2 == arg2.data.u8 ? _mm_srai_epi32 (arg1.data.vs32, 2) : 3 == arg2.data.u8 ? _mm_srai_epi32(arg1 .data.vs32, 3) : 4 == arg2.data.u8 ? _mm_srai_epi32(arg1.data .vs32, 4) : 5 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32 , 5) : 6 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 6) : 7 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 7) : 8 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 8) : 9 == arg2 .data.u8 ? _mm_srai_epi32(arg1.data.vs32, 9) : 10 == arg2.data .u8 ? _mm_srai_epi32(arg1.data.vs32, 10) : 11 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 11) : 12 == arg2.data.u8 ? _mm_srai_epi32 (arg1.data.vs32, 12) : 13 == arg2.data.u8 ? _mm_srai_epi32(arg1 .data.vs32, 13) : 14 == arg2.data.u8 ? _mm_srai_epi32(arg1.data .vs32, 14) : 15 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32 , 15) : 16 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 16 ) : 17 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 17) : 18 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 18) : 19 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 19) : 20 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 20) : 21 == arg2 .data.u8 ? _mm_srai_epi32(arg1.data.vs32, 21) : 22 == arg2.data .u8 ? _mm_srai_epi32(arg1.data.vs32, 22) : 23 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 23) : 24 == arg2.data.u8 ? _mm_srai_epi32 (arg1.data.vs32, 24) : 25 == arg2.data.u8 ? _mm_srai_epi32(arg1 .data.vs32, 25) : 26 == arg2.data.u8 ? _mm_srai_epi32(arg1.data .vs32, 26) : 27 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32 , 27) : 28 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 28 ) : 29 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 29) : 30 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 30) : 31 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 31) : 32 == arg2.data.u8 ? _mm_srai_epi32(arg1.data.vs32, 32) : data.vs32 ; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4389 | #endif // shl_imm | |||
| 4390 | ||||
| 4391 | #line 314 | |||
| 4392 | SIMD_IMPL_INTRIN_2(and_s32, vs32, vs32, vs32)static PyObject *simd__intrin_and_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""and_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_and_si128( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4393 | ||||
| 4394 | #line 314 | |||
| 4395 | SIMD_IMPL_INTRIN_2(or_s32, vs32, vs32, vs32)static PyObject *simd__intrin_or_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""or_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_or_si128( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4396 | ||||
| 4397 | #line 314 | |||
| 4398 | SIMD_IMPL_INTRIN_2(xor_s32, vs32, vs32, vs32)static PyObject *simd__intrin_xor_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_xor_si128( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4399 | ||||
| 4400 | ||||
| 4401 | SIMD_IMPL_INTRIN_1(not_s32, vs32, vs32)static PyObject *simd__intrin_not_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "not_s32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs32 = _mm_xor_si128(arg.data.vs32, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 4402 | ||||
| 4403 | #line 322 | |||
| 4404 | SIMD_IMPL_INTRIN_2(cmpeq_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmpeq_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_cmpeq_epi32( arg1.data.vs32 , arg2.data.vs32 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 4405 | ||||
| 4406 | #line 322 | |||
| 4407 | SIMD_IMPL_INTRIN_2(cmpneq_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmpneq_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(_mm_cmpeq_epi32(arg1 .data.vs32, arg2.data.vs32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 4408 | ||||
| 4409 | #line 322 | |||
| 4410 | SIMD_IMPL_INTRIN_2(cmpgt_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmpgt_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_cmpgt_epi32( arg1.data.vs32 , arg2.data.vs32 )}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 4411 | ||||
| 4412 | #line 322 | |||
| 4413 | SIMD_IMPL_INTRIN_2(cmpge_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmpge_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(_mm_cmpgt_epi32(arg2 .data.vs32, arg1.data.vs32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 4414 | ||||
| 4415 | #line 322 | |||
| 4416 | SIMD_IMPL_INTRIN_2(cmplt_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmplt_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_cmpgt_epi32(arg2.data.vs32, arg1.data.vs32)}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 4417 | ||||
| 4418 | #line 322 | |||
| 4419 | SIMD_IMPL_INTRIN_2(cmple_s32, vb32, vs32, vs32)static PyObject *simd__intrin_cmple_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128(_mm_cmpgt_epi32(arg1 .data.vs32, arg2.data.vs32), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(& ret); } | |||
| 4420 | ||||
| 4421 | ||||
| 4422 | /*************************** | |||
| 4423 | * Conversion | |||
| 4424 | ***************************/ | |||
| 4425 | SIMD_IMPL_INTRIN_1(cvt_s32_b32, vs32, vb32)static PyObject *simd__intrin_cvt_s32_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_s32_b32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs32 = arg.data.vb32}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4426 | SIMD_IMPL_INTRIN_1(cvt_b32_s32, vb32, vs32)static PyObject *simd__intrin_cvt_b32_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b32_s32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb32 = arg.data.vs32}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 4427 | #if 0 | |||
| 4428 | SIMD_IMPL_INTRIN_1(expand_s32_s32, vs32x2, vs32)static PyObject *simd__intrin_expand_s32_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "expand_s32_s32", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vs32x2 = npyv_expand_s32_s32( arg.data .vs32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32x2 }; return simd_arg_to_obj(&ret ); } | |||
| 4429 | #endif // expand_sup | |||
| 4430 | /*************************** | |||
| 4431 | * Arithmetic | |||
| 4432 | ***************************/ | |||
| 4433 | #line 339 | |||
| 4434 | SIMD_IMPL_INTRIN_2(add_s32, vs32, vs32, vs32)static PyObject *simd__intrin_add_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""add_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_add_epi32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4435 | ||||
| 4436 | #line 339 | |||
| 4437 | SIMD_IMPL_INTRIN_2(sub_s32, vs32, vs32, vs32)static PyObject *simd__intrin_sub_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = _mm_sub_epi32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4438 | ||||
| 4439 | ||||
| 4440 | #if 0 | |||
| 4441 | #line 346 | |||
| 4442 | SIMD_IMPL_INTRIN_2(adds_s32, vs32, vs32, vs32)static PyObject *simd__intrin_adds_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_adds_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4443 | ||||
| 4444 | #line 346 | |||
| 4445 | SIMD_IMPL_INTRIN_2(subs_s32, vs32, vs32, vs32)static PyObject *simd__intrin_subs_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_subs_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4446 | ||||
| 4447 | #endif // sat_sup | |||
| 4448 | ||||
| 4449 | #if 1 | |||
| 4450 | SIMD_IMPL_INTRIN_2(mul_s32, vs32, vs32, vs32)static PyObject *simd__intrin_mul_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_mul_u32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4451 | #endif // mul_sup | |||
| 4452 | ||||
| 4453 | #if 0 | |||
| 4454 | SIMD_IMPL_INTRIN_2(div_s32, vs32, vs32, vs32)static PyObject *simd__intrin_div_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""div_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_div_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4455 | #endif // div_sup | |||
| 4456 | ||||
| 4457 | #if 1 | |||
| 4458 | SIMD_IMPL_INTRIN_1(divisor_s32, vs32x3, s32)static PyObject *simd__intrin_divisor_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s32}; if (!PyArg_ParseTuple( args, "O&:" "divisor_s32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs32x3 = npyv_divisor_s32( arg.data.s32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs32x3 }; return simd_arg_to_obj(&ret); } | |||
| 4459 | SIMD_IMPL_INTRIN_2(divc_s32, vs32, vs32, vs32x3)static PyObject *simd__intrin_divc_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_divc_s32( arg1.data.vs32, arg2 .data.vs32x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4460 | #endif // intdiv_sup | |||
| 4461 | ||||
| 4462 | #if 0 | |||
| 4463 | #line 367 | |||
| 4464 | SIMD_IMPL_INTRIN_3(muladd_s32, vs32, vs32, vs32, vs32)static PyObject *simd__intrin_muladd_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_muladd_s32 ( arg1.data.vs32, arg2.data.vs32, arg3.data.vs32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4465 | ||||
| 4466 | #line 367 | |||
| 4467 | SIMD_IMPL_INTRIN_3(mulsub_s32, vs32, vs32, vs32, vs32)static PyObject *simd__intrin_mulsub_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_mulsub_s32 ( arg1.data.vs32, arg2.data.vs32, arg3.data.vs32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4468 | ||||
| 4469 | #line 367 | |||
| 4470 | SIMD_IMPL_INTRIN_3(nmuladd_s32, vs32, vs32, vs32, vs32)static PyObject *simd__intrin_nmuladd_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_nmuladd_s32 ( arg1.data.vs32, arg2.data.vs32, arg3.data.vs32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4471 | ||||
| 4472 | #line 367 | |||
| 4473 | SIMD_IMPL_INTRIN_3(nmulsub_s32, vs32, vs32, vs32, vs32)static PyObject *simd__intrin_nmulsub_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs32 = npyv_nmulsub_s32 ( arg1.data.vs32, arg2.data.vs32, arg3.data.vs32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4474 | ||||
| 4475 | #endif // fused_sup | |||
| 4476 | ||||
| 4477 | #if 0 | |||
| 4478 | SIMD_IMPL_INTRIN_1(sum_s32, s32, vs32)static PyObject *simd__intrin_sum_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "sum_s32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.s32 = npyv_sum_s32( arg.data.vs32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_s32 }; return simd_arg_to_obj(&ret); } | |||
| 4479 | #endif // sum_sup | |||
| 4480 | ||||
| 4481 | #if 0 | |||
| 4482 | SIMD_IMPL_INTRIN_1(sumup_s32, s32, vs32)static PyObject *simd__intrin_sumup_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "sumup_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.s32 = npyv_sumup_s32( arg.data.vs32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_s32 }; return simd_arg_to_obj(&ret); } | |||
| 4483 | #endif // sumup_sup | |||
| 4484 | ||||
| 4485 | /*************************** | |||
| 4486 | * Math | |||
| 4487 | ***************************/ | |||
| 4488 | #if 0 | |||
| 4489 | #line 386 | |||
| 4490 | SIMD_IMPL_INTRIN_1(sqrt_s32, vs32, vs32)static PyObject *simd__intrin_sqrt_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_s32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs32 = npyv_sqrt_s32( arg.data.vs32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4491 | ||||
| 4492 | #line 386 | |||
| 4493 | SIMD_IMPL_INTRIN_1(recip_s32, vs32, vs32)static PyObject *simd__intrin_recip_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "recip_s32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs32 = npyv_recip_s32( arg.data.vs32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4494 | ||||
| 4495 | #line 386 | |||
| 4496 | SIMD_IMPL_INTRIN_1(abs_s32, vs32, vs32)static PyObject *simd__intrin_abs_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "abs_s32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs32 = npyv_abs_s32( arg.data.vs32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4497 | ||||
| 4498 | #line 386 | |||
| 4499 | SIMD_IMPL_INTRIN_1(square_s32, vs32, vs32)static PyObject *simd__intrin_square_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&:" "square_s32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs32 = npyv_square_s32( arg.data.vs32 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4500 | ||||
| 4501 | #endif | |||
| 4502 | ||||
| 4503 | #line 393 | |||
| 4504 | SIMD_IMPL_INTRIN_2(max_s32, vs32, vs32, vs32)static PyObject *simd__intrin_max_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""max_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_max_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4505 | ||||
| 4506 | #line 393 | |||
| 4507 | SIMD_IMPL_INTRIN_2(min_s32, vs32, vs32, vs32)static PyObject *simd__intrin_min_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""min_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_min_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4508 | ||||
| 4509 | ||||
| 4510 | #if 0 | |||
| 4511 | #line 400 | |||
| 4512 | SIMD_IMPL_INTRIN_2(maxp_s32, vs32, vs32, vs32)static PyObject *simd__intrin_maxp_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_maxp_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4513 | ||||
| 4514 | #line 400 | |||
| 4515 | SIMD_IMPL_INTRIN_2(minp_s32, vs32, vs32, vs32)static PyObject *simd__intrin_minp_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_s32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs32 = npyv_minp_s32( arg1.data.vs32, arg2 .data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 4516 | ||||
| 4517 | #endif | |||
| 4518 | ||||
| 4519 | /*************************** | |||
| 4520 | * Mask operations | |||
| 4521 | ***************************/ | |||
| 4522 | #line 410 | |||
| 4523 | SIMD_IMPL_INTRIN_4(ifadd_s32, vs32, vb32, vs32, vs32, vs32)static PyObject *simd__intrin_ifadd_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; simd_arg arg4 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_s32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs32 = npyv_ifadd_s32 ( arg1.data.vb32, arg2.data.vs32, arg3.data.vs32, arg4.data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(& ret); } | |||
| 4524 | ||||
| 4525 | #line 410 | |||
| 4526 | SIMD_IMPL_INTRIN_4(ifsub_s32, vs32, vb32, vs32, vs32, vs32)static PyObject *simd__intrin_ifsub_s32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vs32 }; simd_arg arg3 = {.dtype = simd_data_vs32}; simd_arg arg4 = {.dtype = simd_data_vs32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_s32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs32 = npyv_ifsub_s32 ( arg1.data.vb32, arg2.data.vs32, arg3.data.vs32, arg4.data.vs32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(& ret); } | |||
| 4527 | ||||
| 4528 | ||||
| 4529 | #endif // simd_sup | |||
| 4530 | ||||
| 4531 | #line 34 | |||
| 4532 | #if 1 | |||
| 4533 | /*************************** | |||
| 4534 | * Memory | |||
| 4535 | ***************************/ | |||
| 4536 | #line 41 | |||
| 4537 | SIMD_IMPL_INTRIN_1(load_u64, vu64, qu64)static PyObject *simd__intrin_load_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu64}; if (!PyArg_ParseTuple( args, "O&:" "load_u64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu64 = npyv_load_u64( arg.data.qu64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4538 | ||||
| 4539 | #line 41 | |||
| 4540 | SIMD_IMPL_INTRIN_1(loada_u64, vu64, qu64)static PyObject *simd__intrin_loada_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu64}; if (!PyArg_ParseTuple( args, "O&:" "loada_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu64 = npyv_loada_u64( arg.data.qu64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4541 | ||||
| 4542 | #line 41 | |||
| 4543 | SIMD_IMPL_INTRIN_1(loads_u64, vu64, qu64)static PyObject *simd__intrin_loads_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu64}; if (!PyArg_ParseTuple( args, "O&:" "loads_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu64 = npyv_loads_u64( arg.data.qu64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4544 | ||||
| 4545 | #line 41 | |||
| 4546 | SIMD_IMPL_INTRIN_1(loadl_u64, vu64, qu64)static PyObject *simd__intrin_loadl_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qu64}; if (!PyArg_ParseTuple( args, "O&:" "loadl_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu64 = npyv_loadl_u64( arg.data.qu64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4547 | ||||
| 4548 | #line 46 | |||
| 4549 | // special definition due to the nature of store | |||
| 4550 | static PyObject * | |||
| 4551 | simd__intrin_store_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4552 | { | |||
| 4553 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4554 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4555 | if (!PyArg_ParseTuple( | |||
| 4556 | args, "O&O&:store_u64", | |||
| 4557 | simd_arg_converter, &seq_arg, | |||
| 4558 | simd_arg_converter, &vec_arg | |||
| 4559 | )) { | |||
| 4560 | return NULL((void*)0); | |||
| 4561 | } | |||
| 4562 | npyv_store_u64(seq_arg.data.qu64, vec_arg.data.vu64); | |||
| 4563 | // write-back | |||
| 4564 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4565 | simd_arg_free(&seq_arg); | |||
| 4566 | return NULL((void*)0); | |||
| 4567 | } | |||
| 4568 | simd_arg_free(&seq_arg); | |||
| 4569 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4570 | } | |||
| 4571 | ||||
| 4572 | #line 46 | |||
| 4573 | // special definition due to the nature of storea | |||
| 4574 | static PyObject * | |||
| 4575 | simd__intrin_storea_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4576 | { | |||
| 4577 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4578 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4579 | if (!PyArg_ParseTuple( | |||
| 4580 | args, "O&O&:storea_u64", | |||
| 4581 | simd_arg_converter, &seq_arg, | |||
| 4582 | simd_arg_converter, &vec_arg | |||
| 4583 | )) { | |||
| 4584 | return NULL((void*)0); | |||
| 4585 | } | |||
| 4586 | npyv_storea_u64(seq_arg.data.qu64, vec_arg.data.vu64); | |||
| 4587 | // write-back | |||
| 4588 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4589 | simd_arg_free(&seq_arg); | |||
| 4590 | return NULL((void*)0); | |||
| 4591 | } | |||
| 4592 | simd_arg_free(&seq_arg); | |||
| 4593 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4594 | } | |||
| 4595 | ||||
| 4596 | #line 46 | |||
| 4597 | // special definition due to the nature of stores | |||
| 4598 | static PyObject * | |||
| 4599 | simd__intrin_stores_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4600 | { | |||
| 4601 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4602 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4603 | if (!PyArg_ParseTuple( | |||
| 4604 | args, "O&O&:stores_u64", | |||
| 4605 | simd_arg_converter, &seq_arg, | |||
| 4606 | simd_arg_converter, &vec_arg | |||
| 4607 | )) { | |||
| 4608 | return NULL((void*)0); | |||
| 4609 | } | |||
| 4610 | npyv_stores_u64(seq_arg.data.qu64, vec_arg.data.vu64); | |||
| 4611 | // write-back | |||
| 4612 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4613 | simd_arg_free(&seq_arg); | |||
| 4614 | return NULL((void*)0); | |||
| 4615 | } | |||
| 4616 | simd_arg_free(&seq_arg); | |||
| 4617 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4618 | } | |||
| 4619 | ||||
| 4620 | #line 46 | |||
| 4621 | // special definition due to the nature of storel | |||
| 4622 | static PyObject * | |||
| 4623 | simd__intrin_storel_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4624 | { | |||
| 4625 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4626 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4627 | if (!PyArg_ParseTuple( | |||
| 4628 | args, "O&O&:storel_u64", | |||
| 4629 | simd_arg_converter, &seq_arg, | |||
| 4630 | simd_arg_converter, &vec_arg | |||
| 4631 | )) { | |||
| 4632 | return NULL((void*)0); | |||
| 4633 | } | |||
| 4634 | npyv_storel_u64(seq_arg.data.qu64, vec_arg.data.vu64); | |||
| 4635 | // write-back | |||
| 4636 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4637 | simd_arg_free(&seq_arg); | |||
| 4638 | return NULL((void*)0); | |||
| 4639 | } | |||
| 4640 | simd_arg_free(&seq_arg); | |||
| 4641 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4642 | } | |||
| 4643 | ||||
| 4644 | #line 46 | |||
| 4645 | // special definition due to the nature of storeh | |||
| 4646 | static PyObject * | |||
| 4647 | simd__intrin_storeh_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4648 | { | |||
| 4649 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4650 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4651 | if (!PyArg_ParseTuple( | |||
| 4652 | args, "O&O&:storeh_u64", | |||
| 4653 | simd_arg_converter, &seq_arg, | |||
| 4654 | simd_arg_converter, &vec_arg | |||
| 4655 | )) { | |||
| 4656 | return NULL((void*)0); | |||
| 4657 | } | |||
| 4658 | npyv_storeh_u64(seq_arg.data.qu64, vec_arg.data.vu64); | |||
| 4659 | // write-back | |||
| 4660 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4661 | simd_arg_free(&seq_arg); | |||
| 4662 | return NULL((void*)0); | |||
| 4663 | } | |||
| 4664 | simd_arg_free(&seq_arg); | |||
| 4665 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4666 | } | |||
| 4667 | ||||
| 4668 | ||||
| 4669 | /**************************************** | |||
| 4670 | * Non-contiguous/Partial Memory access | |||
| 4671 | ****************************************/ | |||
| 4672 | #if 1 | |||
| 4673 | // Partial Load | |||
| 4674 | SIMD_IMPL_INTRIN_3(load_till_u64, vu64, qu64, u32, u64)static PyObject *simd__intrin_load_till_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu64}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_u64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_load_till_u64 ( arg1.data.qu64, arg2.data.u32, arg3.data.u64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4675 | SIMD_IMPL_INTRIN_2(load_tillz_u64, vu64, qu64, u32)static PyObject *simd__intrin_load_tillz_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qu64}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_u64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu64 = npyv_load_tillz_u64 ( arg1.data.qu64, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 4676 | ||||
| 4677 | // Partial Store | |||
| 4678 | static PyObject * | |||
| 4679 | simd__intrin_store_till_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4680 | { | |||
| 4681 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4682 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4683 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4684 | if (!PyArg_ParseTuple( | |||
| 4685 | args, "O&O&O&:store_till_u64", | |||
| 4686 | simd_arg_converter, &seq_arg, | |||
| 4687 | simd_arg_converter, &nlane_arg, | |||
| 4688 | simd_arg_converter, &vec_arg | |||
| 4689 | )) { | |||
| 4690 | return NULL((void*)0); | |||
| 4691 | } | |||
| 4692 | npyv_store_till_u64( | |||
| 4693 | seq_arg.data.qu64, nlane_arg.data.u32, vec_arg.data.vu64 | |||
| 4694 | ); | |||
| 4695 | // write-back | |||
| 4696 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4697 | simd_arg_free(&seq_arg); | |||
| 4698 | return NULL((void*)0); | |||
| 4699 | } | |||
| 4700 | simd_arg_free(&seq_arg); | |||
| 4701 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4702 | } | |||
| 4703 | ||||
| 4704 | // Non-contiguous Load | |||
| 4705 | #line 112 | |||
| 4706 | static PyObject * | |||
| 4707 | simd__intrin_loadn_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4708 | { | |||
| 4709 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4710 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4711 | #if 0 | |||
| 4712 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4713 | #endif // till | |||
| 4714 | #if 0 | |||
| 4715 | simd_arg fill_arg = {.dtype = simd_data_u64}; | |||
| 4716 | #endif | |||
| 4717 | if (!PyArg_ParseTuple( | |||
| 4718 | args, "O&O&:loadn_u64", | |||
| 4719 | simd_arg_converter, &seq_arg, | |||
| 4720 | simd_arg_converter, &stride_arg | |||
| 4721 | #if 0 | |||
| 4722 | ,simd_arg_converter, &nlane_arg | |||
| 4723 | #endif | |||
| 4724 | #if 0 | |||
| 4725 | ,simd_arg_converter, &fill_arg | |||
| 4726 | #endif | |||
| 4727 | )) { | |||
| 4728 | return NULL((void*)0); | |||
| 4729 | } | |||
| 4730 | npyv_lanetype_u64 *seq_ptr = seq_arg.data.qu64; | |||
| 4731 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4732 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4733 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u642; | |||
| 4734 | if (stride < 0) { | |||
| 4735 | seq_ptr += cur_seq_len -1; | |||
| 4736 | min_seq_len = -min_seq_len; | |||
| 4737 | } | |||
| 4738 | if (cur_seq_len < min_seq_len) { | |||
| 4739 | PyErr_Format(PyExc_ValueError, | |||
| 4740 | "loadn_u64(), according to provided stride %d, the " | |||
| 4741 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4742 | stride, min_seq_len, cur_seq_len | |||
| 4743 | ); | |||
| 4744 | goto err; | |||
| 4745 | } | |||
| 4746 | npyv_u64 rvec = npyv_loadn_u64( | |||
| 4747 | seq_ptr, stride | |||
| 4748 | #if 0 | |||
| 4749 | , nlane_arg.data.u32 | |||
| 4750 | #endif | |||
| 4751 | #if 0 | |||
| 4752 | , fill_arg.data.u64 | |||
| 4753 | #endif | |||
| 4754 | ); | |||
| 4755 | simd_arg ret = { | |||
| 4756 | .dtype = simd_data_vu64, .data = {.vu64=rvec} | |||
| 4757 | }; | |||
| 4758 | simd_arg_free(&seq_arg); | |||
| 4759 | return simd_arg_to_obj(&ret); | |||
| 4760 | err: | |||
| 4761 | simd_arg_free(&seq_arg); | |||
| 4762 | return NULL((void*)0); | |||
| 4763 | } | |||
| 4764 | ||||
| 4765 | #line 112 | |||
| 4766 | static PyObject * | |||
| 4767 | simd__intrin_loadn_till_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4768 | { | |||
| 4769 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4770 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4771 | #if 1 | |||
| 4772 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4773 | #endif // till | |||
| 4774 | #if 1 | |||
| 4775 | simd_arg fill_arg = {.dtype = simd_data_u64}; | |||
| 4776 | #endif | |||
| 4777 | if (!PyArg_ParseTuple( | |||
| 4778 | args, "O&O&O&O&:loadn_till_u64", | |||
| 4779 | simd_arg_converter, &seq_arg, | |||
| 4780 | simd_arg_converter, &stride_arg | |||
| 4781 | #if 1 | |||
| 4782 | ,simd_arg_converter, &nlane_arg | |||
| 4783 | #endif | |||
| 4784 | #if 1 | |||
| 4785 | ,simd_arg_converter, &fill_arg | |||
| 4786 | #endif | |||
| 4787 | )) { | |||
| 4788 | return NULL((void*)0); | |||
| 4789 | } | |||
| 4790 | npyv_lanetype_u64 *seq_ptr = seq_arg.data.qu64; | |||
| 4791 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4792 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4793 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u642; | |||
| 4794 | if (stride < 0) { | |||
| 4795 | seq_ptr += cur_seq_len -1; | |||
| 4796 | min_seq_len = -min_seq_len; | |||
| 4797 | } | |||
| 4798 | if (cur_seq_len < min_seq_len) { | |||
| 4799 | PyErr_Format(PyExc_ValueError, | |||
| 4800 | "loadn_till_u64(), according to provided stride %d, the " | |||
| 4801 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4802 | stride, min_seq_len, cur_seq_len | |||
| 4803 | ); | |||
| 4804 | goto err; | |||
| 4805 | } | |||
| 4806 | npyv_u64 rvec = npyv_loadn_till_u64( | |||
| 4807 | seq_ptr, stride | |||
| 4808 | #if 1 | |||
| 4809 | , nlane_arg.data.u32 | |||
| 4810 | #endif | |||
| 4811 | #if 1 | |||
| 4812 | , fill_arg.data.u64 | |||
| 4813 | #endif | |||
| 4814 | ); | |||
| 4815 | simd_arg ret = { | |||
| 4816 | .dtype = simd_data_vu64, .data = {.vu64=rvec} | |||
| 4817 | }; | |||
| 4818 | simd_arg_free(&seq_arg); | |||
| 4819 | return simd_arg_to_obj(&ret); | |||
| 4820 | err: | |||
| 4821 | simd_arg_free(&seq_arg); | |||
| 4822 | return NULL((void*)0); | |||
| 4823 | } | |||
| 4824 | ||||
| 4825 | #line 112 | |||
| 4826 | static PyObject * | |||
| 4827 | simd__intrin_loadn_tillz_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4828 | { | |||
| 4829 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4830 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4831 | #if 1 | |||
| 4832 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4833 | #endif // till | |||
| 4834 | #if 0 | |||
| 4835 | simd_arg fill_arg = {.dtype = simd_data_u64}; | |||
| 4836 | #endif | |||
| 4837 | if (!PyArg_ParseTuple( | |||
| 4838 | args, "O&O&O&:loadn_tillz_u64", | |||
| 4839 | simd_arg_converter, &seq_arg, | |||
| 4840 | simd_arg_converter, &stride_arg | |||
| 4841 | #if 1 | |||
| 4842 | ,simd_arg_converter, &nlane_arg | |||
| 4843 | #endif | |||
| 4844 | #if 0 | |||
| 4845 | ,simd_arg_converter, &fill_arg | |||
| 4846 | #endif | |||
| 4847 | )) { | |||
| 4848 | return NULL((void*)0); | |||
| 4849 | } | |||
| 4850 | npyv_lanetype_u64 *seq_ptr = seq_arg.data.qu64; | |||
| 4851 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4852 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4853 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u642; | |||
| 4854 | if (stride < 0) { | |||
| 4855 | seq_ptr += cur_seq_len -1; | |||
| 4856 | min_seq_len = -min_seq_len; | |||
| 4857 | } | |||
| 4858 | if (cur_seq_len < min_seq_len) { | |||
| 4859 | PyErr_Format(PyExc_ValueError, | |||
| 4860 | "loadn_tillz_u64(), according to provided stride %d, the " | |||
| 4861 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4862 | stride, min_seq_len, cur_seq_len | |||
| 4863 | ); | |||
| 4864 | goto err; | |||
| 4865 | } | |||
| 4866 | npyv_u64 rvec = npyv_loadn_tillz_u64( | |||
| 4867 | seq_ptr, stride | |||
| 4868 | #if 1 | |||
| 4869 | , nlane_arg.data.u32 | |||
| 4870 | #endif | |||
| 4871 | #if 0 | |||
| 4872 | , fill_arg.data.u64 | |||
| 4873 | #endif | |||
| 4874 | ); | |||
| 4875 | simd_arg ret = { | |||
| 4876 | .dtype = simd_data_vu64, .data = {.vu64=rvec} | |||
| 4877 | }; | |||
| 4878 | simd_arg_free(&seq_arg); | |||
| 4879 | return simd_arg_to_obj(&ret); | |||
| 4880 | err: | |||
| 4881 | simd_arg_free(&seq_arg); | |||
| 4882 | return NULL((void*)0); | |||
| 4883 | } | |||
| 4884 | ||||
| 4885 | ||||
| 4886 | // Non-contiguous Store | |||
| 4887 | #line 178 | |||
| 4888 | static PyObject * | |||
| 4889 | simd__intrin_storen_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4890 | { | |||
| 4891 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4892 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4893 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4894 | #if 0 | |||
| 4895 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4896 | #endif | |||
| 4897 | if (!PyArg_ParseTuple( | |||
| 4898 | args, "O&O&O&:storen_u64", | |||
| 4899 | simd_arg_converter, &seq_arg, | |||
| 4900 | simd_arg_converter, &stride_arg | |||
| 4901 | #if 0 | |||
| 4902 | ,simd_arg_converter, &nlane_arg | |||
| 4903 | #endif | |||
| 4904 | ,simd_arg_converter, &vec_arg | |||
| 4905 | )) { | |||
| 4906 | return NULL((void*)0); | |||
| 4907 | } | |||
| 4908 | npyv_lanetype_u64 *seq_ptr = seq_arg.data.qu64; | |||
| 4909 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4910 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4911 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u642; | |||
| 4912 | if (stride < 0) { | |||
| 4913 | seq_ptr += cur_seq_len -1; | |||
| 4914 | min_seq_len = -min_seq_len; | |||
| 4915 | } | |||
| 4916 | // overflow guard | |||
| 4917 | if (cur_seq_len < min_seq_len) { | |||
| 4918 | PyErr_Format(PyExc_ValueError, | |||
| 4919 | "storen_u64(), according to provided stride %d, the" | |||
| 4920 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4921 | stride, min_seq_len, cur_seq_len | |||
| 4922 | ); | |||
| 4923 | goto err; | |||
| 4924 | } | |||
| 4925 | npyv_storen_u64( | |||
| 4926 | seq_ptr, stride | |||
| 4927 | #if 0 | |||
| 4928 | ,nlane_arg.data.u32 | |||
| 4929 | #endif | |||
| 4930 | ,vec_arg.data.vu64 | |||
| 4931 | ); | |||
| 4932 | // write-back | |||
| 4933 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4934 | goto err; | |||
| 4935 | } | |||
| 4936 | simd_arg_free(&seq_arg); | |||
| 4937 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4938 | err: | |||
| 4939 | simd_arg_free(&seq_arg); | |||
| 4940 | return NULL((void*)0); | |||
| 4941 | } | |||
| 4942 | ||||
| 4943 | #line 178 | |||
| 4944 | static PyObject * | |||
| 4945 | simd__intrin_storen_till_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 4946 | { | |||
| 4947 | simd_arg seq_arg = {.dtype = simd_data_qu64}; | |||
| 4948 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 4949 | simd_arg vec_arg = {.dtype = simd_data_vu64}; | |||
| 4950 | #if 1 | |||
| 4951 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 4952 | #endif | |||
| 4953 | if (!PyArg_ParseTuple( | |||
| 4954 | args, "O&O&O&O&:storen_u64", | |||
| 4955 | simd_arg_converter, &seq_arg, | |||
| 4956 | simd_arg_converter, &stride_arg | |||
| 4957 | #if 1 | |||
| 4958 | ,simd_arg_converter, &nlane_arg | |||
| 4959 | #endif | |||
| 4960 | ,simd_arg_converter, &vec_arg | |||
| 4961 | )) { | |||
| 4962 | return NULL((void*)0); | |||
| 4963 | } | |||
| 4964 | npyv_lanetype_u64 *seq_ptr = seq_arg.data.qu64; | |||
| 4965 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 4966 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 4967 | Py_ssize_t min_seq_len = stride * npyv_nlanes_u642; | |||
| 4968 | if (stride < 0) { | |||
| 4969 | seq_ptr += cur_seq_len -1; | |||
| 4970 | min_seq_len = -min_seq_len; | |||
| 4971 | } | |||
| 4972 | // overflow guard | |||
| 4973 | if (cur_seq_len < min_seq_len) { | |||
| 4974 | PyErr_Format(PyExc_ValueError, | |||
| 4975 | "storen_till_u64(), according to provided stride %d, the" | |||
| 4976 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 4977 | stride, min_seq_len, cur_seq_len | |||
| 4978 | ); | |||
| 4979 | goto err; | |||
| 4980 | } | |||
| 4981 | npyv_storen_till_u64( | |||
| 4982 | seq_ptr, stride | |||
| 4983 | #if 1 | |||
| 4984 | ,nlane_arg.data.u32 | |||
| 4985 | #endif | |||
| 4986 | ,vec_arg.data.vu64 | |||
| 4987 | ); | |||
| 4988 | // write-back | |||
| 4989 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qu64, simd_data_qu64)) { | |||
| 4990 | goto err; | |||
| 4991 | } | |||
| 4992 | simd_arg_free(&seq_arg); | |||
| 4993 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 4994 | err: | |||
| 4995 | simd_arg_free(&seq_arg); | |||
| 4996 | return NULL((void*)0); | |||
| 4997 | } | |||
| 4998 | ||||
| 4999 | #endif // 1 | |||
| 5000 | ||||
| 5001 | /*************************** | |||
| 5002 | * Misc | |||
| 5003 | ***************************/ | |||
| 5004 | SIMD_IMPL_INTRIN_0(zero_u64, vu64)static PyObject *simd__intrin_zero_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_u64") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vu64, .data = {.vu64 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 5005 | SIMD_IMPL_INTRIN_1(setall_u64, vu64, u64)static PyObject *simd__intrin_setall_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u64}; if (!PyArg_ParseTuple( args, "O&:" "setall_u64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu64 = _mm_set1_epi64x((npy_int64)(arg .data.u64))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(& ret); } | |||
| 5006 | SIMD_IMPL_INTRIN_3(select_u64, vu64, vb64, vu64, vu64)static PyObject *simd__intrin_select_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_select_u8 ( arg1.data.vb64, arg2.data.vu64, arg3.data.vu64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5007 | ||||
| 5008 | #line 246 | |||
| 5009 | #if 1 | |||
| 5010 | SIMD_IMPL_INTRIN_1(reinterpret_u8_u64, vu8, vu64)static PyObject *simd__intrin_reinterpret_u8_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vu64 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 5011 | #endif // simd_sup2 | |||
| 5012 | ||||
| 5013 | #line 246 | |||
| 5014 | #if 1 | |||
| 5015 | SIMD_IMPL_INTRIN_1(reinterpret_s8_u64, vs8, vu64)static PyObject *simd__intrin_reinterpret_s8_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vu64 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 5016 | #endif // simd_sup2 | |||
| 5017 | ||||
| 5018 | #line 246 | |||
| 5019 | #if 1 | |||
| 5020 | SIMD_IMPL_INTRIN_1(reinterpret_u16_u64, vu16, vu64)static PyObject *simd__intrin_reinterpret_u16_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 5021 | #endif // simd_sup2 | |||
| 5022 | ||||
| 5023 | #line 246 | |||
| 5024 | #if 1 | |||
| 5025 | SIMD_IMPL_INTRIN_1(reinterpret_s16_u64, vs16, vu64)static PyObject *simd__intrin_reinterpret_s16_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 5026 | #endif // simd_sup2 | |||
| 5027 | ||||
| 5028 | #line 246 | |||
| 5029 | #if 1 | |||
| 5030 | SIMD_IMPL_INTRIN_1(reinterpret_u32_u64, vu32, vu64)static PyObject *simd__intrin_reinterpret_u32_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 5031 | #endif // simd_sup2 | |||
| 5032 | ||||
| 5033 | #line 246 | |||
| 5034 | #if 1 | |||
| 5035 | SIMD_IMPL_INTRIN_1(reinterpret_s32_u64, vs32, vu64)static PyObject *simd__intrin_reinterpret_s32_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 5036 | #endif // simd_sup2 | |||
| 5037 | ||||
| 5038 | #line 246 | |||
| 5039 | #if 1 | |||
| 5040 | SIMD_IMPL_INTRIN_1(reinterpret_u64_u64, vu64, vu64)static PyObject *simd__intrin_reinterpret_u64_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5041 | #endif // simd_sup2 | |||
| 5042 | ||||
| 5043 | #line 246 | |||
| 5044 | #if 1 | |||
| 5045 | SIMD_IMPL_INTRIN_1(reinterpret_s64_u64, vs64, vu64)static PyObject *simd__intrin_reinterpret_s64_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vu64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5046 | #endif // simd_sup2 | |||
| 5047 | ||||
| 5048 | #line 246 | |||
| 5049 | #if 1 | |||
| 5050 | SIMD_IMPL_INTRIN_1(reinterpret_f32_u64, vf32, vu64)static PyObject *simd__intrin_reinterpret_f32_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vu64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 5051 | #endif // simd_sup2 | |||
| 5052 | ||||
| 5053 | #line 246 | |||
| 5054 | #if NPY_SIMD_F641 | |||
| 5055 | SIMD_IMPL_INTRIN_1(reinterpret_f64_u64, vf64, vu64)static PyObject *simd__intrin_reinterpret_f64_u64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_u64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vu64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 5056 | #endif // simd_sup2 | |||
| 5057 | ||||
| 5058 | ||||
| 5059 | /** | |||
| 5060 | * special definition due to the nature of intrinsics | |||
| 5061 | * npyv_setf_u64 and npy_set_u64. | |||
| 5062 | */ | |||
| 5063 | #line 258 | |||
| 5064 | static PyObject * | |||
| 5065 | simd__intrin_setf_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5066 | { | |||
| 5067 | npyv_lanetype_u64 *data = simd_sequence_from_iterable(args, simd_data_qu64, npyv_nlanes_u642); | |||
| 5068 | if (data == NULL((void*)0)) { | |||
| 5069 | return NULL((void*)0); | |||
| 5070 | } | |||
| 5071 | simd_data r = {.vu64 = npyv_setf_u64(npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5072 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5073 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5074 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5075 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5076 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5077 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5078 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5079 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5080 | data[64] // for setfnpyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5081 | )npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2]))}; | |||
| 5082 | simd_sequence_free(data); | |||
| 5083 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu64); | |||
| 5084 | } | |||
| 5085 | ||||
| 5086 | #line 258 | |||
| 5087 | static PyObject * | |||
| 5088 | simd__intrin_set_u64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5089 | { | |||
| 5090 | npyv_lanetype_u64 *data = simd_sequence_from_iterable(args, simd_data_qu64, npyv_nlanes_u642); | |||
| 5091 | if (data == NULL((void*)0)) { | |||
| 5092 | return NULL((void*)0); | |||
| 5093 | } | |||
| 5094 | simd_data r = {.vu64 = npyv_set_u64(npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5095 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5096 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5097 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5098 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5099 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5100 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5101 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5102 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5103 | data[64] // for setfnpyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5104 | )npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1]))}; | |||
| 5105 | simd_sequence_free(data); | |||
| 5106 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vu64); | |||
| 5107 | } | |||
| 5108 | ||||
| 5109 | ||||
| 5110 | /*************************** | |||
| 5111 | * Reorder | |||
| 5112 | ***************************/ | |||
| 5113 | #line 287 | |||
| 5114 | SIMD_IMPL_INTRIN_2(combinel_u64, vu64, vu64, vu64)static PyObject *simd__intrin_combinel_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_u64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu64 = _mm_unpacklo_epi64 ( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5115 | ||||
| 5116 | #line 287 | |||
| 5117 | SIMD_IMPL_INTRIN_2(combineh_u64, vu64, vu64, vu64)static PyObject *simd__intrin_combineh_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_u64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu64 = _mm_unpackhi_epi64 ( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5118 | ||||
| 5119 | ||||
| 5120 | #line 293 | |||
| 5121 | SIMD_IMPL_INTRIN_2(combine_u64, vu64x2, vu64, vu64)static PyObject *simd__intrin_combine_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_u64", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vu64x2 = npyv__combine ( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64x2 }; return simd_arg_to_obj(&ret); } | |||
| 5122 | ||||
| 5123 | #line 293 | |||
| 5124 | SIMD_IMPL_INTRIN_2(zip_u64, vu64x2, vu64, vu64)static PyObject *simd__intrin_zip_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64x2 = npyv_zip_u64( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64x2 }; return simd_arg_to_obj(&ret); } | |||
| 5125 | ||||
| 5126 | ||||
| 5127 | #if 0 | |||
| 5128 | SIMD_IMPL_INTRIN_1(rev64_u64, vu64, vu64)static PyObject *simd__intrin_rev64_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "rev64_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu64 = npyv_rev64_u64( arg.data.vu64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5129 | #endif | |||
| 5130 | ||||
| 5131 | /*************************** | |||
| 5132 | * Operators | |||
| 5133 | ***************************/ | |||
| 5134 | #if 63 > 0 | |||
| 5135 | SIMD_IMPL_INTRIN_2(shl_u64, vu64, vu64, u8)static PyObject *simd__intrin_shl_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_sll_epi64(arg1.data.vu64, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5136 | SIMD_IMPL_INTRIN_2(shr_u64, vu64, vu64, u8)static PyObject *simd__intrin_shr_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_srl_epi64(arg1.data.vu64, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5137 | // immediate constant | |||
| 5138 | SIMD_IMPL_INTRIN_2IMM(shli_u64, vu64, vu64, 63)static PyObject *simd__intrin_shli_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu64 = 0 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 0) : 1 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 1) : 2 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 2) : 3 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 3) : 4 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 5) : 6 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 6) : 7 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vu64, 8) : 9 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vu64, 9) : 10 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 10) : 11 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 11) : 12 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 12) : 13 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 13) : 14 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 15 ) : 16 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 16) : 17 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 17) : 18 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 18) : 19 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 19) : 20 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vu64, 20) : 21 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vu64, 21) : 22 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 22) : 23 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 23) : 24 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 24) : 25 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 25) : 26 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 26) : 27 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 27 ) : 28 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 28) : 29 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 29) : 30 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 30) : 31 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 31) : 32 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vu64, 32) : 33 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vu64, 33) : 34 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 34) : 35 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 35) : 36 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 36) : 37 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 37) : 38 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 38) : 39 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 39 ) : 40 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 40) : 41 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 41) : 42 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 42) : 43 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 43) : 44 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vu64, 44) : 45 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vu64, 45) : 46 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 46) : 47 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 47) : 48 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 48) : 49 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 49) : 50 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 50) : 51 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 51 ) : 52 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 52) : 53 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 53) : 54 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 54) : 55 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 55) : 56 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vu64, 56) : 57 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vu64, 57) : 58 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 58) : 59 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vu64, 59) : 60 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vu64, 60) : 61 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vu64, 61) : 62 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64 , 62) : 63 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vu64, 63 ) : data.vu64; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(& ret); } | |||
| 5139 | SIMD_IMPL_INTRIN_2IMM(shri_u64, vu64, vu64, 64)static PyObject *simd__intrin_shri_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vu64 = 1 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 1) : 2 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 2) : 3 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 3) : 4 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 4) : 5 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 5) : 6 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 6) : 7 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 7) : 8 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 8) : 9 == arg2 .data.u8 ? _mm_srli_epi64(arg1.data.vu64, 9) : 10 == arg2.data .u8 ? _mm_srli_epi64(arg1.data.vu64, 10) : 11 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 11) : 12 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 12) : 13 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 13) : 14 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 14) : 15 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 15) : 16 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 16 ) : 17 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 17) : 18 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 18) : 19 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 19) : 20 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 20) : 21 == arg2 .data.u8 ? _mm_srli_epi64(arg1.data.vu64, 21) : 22 == arg2.data .u8 ? _mm_srli_epi64(arg1.data.vu64, 22) : 23 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 23) : 24 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 24) : 25 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 25) : 26 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 26) : 27 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 27) : 28 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 28 ) : 29 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 29) : 30 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 30) : 31 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 31) : 32 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 32) : 33 == arg2 .data.u8 ? _mm_srli_epi64(arg1.data.vu64, 33) : 34 == arg2.data .u8 ? _mm_srli_epi64(arg1.data.vu64, 34) : 35 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 35) : 36 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 36) : 37 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 37) : 38 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 38) : 39 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 39) : 40 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 40 ) : 41 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 41) : 42 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 42) : 43 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 43) : 44 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 44) : 45 == arg2 .data.u8 ? _mm_srli_epi64(arg1.data.vu64, 45) : 46 == arg2.data .u8 ? _mm_srli_epi64(arg1.data.vu64, 46) : 47 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 47) : 48 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 48) : 49 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 49) : 50 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 50) : 51 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 51) : 52 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 52 ) : 53 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 53) : 54 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 54) : 55 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 55) : 56 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 56) : 57 == arg2 .data.u8 ? _mm_srli_epi64(arg1.data.vu64, 57) : 58 == arg2.data .u8 ? _mm_srli_epi64(arg1.data.vu64, 58) : 59 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 59) : 60 == arg2.data.u8 ? _mm_srli_epi64 (arg1.data.vu64, 60) : 61 == arg2.data.u8 ? _mm_srli_epi64(arg1 .data.vu64, 61) : 62 == arg2.data.u8 ? _mm_srli_epi64(arg1.data .vu64, 62) : 63 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64 , 63) : 64 == arg2.data.u8 ? _mm_srli_epi64(arg1.data.vu64, 64 ) : data.vu64; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(& ret); } | |||
| 5140 | #endif // shl_imm | |||
| 5141 | ||||
| 5142 | #line 314 | |||
| 5143 | SIMD_IMPL_INTRIN_2(and_u64, vu64, vu64, vu64)static PyObject *simd__intrin_and_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""and_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_and_si128( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5144 | ||||
| 5145 | #line 314 | |||
| 5146 | SIMD_IMPL_INTRIN_2(or_u64, vu64, vu64, vu64)static PyObject *simd__intrin_or_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""or_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_or_si128( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5147 | ||||
| 5148 | #line 314 | |||
| 5149 | SIMD_IMPL_INTRIN_2(xor_u64, vu64, vu64, vu64)static PyObject *simd__intrin_xor_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_xor_si128( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5150 | ||||
| 5151 | ||||
| 5152 | SIMD_IMPL_INTRIN_1(not_u64, vu64, vu64)static PyObject *simd__intrin_not_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "not_u64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu64 = _mm_xor_si128(arg.data.vu64, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5153 | ||||
| 5154 | #line 322 | |||
| 5155 | SIMD_IMPL_INTRIN_2(cmpeq_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmpeq_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpeq_u64( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5156 | ||||
| 5157 | #line 322 | |||
| 5158 | SIMD_IMPL_INTRIN_2(cmpneq_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmpneq_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpeq_u64(arg1 .data.vu64, arg2.data.vu64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5159 | ||||
| 5160 | #line 322 | |||
| 5161 | SIMD_IMPL_INTRIN_2(cmpgt_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmpgt_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpgt_u64( arg1.data.vu64, arg2.data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5162 | ||||
| 5163 | #line 322 | |||
| 5164 | SIMD_IMPL_INTRIN_2(cmpge_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmpge_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpgt_u64(arg2 .data.vu64, arg1.data.vu64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5165 | ||||
| 5166 | #line 322 | |||
| 5167 | SIMD_IMPL_INTRIN_2(cmplt_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmplt_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpgt_u64(arg2.data.vu64, arg1 .data.vu64)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5168 | ||||
| 5169 | #line 322 | |||
| 5170 | SIMD_IMPL_INTRIN_2(cmple_u64, vb64, vu64, vu64)static PyObject *simd__intrin_cmple_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpgt_u64(arg1 .data.vu64, arg2.data.vu64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5171 | ||||
| 5172 | ||||
| 5173 | /*************************** | |||
| 5174 | * Conversion | |||
| 5175 | ***************************/ | |||
| 5176 | SIMD_IMPL_INTRIN_1(cvt_u64_b64, vu64, vb64)static PyObject *simd__intrin_cvt_u64_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_u64_b64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu64 = arg.data.vb64}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5177 | SIMD_IMPL_INTRIN_1(cvt_b64_u64, vb64, vu64)static PyObject *simd__intrin_cvt_b64_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b64_u64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb64 = arg.data.vu64}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5178 | #if 0 | |||
| 5179 | SIMD_IMPL_INTRIN_1(expand_u64_u64, vu64x2, vu64)static PyObject *simd__intrin_expand_u64_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "expand_u64_u64", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vu64x2 = npyv_expand_u64_u64( arg.data .vu64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64x2 }; return simd_arg_to_obj(&ret ); } | |||
| 5180 | #endif // expand_sup | |||
| 5181 | /*************************** | |||
| 5182 | * Arithmetic | |||
| 5183 | ***************************/ | |||
| 5184 | #line 339 | |||
| 5185 | SIMD_IMPL_INTRIN_2(add_u64, vu64, vu64, vu64)static PyObject *simd__intrin_add_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""add_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_add_epi64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5186 | ||||
| 5187 | #line 339 | |||
| 5188 | SIMD_IMPL_INTRIN_2(sub_u64, vu64, vu64, vu64)static PyObject *simd__intrin_sub_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = _mm_sub_epi64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5189 | ||||
| 5190 | ||||
| 5191 | #if 0 | |||
| 5192 | #line 346 | |||
| 5193 | SIMD_IMPL_INTRIN_2(adds_u64, vu64, vu64, vu64)static PyObject *simd__intrin_adds_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_adds_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5194 | ||||
| 5195 | #line 346 | |||
| 5196 | SIMD_IMPL_INTRIN_2(subs_u64, vu64, vu64, vu64)static PyObject *simd__intrin_subs_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_subs_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5197 | ||||
| 5198 | #endif // sat_sup | |||
| 5199 | ||||
| 5200 | #if 0 | |||
| 5201 | SIMD_IMPL_INTRIN_2(mul_u64, vu64, vu64, vu64)static PyObject *simd__intrin_mul_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_mul_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5202 | #endif // mul_sup | |||
| 5203 | ||||
| 5204 | #if 0 | |||
| 5205 | SIMD_IMPL_INTRIN_2(div_u64, vu64, vu64, vu64)static PyObject *simd__intrin_div_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""div_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_div_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5206 | #endif // div_sup | |||
| 5207 | ||||
| 5208 | #if 1 | |||
| 5209 | SIMD_IMPL_INTRIN_1(divisor_u64, vu64x3, u64)static PyObject *simd__intrin_divisor_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_u64}; if (!PyArg_ParseTuple( args, "O&:" "divisor_u64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vu64x3 = npyv_divisor_u64( arg.data.u64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vu64x3 }; return simd_arg_to_obj(&ret); } | |||
| 5210 | SIMD_IMPL_INTRIN_2(divc_u64, vu64, vu64, vu64x3)static PyObject *simd__intrin_divc_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_divc_u64( arg1.data.vu64, arg2 .data.vu64x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5211 | #endif // intdiv_sup | |||
| 5212 | ||||
| 5213 | #if 0 | |||
| 5214 | #line 367 | |||
| 5215 | SIMD_IMPL_INTRIN_3(muladd_u64, vu64, vu64, vu64, vu64)static PyObject *simd__intrin_muladd_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_muladd_u64 ( arg1.data.vu64, arg2.data.vu64, arg3.data.vu64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5216 | ||||
| 5217 | #line 367 | |||
| 5218 | SIMD_IMPL_INTRIN_3(mulsub_u64, vu64, vu64, vu64, vu64)static PyObject *simd__intrin_mulsub_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_mulsub_u64 ( arg1.data.vu64, arg2.data.vu64, arg3.data.vu64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5219 | ||||
| 5220 | #line 367 | |||
| 5221 | SIMD_IMPL_INTRIN_3(nmuladd_u64, vu64, vu64, vu64, vu64)static PyObject *simd__intrin_nmuladd_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_nmuladd_u64 ( arg1.data.vu64, arg2.data.vu64, arg3.data.vu64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5222 | ||||
| 5223 | #line 367 | |||
| 5224 | SIMD_IMPL_INTRIN_3(nmulsub_u64, vu64, vu64, vu64, vu64)static PyObject *simd__intrin_nmulsub_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vu64 = npyv_nmulsub_u64 ( arg1.data.vu64, arg2.data.vu64, arg3.data.vu64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5225 | ||||
| 5226 | #endif // fused_sup | |||
| 5227 | ||||
| 5228 | #if 1 | |||
| 5229 | SIMD_IMPL_INTRIN_1(sum_u64, u64, vu64)static PyObject *simd__intrin_sum_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "sum_u64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.u64 = npyv_sum_u64( arg.data.vu64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 5230 | #endif // sum_sup | |||
| 5231 | ||||
| 5232 | #if 0 | |||
| 5233 | SIMD_IMPL_INTRIN_1(sumup_u64, u64, vu64)static PyObject *simd__intrin_sumup_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "sumup_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.u64 = npyv_sumup_u64( arg.data.vu64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 5234 | #endif // sumup_sup | |||
| 5235 | ||||
| 5236 | /*************************** | |||
| 5237 | * Math | |||
| 5238 | ***************************/ | |||
| 5239 | #if 0 | |||
| 5240 | #line 386 | |||
| 5241 | SIMD_IMPL_INTRIN_1(sqrt_u64, vu64, vu64)static PyObject *simd__intrin_sqrt_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_u64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vu64 = npyv_sqrt_u64( arg.data.vu64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5242 | ||||
| 5243 | #line 386 | |||
| 5244 | SIMD_IMPL_INTRIN_1(recip_u64, vu64, vu64)static PyObject *simd__intrin_recip_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "recip_u64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vu64 = npyv_recip_u64( arg.data.vu64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5245 | ||||
| 5246 | #line 386 | |||
| 5247 | SIMD_IMPL_INTRIN_1(abs_u64, vu64, vu64)static PyObject *simd__intrin_abs_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "abs_u64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vu64 = npyv_abs_u64( arg.data.vu64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5248 | ||||
| 5249 | #line 386 | |||
| 5250 | SIMD_IMPL_INTRIN_1(square_u64, vu64, vu64)static PyObject *simd__intrin_square_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&:" "square_u64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vu64 = npyv_square_u64( arg.data.vu64 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5251 | ||||
| 5252 | #endif | |||
| 5253 | ||||
| 5254 | #line 393 | |||
| 5255 | SIMD_IMPL_INTRIN_2(max_u64, vu64, vu64, vu64)static PyObject *simd__intrin_max_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""max_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_max_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5256 | ||||
| 5257 | #line 393 | |||
| 5258 | SIMD_IMPL_INTRIN_2(min_u64, vu64, vu64, vu64)static PyObject *simd__intrin_min_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""min_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_min_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5259 | ||||
| 5260 | ||||
| 5261 | #if 0 | |||
| 5262 | #line 400 | |||
| 5263 | SIMD_IMPL_INTRIN_2(maxp_u64, vu64, vu64, vu64)static PyObject *simd__intrin_maxp_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_maxp_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5264 | ||||
| 5265 | #line 400 | |||
| 5266 | SIMD_IMPL_INTRIN_2(minp_u64, vu64, vu64, vu64)static PyObject *simd__intrin_minp_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vu64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_u64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vu64 = npyv_minp_u64( arg1.data.vu64, arg2 .data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret); } | |||
| 5267 | ||||
| 5268 | #endif | |||
| 5269 | ||||
| 5270 | /*************************** | |||
| 5271 | * Mask operations | |||
| 5272 | ***************************/ | |||
| 5273 | #line 410 | |||
| 5274 | SIMD_IMPL_INTRIN_4(ifadd_u64, vu64, vb64, vu64, vu64, vu64)static PyObject *simd__intrin_ifadd_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; simd_arg arg4 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_u64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu64 = npyv_ifadd_u64 ( arg1.data.vb64, arg2.data.vu64, arg3.data.vu64, arg4.data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(& ret); } | |||
| 5275 | ||||
| 5276 | #line 410 | |||
| 5277 | SIMD_IMPL_INTRIN_4(ifsub_u64, vu64, vb64, vu64, vu64, vu64)static PyObject *simd__intrin_ifsub_u64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vu64 }; simd_arg arg3 = {.dtype = simd_data_vu64}; simd_arg arg4 = {.dtype = simd_data_vu64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_u64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vu64 = npyv_ifsub_u64 ( arg1.data.vb64, arg2.data.vu64, arg3.data.vu64, arg4.data.vu64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj(& ret); } | |||
| 5278 | ||||
| 5279 | ||||
| 5280 | #endif // simd_sup | |||
| 5281 | ||||
| 5282 | #line 34 | |||
| 5283 | #if 1 | |||
| 5284 | /*************************** | |||
| 5285 | * Memory | |||
| 5286 | ***************************/ | |||
| 5287 | #line 41 | |||
| 5288 | SIMD_IMPL_INTRIN_1(load_s64, vs64, qs64)static PyObject *simd__intrin_load_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs64}; if (!PyArg_ParseTuple( args, "O&:" "load_s64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs64 = npyv_load_s64( arg.data.qs64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5289 | ||||
| 5290 | #line 41 | |||
| 5291 | SIMD_IMPL_INTRIN_1(loada_s64, vs64, qs64)static PyObject *simd__intrin_loada_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs64}; if (!PyArg_ParseTuple( args, "O&:" "loada_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs64 = npyv_loada_s64( arg.data.qs64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5292 | ||||
| 5293 | #line 41 | |||
| 5294 | SIMD_IMPL_INTRIN_1(loads_s64, vs64, qs64)static PyObject *simd__intrin_loads_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs64}; if (!PyArg_ParseTuple( args, "O&:" "loads_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs64 = npyv_loads_s64( arg.data.qs64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5295 | ||||
| 5296 | #line 41 | |||
| 5297 | SIMD_IMPL_INTRIN_1(loadl_s64, vs64, qs64)static PyObject *simd__intrin_loadl_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qs64}; if (!PyArg_ParseTuple( args, "O&:" "loadl_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs64 = npyv_loadl_s64( arg.data.qs64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5298 | ||||
| 5299 | #line 46 | |||
| 5300 | // special definition due to the nature of store | |||
| 5301 | static PyObject * | |||
| 5302 | simd__intrin_store_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5303 | { | |||
| 5304 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5305 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5306 | if (!PyArg_ParseTuple( | |||
| 5307 | args, "O&O&:store_s64", | |||
| 5308 | simd_arg_converter, &seq_arg, | |||
| 5309 | simd_arg_converter, &vec_arg | |||
| 5310 | )) { | |||
| 5311 | return NULL((void*)0); | |||
| 5312 | } | |||
| 5313 | npyv_store_s64(seq_arg.data.qs64, vec_arg.data.vs64); | |||
| 5314 | // write-back | |||
| 5315 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5316 | simd_arg_free(&seq_arg); | |||
| 5317 | return NULL((void*)0); | |||
| 5318 | } | |||
| 5319 | simd_arg_free(&seq_arg); | |||
| 5320 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5321 | } | |||
| 5322 | ||||
| 5323 | #line 46 | |||
| 5324 | // special definition due to the nature of storea | |||
| 5325 | static PyObject * | |||
| 5326 | simd__intrin_storea_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5327 | { | |||
| 5328 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5329 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5330 | if (!PyArg_ParseTuple( | |||
| 5331 | args, "O&O&:storea_s64", | |||
| 5332 | simd_arg_converter, &seq_arg, | |||
| 5333 | simd_arg_converter, &vec_arg | |||
| 5334 | )) { | |||
| 5335 | return NULL((void*)0); | |||
| 5336 | } | |||
| 5337 | npyv_storea_s64(seq_arg.data.qs64, vec_arg.data.vs64); | |||
| 5338 | // write-back | |||
| 5339 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5340 | simd_arg_free(&seq_arg); | |||
| 5341 | return NULL((void*)0); | |||
| 5342 | } | |||
| 5343 | simd_arg_free(&seq_arg); | |||
| 5344 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5345 | } | |||
| 5346 | ||||
| 5347 | #line 46 | |||
| 5348 | // special definition due to the nature of stores | |||
| 5349 | static PyObject * | |||
| 5350 | simd__intrin_stores_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5351 | { | |||
| 5352 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5353 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5354 | if (!PyArg_ParseTuple( | |||
| 5355 | args, "O&O&:stores_s64", | |||
| 5356 | simd_arg_converter, &seq_arg, | |||
| 5357 | simd_arg_converter, &vec_arg | |||
| 5358 | )) { | |||
| 5359 | return NULL((void*)0); | |||
| 5360 | } | |||
| 5361 | npyv_stores_s64(seq_arg.data.qs64, vec_arg.data.vs64); | |||
| 5362 | // write-back | |||
| 5363 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5364 | simd_arg_free(&seq_arg); | |||
| 5365 | return NULL((void*)0); | |||
| 5366 | } | |||
| 5367 | simd_arg_free(&seq_arg); | |||
| 5368 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5369 | } | |||
| 5370 | ||||
| 5371 | #line 46 | |||
| 5372 | // special definition due to the nature of storel | |||
| 5373 | static PyObject * | |||
| 5374 | simd__intrin_storel_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5375 | { | |||
| 5376 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5377 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5378 | if (!PyArg_ParseTuple( | |||
| 5379 | args, "O&O&:storel_s64", | |||
| 5380 | simd_arg_converter, &seq_arg, | |||
| 5381 | simd_arg_converter, &vec_arg | |||
| 5382 | )) { | |||
| 5383 | return NULL((void*)0); | |||
| 5384 | } | |||
| 5385 | npyv_storel_s64(seq_arg.data.qs64, vec_arg.data.vs64); | |||
| 5386 | // write-back | |||
| 5387 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5388 | simd_arg_free(&seq_arg); | |||
| 5389 | return NULL((void*)0); | |||
| 5390 | } | |||
| 5391 | simd_arg_free(&seq_arg); | |||
| 5392 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5393 | } | |||
| 5394 | ||||
| 5395 | #line 46 | |||
| 5396 | // special definition due to the nature of storeh | |||
| 5397 | static PyObject * | |||
| 5398 | simd__intrin_storeh_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5399 | { | |||
| 5400 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5401 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5402 | if (!PyArg_ParseTuple( | |||
| 5403 | args, "O&O&:storeh_s64", | |||
| 5404 | simd_arg_converter, &seq_arg, | |||
| 5405 | simd_arg_converter, &vec_arg | |||
| 5406 | )) { | |||
| 5407 | return NULL((void*)0); | |||
| 5408 | } | |||
| 5409 | npyv_storeh_s64(seq_arg.data.qs64, vec_arg.data.vs64); | |||
| 5410 | // write-back | |||
| 5411 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5412 | simd_arg_free(&seq_arg); | |||
| 5413 | return NULL((void*)0); | |||
| 5414 | } | |||
| 5415 | simd_arg_free(&seq_arg); | |||
| 5416 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5417 | } | |||
| 5418 | ||||
| 5419 | ||||
| 5420 | /**************************************** | |||
| 5421 | * Non-contiguous/Partial Memory access | |||
| 5422 | ****************************************/ | |||
| 5423 | #if 1 | |||
| 5424 | // Partial Load | |||
| 5425 | SIMD_IMPL_INTRIN_3(load_till_s64, vs64, qs64, u32, s64)static PyObject *simd__intrin_load_till_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs64}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_s64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_load_till_s64 ( arg1.data.qs64, arg2.data.u32, arg3.data.s64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5426 | SIMD_IMPL_INTRIN_2(load_tillz_s64, vs64, qs64, u32)static PyObject *simd__intrin_load_tillz_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qs64}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_s64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs64 = npyv_load_tillz_s64 ( arg1.data.qs64, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5427 | ||||
| 5428 | // Partial Store | |||
| 5429 | static PyObject * | |||
| 5430 | simd__intrin_store_till_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5431 | { | |||
| 5432 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5433 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5434 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5435 | if (!PyArg_ParseTuple( | |||
| 5436 | args, "O&O&O&:store_till_s64", | |||
| 5437 | simd_arg_converter, &seq_arg, | |||
| 5438 | simd_arg_converter, &nlane_arg, | |||
| 5439 | simd_arg_converter, &vec_arg | |||
| 5440 | )) { | |||
| 5441 | return NULL((void*)0); | |||
| 5442 | } | |||
| 5443 | npyv_store_till_s64( | |||
| 5444 | seq_arg.data.qs64, nlane_arg.data.u32, vec_arg.data.vs64 | |||
| 5445 | ); | |||
| 5446 | // write-back | |||
| 5447 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5448 | simd_arg_free(&seq_arg); | |||
| 5449 | return NULL((void*)0); | |||
| 5450 | } | |||
| 5451 | simd_arg_free(&seq_arg); | |||
| 5452 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5453 | } | |||
| 5454 | ||||
| 5455 | // Non-contiguous Load | |||
| 5456 | #line 112 | |||
| 5457 | static PyObject * | |||
| 5458 | simd__intrin_loadn_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5459 | { | |||
| 5460 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5461 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 5462 | #if 0 | |||
| 5463 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5464 | #endif // till | |||
| 5465 | #if 0 | |||
| 5466 | simd_arg fill_arg = {.dtype = simd_data_s64}; | |||
| 5467 | #endif | |||
| 5468 | if (!PyArg_ParseTuple( | |||
| 5469 | args, "O&O&:loadn_s64", | |||
| 5470 | simd_arg_converter, &seq_arg, | |||
| 5471 | simd_arg_converter, &stride_arg | |||
| 5472 | #if 0 | |||
| 5473 | ,simd_arg_converter, &nlane_arg | |||
| 5474 | #endif | |||
| 5475 | #if 0 | |||
| 5476 | ,simd_arg_converter, &fill_arg | |||
| 5477 | #endif | |||
| 5478 | )) { | |||
| 5479 | return NULL((void*)0); | |||
| 5480 | } | |||
| 5481 | npyv_lanetype_s64 *seq_ptr = seq_arg.data.qs64; | |||
| 5482 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 5483 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 5484 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s642; | |||
| 5485 | if (stride < 0) { | |||
| 5486 | seq_ptr += cur_seq_len -1; | |||
| 5487 | min_seq_len = -min_seq_len; | |||
| 5488 | } | |||
| 5489 | if (cur_seq_len < min_seq_len) { | |||
| 5490 | PyErr_Format(PyExc_ValueError, | |||
| 5491 | "loadn_s64(), according to provided stride %d, the " | |||
| 5492 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 5493 | stride, min_seq_len, cur_seq_len | |||
| 5494 | ); | |||
| 5495 | goto err; | |||
| 5496 | } | |||
| 5497 | npyv_s64 rvec = npyv_loadn_s64( | |||
| 5498 | seq_ptr, stride | |||
| 5499 | #if 0 | |||
| 5500 | , nlane_arg.data.u32 | |||
| 5501 | #endif | |||
| 5502 | #if 0 | |||
| 5503 | , fill_arg.data.s64 | |||
| 5504 | #endif | |||
| 5505 | ); | |||
| 5506 | simd_arg ret = { | |||
| 5507 | .dtype = simd_data_vs64, .data = {.vs64=rvec} | |||
| 5508 | }; | |||
| 5509 | simd_arg_free(&seq_arg); | |||
| 5510 | return simd_arg_to_obj(&ret); | |||
| 5511 | err: | |||
| 5512 | simd_arg_free(&seq_arg); | |||
| 5513 | return NULL((void*)0); | |||
| 5514 | } | |||
| 5515 | ||||
| 5516 | #line 112 | |||
| 5517 | static PyObject * | |||
| 5518 | simd__intrin_loadn_till_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5519 | { | |||
| 5520 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5521 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 5522 | #if 1 | |||
| 5523 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5524 | #endif // till | |||
| 5525 | #if 1 | |||
| 5526 | simd_arg fill_arg = {.dtype = simd_data_s64}; | |||
| 5527 | #endif | |||
| 5528 | if (!PyArg_ParseTuple( | |||
| 5529 | args, "O&O&O&O&:loadn_till_s64", | |||
| 5530 | simd_arg_converter, &seq_arg, | |||
| 5531 | simd_arg_converter, &stride_arg | |||
| 5532 | #if 1 | |||
| 5533 | ,simd_arg_converter, &nlane_arg | |||
| 5534 | #endif | |||
| 5535 | #if 1 | |||
| 5536 | ,simd_arg_converter, &fill_arg | |||
| 5537 | #endif | |||
| 5538 | )) { | |||
| 5539 | return NULL((void*)0); | |||
| 5540 | } | |||
| 5541 | npyv_lanetype_s64 *seq_ptr = seq_arg.data.qs64; | |||
| 5542 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 5543 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 5544 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s642; | |||
| 5545 | if (stride < 0) { | |||
| 5546 | seq_ptr += cur_seq_len -1; | |||
| 5547 | min_seq_len = -min_seq_len; | |||
| 5548 | } | |||
| 5549 | if (cur_seq_len < min_seq_len) { | |||
| 5550 | PyErr_Format(PyExc_ValueError, | |||
| 5551 | "loadn_till_s64(), according to provided stride %d, the " | |||
| 5552 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 5553 | stride, min_seq_len, cur_seq_len | |||
| 5554 | ); | |||
| 5555 | goto err; | |||
| 5556 | } | |||
| 5557 | npyv_s64 rvec = npyv_loadn_till_s64( | |||
| 5558 | seq_ptr, stride | |||
| 5559 | #if 1 | |||
| 5560 | , nlane_arg.data.u32 | |||
| 5561 | #endif | |||
| 5562 | #if 1 | |||
| 5563 | , fill_arg.data.s64 | |||
| 5564 | #endif | |||
| 5565 | ); | |||
| 5566 | simd_arg ret = { | |||
| 5567 | .dtype = simd_data_vs64, .data = {.vs64=rvec} | |||
| 5568 | }; | |||
| 5569 | simd_arg_free(&seq_arg); | |||
| 5570 | return simd_arg_to_obj(&ret); | |||
| 5571 | err: | |||
| 5572 | simd_arg_free(&seq_arg); | |||
| 5573 | return NULL((void*)0); | |||
| 5574 | } | |||
| 5575 | ||||
| 5576 | #line 112 | |||
| 5577 | static PyObject * | |||
| 5578 | simd__intrin_loadn_tillz_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5579 | { | |||
| 5580 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5581 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 5582 | #if 1 | |||
| 5583 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5584 | #endif // till | |||
| 5585 | #if 0 | |||
| 5586 | simd_arg fill_arg = {.dtype = simd_data_s64}; | |||
| 5587 | #endif | |||
| 5588 | if (!PyArg_ParseTuple( | |||
| 5589 | args, "O&O&O&:loadn_tillz_s64", | |||
| 5590 | simd_arg_converter, &seq_arg, | |||
| 5591 | simd_arg_converter, &stride_arg | |||
| 5592 | #if 1 | |||
| 5593 | ,simd_arg_converter, &nlane_arg | |||
| 5594 | #endif | |||
| 5595 | #if 0 | |||
| 5596 | ,simd_arg_converter, &fill_arg | |||
| 5597 | #endif | |||
| 5598 | )) { | |||
| 5599 | return NULL((void*)0); | |||
| 5600 | } | |||
| 5601 | npyv_lanetype_s64 *seq_ptr = seq_arg.data.qs64; | |||
| 5602 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 5603 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 5604 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s642; | |||
| 5605 | if (stride < 0) { | |||
| 5606 | seq_ptr += cur_seq_len -1; | |||
| 5607 | min_seq_len = -min_seq_len; | |||
| 5608 | } | |||
| 5609 | if (cur_seq_len < min_seq_len) { | |||
| 5610 | PyErr_Format(PyExc_ValueError, | |||
| 5611 | "loadn_tillz_s64(), according to provided stride %d, the " | |||
| 5612 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 5613 | stride, min_seq_len, cur_seq_len | |||
| 5614 | ); | |||
| 5615 | goto err; | |||
| 5616 | } | |||
| 5617 | npyv_s64 rvec = npyv_loadn_tillz_s64( | |||
| 5618 | seq_ptr, stride | |||
| 5619 | #if 1 | |||
| 5620 | , nlane_arg.data.u32 | |||
| 5621 | #endif | |||
| 5622 | #if 0 | |||
| 5623 | , fill_arg.data.s64 | |||
| 5624 | #endif | |||
| 5625 | ); | |||
| 5626 | simd_arg ret = { | |||
| 5627 | .dtype = simd_data_vs64, .data = {.vs64=rvec} | |||
| 5628 | }; | |||
| 5629 | simd_arg_free(&seq_arg); | |||
| 5630 | return simd_arg_to_obj(&ret); | |||
| 5631 | err: | |||
| 5632 | simd_arg_free(&seq_arg); | |||
| 5633 | return NULL((void*)0); | |||
| 5634 | } | |||
| 5635 | ||||
| 5636 | ||||
| 5637 | // Non-contiguous Store | |||
| 5638 | #line 178 | |||
| 5639 | static PyObject * | |||
| 5640 | simd__intrin_storen_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5641 | { | |||
| 5642 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5643 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 5644 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5645 | #if 0 | |||
| 5646 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5647 | #endif | |||
| 5648 | if (!PyArg_ParseTuple( | |||
| 5649 | args, "O&O&O&:storen_s64", | |||
| 5650 | simd_arg_converter, &seq_arg, | |||
| 5651 | simd_arg_converter, &stride_arg | |||
| 5652 | #if 0 | |||
| 5653 | ,simd_arg_converter, &nlane_arg | |||
| 5654 | #endif | |||
| 5655 | ,simd_arg_converter, &vec_arg | |||
| 5656 | )) { | |||
| 5657 | return NULL((void*)0); | |||
| 5658 | } | |||
| 5659 | npyv_lanetype_s64 *seq_ptr = seq_arg.data.qs64; | |||
| 5660 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 5661 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 5662 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s642; | |||
| 5663 | if (stride < 0) { | |||
| 5664 | seq_ptr += cur_seq_len -1; | |||
| 5665 | min_seq_len = -min_seq_len; | |||
| 5666 | } | |||
| 5667 | // overflow guard | |||
| 5668 | if (cur_seq_len < min_seq_len) { | |||
| 5669 | PyErr_Format(PyExc_ValueError, | |||
| 5670 | "storen_s64(), according to provided stride %d, the" | |||
| 5671 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 5672 | stride, min_seq_len, cur_seq_len | |||
| 5673 | ); | |||
| 5674 | goto err; | |||
| 5675 | } | |||
| 5676 | npyv_storen_s64( | |||
| 5677 | seq_ptr, stride | |||
| 5678 | #if 0 | |||
| 5679 | ,nlane_arg.data.u32 | |||
| 5680 | #endif | |||
| 5681 | ,vec_arg.data.vs64 | |||
| 5682 | ); | |||
| 5683 | // write-back | |||
| 5684 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5685 | goto err; | |||
| 5686 | } | |||
| 5687 | simd_arg_free(&seq_arg); | |||
| 5688 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5689 | err: | |||
| 5690 | simd_arg_free(&seq_arg); | |||
| 5691 | return NULL((void*)0); | |||
| 5692 | } | |||
| 5693 | ||||
| 5694 | #line 178 | |||
| 5695 | static PyObject * | |||
| 5696 | simd__intrin_storen_till_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5697 | { | |||
| 5698 | simd_arg seq_arg = {.dtype = simd_data_qs64}; | |||
| 5699 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 5700 | simd_arg vec_arg = {.dtype = simd_data_vs64}; | |||
| 5701 | #if 1 | |||
| 5702 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 5703 | #endif | |||
| 5704 | if (!PyArg_ParseTuple( | |||
| 5705 | args, "O&O&O&O&:storen_s64", | |||
| 5706 | simd_arg_converter, &seq_arg, | |||
| 5707 | simd_arg_converter, &stride_arg | |||
| 5708 | #if 1 | |||
| 5709 | ,simd_arg_converter, &nlane_arg | |||
| 5710 | #endif | |||
| 5711 | ,simd_arg_converter, &vec_arg | |||
| 5712 | )) { | |||
| 5713 | return NULL((void*)0); | |||
| 5714 | } | |||
| 5715 | npyv_lanetype_s64 *seq_ptr = seq_arg.data.qs64; | |||
| 5716 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 5717 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 5718 | Py_ssize_t min_seq_len = stride * npyv_nlanes_s642; | |||
| 5719 | if (stride < 0) { | |||
| 5720 | seq_ptr += cur_seq_len -1; | |||
| 5721 | min_seq_len = -min_seq_len; | |||
| 5722 | } | |||
| 5723 | // overflow guard | |||
| 5724 | if (cur_seq_len < min_seq_len) { | |||
| 5725 | PyErr_Format(PyExc_ValueError, | |||
| 5726 | "storen_till_s64(), according to provided stride %d, the" | |||
| 5727 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 5728 | stride, min_seq_len, cur_seq_len | |||
| 5729 | ); | |||
| 5730 | goto err; | |||
| 5731 | } | |||
| 5732 | npyv_storen_till_s64( | |||
| 5733 | seq_ptr, stride | |||
| 5734 | #if 1 | |||
| 5735 | ,nlane_arg.data.u32 | |||
| 5736 | #endif | |||
| 5737 | ,vec_arg.data.vs64 | |||
| 5738 | ); | |||
| 5739 | // write-back | |||
| 5740 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qs64, simd_data_qs64)) { | |||
| 5741 | goto err; | |||
| 5742 | } | |||
| 5743 | simd_arg_free(&seq_arg); | |||
| 5744 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 5745 | err: | |||
| 5746 | simd_arg_free(&seq_arg); | |||
| 5747 | return NULL((void*)0); | |||
| 5748 | } | |||
| 5749 | ||||
| 5750 | #endif // 1 | |||
| 5751 | ||||
| 5752 | /*************************** | |||
| 5753 | * Misc | |||
| 5754 | ***************************/ | |||
| 5755 | SIMD_IMPL_INTRIN_0(zero_s64, vs64)static PyObject *simd__intrin_zero_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_s64") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vs64, .data = {.vs64 = _mm_setzero_si128()} , }; return simd_arg_to_obj(&a); } | |||
| 5756 | SIMD_IMPL_INTRIN_1(setall_s64, vs64, s64)static PyObject *simd__intrin_setall_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s64}; if (!PyArg_ParseTuple( args, "O&:" "setall_s64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs64 = _mm_set1_epi64x((npy_int64)(arg .data.s64))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(& ret); } | |||
| 5757 | SIMD_IMPL_INTRIN_3(select_s64, vs64, vb64, vs64, vs64)static PyObject *simd__intrin_select_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_select_u8 ( arg1.data.vb64, arg2.data.vs64, arg3.data.vs64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5758 | ||||
| 5759 | #line 246 | |||
| 5760 | #if 1 | |||
| 5761 | SIMD_IMPL_INTRIN_1(reinterpret_u8_s64, vu8, vs64)static PyObject *simd__intrin_reinterpret_u8_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = arg.data.vs64 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj(&ret); } | |||
| 5762 | #endif // simd_sup2 | |||
| 5763 | ||||
| 5764 | #line 246 | |||
| 5765 | #if 1 | |||
| 5766 | SIMD_IMPL_INTRIN_1(reinterpret_s8_s64, vs8, vs64)static PyObject *simd__intrin_reinterpret_s8_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = arg.data.vs64 }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj(&ret); } | |||
| 5767 | #endif // simd_sup2 | |||
| 5768 | ||||
| 5769 | #line 246 | |||
| 5770 | #if 1 | |||
| 5771 | SIMD_IMPL_INTRIN_1(reinterpret_u16_s64, vu16, vs64)static PyObject *simd__intrin_reinterpret_u16_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu16 }; return simd_arg_to_obj(&ret) ; } | |||
| 5772 | #endif // simd_sup2 | |||
| 5773 | ||||
| 5774 | #line 246 | |||
| 5775 | #if 1 | |||
| 5776 | SIMD_IMPL_INTRIN_1(reinterpret_s16_s64, vs16, vs64)static PyObject *simd__intrin_reinterpret_s16_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs16 }; return simd_arg_to_obj(&ret) ; } | |||
| 5777 | #endif // simd_sup2 | |||
| 5778 | ||||
| 5779 | #line 246 | |||
| 5780 | #if 1 | |||
| 5781 | SIMD_IMPL_INTRIN_1(reinterpret_u32_s64, vu32, vs64)static PyObject *simd__intrin_reinterpret_u32_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu32 }; return simd_arg_to_obj(&ret) ; } | |||
| 5782 | #endif // simd_sup2 | |||
| 5783 | ||||
| 5784 | #line 246 | |||
| 5785 | #if 1 | |||
| 5786 | SIMD_IMPL_INTRIN_1(reinterpret_s32_s64, vs32, vs64)static PyObject *simd__intrin_reinterpret_s32_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret) ; } | |||
| 5787 | #endif // simd_sup2 | |||
| 5788 | ||||
| 5789 | #line 246 | |||
| 5790 | #if 1 | |||
| 5791 | SIMD_IMPL_INTRIN_1(reinterpret_u64_s64, vu64, vs64)static PyObject *simd__intrin_reinterpret_u64_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vu64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5792 | #endif // simd_sup2 | |||
| 5793 | ||||
| 5794 | #line 246 | |||
| 5795 | #if 1 | |||
| 5796 | SIMD_IMPL_INTRIN_1(reinterpret_s64_s64, vs64, vs64)static PyObject *simd__intrin_reinterpret_s64_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = arg.data. vs64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5797 | #endif // simd_sup2 | |||
| 5798 | ||||
| 5799 | #line 246 | |||
| 5800 | #if 1 | |||
| 5801 | SIMD_IMPL_INTRIN_1(reinterpret_f32_s64, vf32, vs64)static PyObject *simd__intrin_reinterpret_f32_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps ( arg.data.vs64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 5802 | #endif // simd_sup2 | |||
| 5803 | ||||
| 5804 | #line 246 | |||
| 5805 | #if NPY_SIMD_F641 | |||
| 5806 | SIMD_IMPL_INTRIN_1(reinterpret_f64_s64, vf64, vs64)static PyObject *simd__intrin_reinterpret_f64_s64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_s64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castsi128_pd ( arg.data.vs64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 5807 | #endif // simd_sup2 | |||
| 5808 | ||||
| 5809 | ||||
| 5810 | /** | |||
| 5811 | * special definition due to the nature of intrinsics | |||
| 5812 | * npyv_setf_s64 and npy_set_s64. | |||
| 5813 | */ | |||
| 5814 | #line 258 | |||
| 5815 | static PyObject * | |||
| 5816 | simd__intrin_setf_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5817 | { | |||
| 5818 | npyv_lanetype_s64 *data = simd_sequence_from_iterable(args, simd_data_qs64, npyv_nlanes_s642); | |||
| 5819 | if (data == NULL((void*)0)) { | |||
| 5820 | return NULL((void*)0); | |||
| 5821 | } | |||
| 5822 | simd_data r = {.vs64 = npyv_setf_s64(npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5823 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5824 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5825 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5826 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5827 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5828 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5829 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5830 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5831 | data[64] // for setfnpyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2])) | |||
| 5832 | )npyv__setr_epi64((npy_int64)(data[1]), (npy_int64)(data[2]))}; | |||
| 5833 | simd_sequence_free(data); | |||
| 5834 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs64); | |||
| 5835 | } | |||
| 5836 | ||||
| 5837 | #line 258 | |||
| 5838 | static PyObject * | |||
| 5839 | simd__intrin_set_s64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 5840 | { | |||
| 5841 | npyv_lanetype_s64 *data = simd_sequence_from_iterable(args, simd_data_qs64, npyv_nlanes_s642); | |||
| 5842 | if (data == NULL((void*)0)) { | |||
| 5843 | return NULL((void*)0); | |||
| 5844 | } | |||
| 5845 | simd_data r = {.vs64 = npyv_set_s64(npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5846 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5847 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5848 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5849 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5850 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5851 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5852 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5853 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5854 | data[64] // for setfnpyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1])) | |||
| 5855 | )npyv__setr_epi64((npy_int64)(data[0]), (npy_int64)(data[1]))}; | |||
| 5856 | simd_sequence_free(data); | |||
| 5857 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vs64); | |||
| 5858 | } | |||
| 5859 | ||||
| 5860 | ||||
| 5861 | /*************************** | |||
| 5862 | * Reorder | |||
| 5863 | ***************************/ | |||
| 5864 | #line 287 | |||
| 5865 | SIMD_IMPL_INTRIN_2(combinel_s64, vs64, vs64, vs64)static PyObject *simd__intrin_combinel_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_s64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs64 = _mm_unpacklo_epi64 ( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5866 | ||||
| 5867 | #line 287 | |||
| 5868 | SIMD_IMPL_INTRIN_2(combineh_s64, vs64, vs64, vs64)static PyObject *simd__intrin_combineh_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_s64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs64 = _mm_unpackhi_epi64 ( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5869 | ||||
| 5870 | ||||
| 5871 | #line 293 | |||
| 5872 | SIMD_IMPL_INTRIN_2(combine_s64, vs64x2, vs64, vs64)static PyObject *simd__intrin_combine_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_s64", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs64x2 = npyv__combine ( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64x2 }; return simd_arg_to_obj(&ret); } | |||
| 5873 | ||||
| 5874 | #line 293 | |||
| 5875 | SIMD_IMPL_INTRIN_2(zip_s64, vs64x2, vs64, vs64)static PyObject *simd__intrin_zip_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64x2 = npyv_zip_s64( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64x2 }; return simd_arg_to_obj(&ret); } | |||
| 5876 | ||||
| 5877 | ||||
| 5878 | #if 0 | |||
| 5879 | SIMD_IMPL_INTRIN_1(rev64_s64, vs64, vs64)static PyObject *simd__intrin_rev64_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "rev64_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs64 = npyv_rev64_s64( arg.data.vs64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5880 | #endif | |||
| 5881 | ||||
| 5882 | /*************************** | |||
| 5883 | * Operators | |||
| 5884 | ***************************/ | |||
| 5885 | #if 63 > 0 | |||
| 5886 | SIMD_IMPL_INTRIN_2(shl_s64, vs64, vs64, u8)static PyObject *simd__intrin_shl_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_sll_epi64(arg1.data.vs64, _mm_cvtsi32_si128 (arg2.data.u8))}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5887 | SIMD_IMPL_INTRIN_2(shr_s64, vs64, vs64, u8)static PyObject *simd__intrin_shr_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_shr_s64( arg1.data.vs64, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5888 | // immediate constant | |||
| 5889 | SIMD_IMPL_INTRIN_2IMM(shli_s64, vs64, vs64, 63)static PyObject *simd__intrin_shli_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs64 = 0 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 0) : 1 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 1) : 2 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 2) : 3 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 3) : 4 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 4) : 5 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 5) : 6 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 6) : 7 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 7) : 8 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vs64, 8) : 9 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vs64, 9) : 10 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 10) : 11 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 11) : 12 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 12) : 13 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 13) : 14 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 14) : 15 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 15 ) : 16 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 16) : 17 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 17) : 18 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 18) : 19 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 19) : 20 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vs64, 20) : 21 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vs64, 21) : 22 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 22) : 23 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 23) : 24 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 24) : 25 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 25) : 26 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 26) : 27 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 27 ) : 28 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 28) : 29 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 29) : 30 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 30) : 31 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 31) : 32 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vs64, 32) : 33 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vs64, 33) : 34 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 34) : 35 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 35) : 36 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 36) : 37 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 37) : 38 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 38) : 39 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 39 ) : 40 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 40) : 41 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 41) : 42 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 42) : 43 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 43) : 44 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vs64, 44) : 45 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vs64, 45) : 46 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 46) : 47 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 47) : 48 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 48) : 49 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 49) : 50 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 50) : 51 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 51 ) : 52 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 52) : 53 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 53) : 54 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 54) : 55 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 55) : 56 == arg2 .data.u8 ? _mm_slli_epi64(arg1.data.vs64, 56) : 57 == arg2.data .u8 ? _mm_slli_epi64(arg1.data.vs64, 57) : 58 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 58) : 59 == arg2.data.u8 ? _mm_slli_epi64 (arg1.data.vs64, 59) : 60 == arg2.data.u8 ? _mm_slli_epi64(arg1 .data.vs64, 60) : 61 == arg2.data.u8 ? _mm_slli_epi64(arg1.data .vs64, 61) : 62 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64 , 62) : 63 == arg2.data.u8 ? _mm_slli_epi64(arg1.data.vs64, 63 ) : data.vs64; simd_arg_free(&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(& ret); } | |||
| 5890 | SIMD_IMPL_INTRIN_2IMM(shri_s64, vs64, vs64, 64)static PyObject *simd__intrin_shri_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vs64 = 1 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 1) : 2 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 2) : 3 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 3) : 4 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 4 ) : 5 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 5) : 6 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 6) : 7 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 7) : 8 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 8) : 9 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 9) : 10 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 10) : 11 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 11) : 12 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 12) : 13 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 13) : 14 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 14) : 15 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 15) : 16 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 16) : 17 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 17) : 18 == arg2.data.u8 ? npyv_shr_s64(arg1 .data.vs64, 18) : 19 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 19) : 20 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64 , 20) : 21 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 21) : 22 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 22) : 23 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 23) : 24 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 24) : 25 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 25) : 26 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 26) : 27 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 27) : 28 == arg2.data.u8 ? npyv_shr_s64(arg1 .data.vs64, 28) : 29 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 29) : 30 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64 , 30) : 31 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 31) : 32 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 32) : 33 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 33) : 34 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 34) : 35 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 35) : 36 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 36) : 37 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 37) : 38 == arg2.data.u8 ? npyv_shr_s64(arg1 .data.vs64, 38) : 39 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 39) : 40 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64 , 40) : 41 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 41) : 42 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 42) : 43 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 43) : 44 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 44) : 45 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 45) : 46 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 46) : 47 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 47) : 48 == arg2.data.u8 ? npyv_shr_s64(arg1 .data.vs64, 48) : 49 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 49) : 50 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64 , 50) : 51 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 51) : 52 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 52) : 53 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 53) : 54 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 54) : 55 == arg2.data .u8 ? npyv_shr_s64(arg1.data.vs64, 55) : 56 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 56) : 57 == arg2.data.u8 ? npyv_shr_s64 (arg1.data.vs64, 57) : 58 == arg2.data.u8 ? npyv_shr_s64(arg1 .data.vs64, 58) : 59 == arg2.data.u8 ? npyv_shr_s64(arg1.data .vs64, 59) : 60 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64 , 60) : 61 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 61) : 62 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 62) : 63 == arg2.data.u8 ? npyv_shr_s64(arg1.data.vs64, 63) : 64 == arg2 .data.u8 ? npyv_shr_s64(arg1.data.vs64, 64) : data.vs64; simd_arg_free (&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5891 | #endif // shl_imm | |||
| 5892 | ||||
| 5893 | #line 314 | |||
| 5894 | SIMD_IMPL_INTRIN_2(and_s64, vs64, vs64, vs64)static PyObject *simd__intrin_and_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""and_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_and_si128( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5895 | ||||
| 5896 | #line 314 | |||
| 5897 | SIMD_IMPL_INTRIN_2(or_s64, vs64, vs64, vs64)static PyObject *simd__intrin_or_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""or_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_or_si128( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5898 | ||||
| 5899 | #line 314 | |||
| 5900 | SIMD_IMPL_INTRIN_2(xor_s64, vs64, vs64, vs64)static PyObject *simd__intrin_xor_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_xor_si128( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5901 | ||||
| 5902 | ||||
| 5903 | SIMD_IMPL_INTRIN_1(not_s64, vs64, vs64)static PyObject *simd__intrin_not_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "not_s64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs64 = _mm_xor_si128(arg.data.vs64, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret) ; } | |||
| 5904 | ||||
| 5905 | #line 322 | |||
| 5906 | SIMD_IMPL_INTRIN_2(cmpeq_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmpeq_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpeq_u64( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5907 | ||||
| 5908 | #line 322 | |||
| 5909 | SIMD_IMPL_INTRIN_2(cmpneq_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmpneq_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpeq_u64(arg1 .data.vs64, arg2.data.vs64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5910 | ||||
| 5911 | #line 322 | |||
| 5912 | SIMD_IMPL_INTRIN_2(cmpgt_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmpgt_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpgt_s64( arg1.data.vs64, arg2.data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5913 | ||||
| 5914 | #line 322 | |||
| 5915 | SIMD_IMPL_INTRIN_2(cmpge_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmpge_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpgt_s64(arg2 .data.vs64, arg1.data.vs64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5916 | ||||
| 5917 | #line 322 | |||
| 5918 | SIMD_IMPL_INTRIN_2(cmplt_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmplt_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = npyv_cmpgt_s64(arg2.data.vs64, arg1 .data.vs64)}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5919 | ||||
| 5920 | #line 322 | |||
| 5921 | SIMD_IMPL_INTRIN_2(cmple_s64, vb64, vs64, vs64)static PyObject *simd__intrin_cmple_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128(npyv_cmpgt_s64(arg1 .data.vs64, arg2.data.vs64), _mm_set1_epi32(-1))}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(& ret); } | |||
| 5922 | ||||
| 5923 | ||||
| 5924 | /*************************** | |||
| 5925 | * Conversion | |||
| 5926 | ***************************/ | |||
| 5927 | SIMD_IMPL_INTRIN_1(cvt_s64_b64, vs64, vb64)static PyObject *simd__intrin_cvt_s64_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_s64_b64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs64 = arg.data.vb64}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5928 | SIMD_IMPL_INTRIN_1(cvt_b64_s64, vb64, vs64)static PyObject *simd__intrin_cvt_b64_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b64_s64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb64 = arg.data.vs64}; simd_arg_free( &arg); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 5929 | #if 0 | |||
| 5930 | SIMD_IMPL_INTRIN_1(expand_s64_s64, vs64x2, vs64)static PyObject *simd__intrin_expand_s64_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "expand_s64_s64", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vs64x2 = npyv_expand_s64_s64( arg.data .vs64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vs64x2 }; return simd_arg_to_obj(&ret ); } | |||
| 5931 | #endif // expand_sup | |||
| 5932 | /*************************** | |||
| 5933 | * Arithmetic | |||
| 5934 | ***************************/ | |||
| 5935 | #line 339 | |||
| 5936 | SIMD_IMPL_INTRIN_2(add_s64, vs64, vs64, vs64)static PyObject *simd__intrin_add_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""add_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_add_epi64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5937 | ||||
| 5938 | #line 339 | |||
| 5939 | SIMD_IMPL_INTRIN_2(sub_s64, vs64, vs64, vs64)static PyObject *simd__intrin_sub_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = _mm_sub_epi64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5940 | ||||
| 5941 | ||||
| 5942 | #if 0 | |||
| 5943 | #line 346 | |||
| 5944 | SIMD_IMPL_INTRIN_2(adds_s64, vs64, vs64, vs64)static PyObject *simd__intrin_adds_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_adds_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5945 | ||||
| 5946 | #line 346 | |||
| 5947 | SIMD_IMPL_INTRIN_2(subs_s64, vs64, vs64, vs64)static PyObject *simd__intrin_subs_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_subs_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5948 | ||||
| 5949 | #endif // sat_sup | |||
| 5950 | ||||
| 5951 | #if 0 | |||
| 5952 | SIMD_IMPL_INTRIN_2(mul_s64, vs64, vs64, vs64)static PyObject *simd__intrin_mul_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_mul_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5953 | #endif // mul_sup | |||
| 5954 | ||||
| 5955 | #if 0 | |||
| 5956 | SIMD_IMPL_INTRIN_2(div_s64, vs64, vs64, vs64)static PyObject *simd__intrin_div_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""div_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_div_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5957 | #endif // div_sup | |||
| 5958 | ||||
| 5959 | #if 1 | |||
| 5960 | SIMD_IMPL_INTRIN_1(divisor_s64, vs64x3, s64)static PyObject *simd__intrin_divisor_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_s64}; if (!PyArg_ParseTuple( args, "O&:" "divisor_s64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vs64x3 = npyv_divisor_s64( arg.data.s64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs64x3 }; return simd_arg_to_obj(&ret); } | |||
| 5961 | SIMD_IMPL_INTRIN_2(divc_s64, vs64, vs64, vs64x3)static PyObject *simd__intrin_divc_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_divc_s64( arg1.data.vs64, arg2 .data.vs64x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5962 | #endif // intdiv_sup | |||
| 5963 | ||||
| 5964 | #if 0 | |||
| 5965 | #line 367 | |||
| 5966 | SIMD_IMPL_INTRIN_3(muladd_s64, vs64, vs64, vs64, vs64)static PyObject *simd__intrin_muladd_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_muladd_s64 ( arg1.data.vs64, arg2.data.vs64, arg3.data.vs64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5967 | ||||
| 5968 | #line 367 | |||
| 5969 | SIMD_IMPL_INTRIN_3(mulsub_s64, vs64, vs64, vs64, vs64)static PyObject *simd__intrin_mulsub_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_mulsub_s64 ( arg1.data.vs64, arg2.data.vs64, arg3.data.vs64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5970 | ||||
| 5971 | #line 367 | |||
| 5972 | SIMD_IMPL_INTRIN_3(nmuladd_s64, vs64, vs64, vs64, vs64)static PyObject *simd__intrin_nmuladd_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_nmuladd_s64 ( arg1.data.vs64, arg2.data.vs64, arg3.data.vs64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5973 | ||||
| 5974 | #line 367 | |||
| 5975 | SIMD_IMPL_INTRIN_3(nmulsub_s64, vs64, vs64, vs64, vs64)static PyObject *simd__intrin_nmulsub_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vs64 = npyv_nmulsub_s64 ( arg1.data.vs64, arg2.data.vs64, arg3.data.vs64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5976 | ||||
| 5977 | #endif // fused_sup | |||
| 5978 | ||||
| 5979 | #if 0 | |||
| 5980 | SIMD_IMPL_INTRIN_1(sum_s64, s64, vs64)static PyObject *simd__intrin_sum_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "sum_s64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.s64 = npyv_sum_s64( arg.data.vs64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_s64 }; return simd_arg_to_obj(&ret); } | |||
| 5981 | #endif // sum_sup | |||
| 5982 | ||||
| 5983 | #if 0 | |||
| 5984 | SIMD_IMPL_INTRIN_1(sumup_s64, s64, vs64)static PyObject *simd__intrin_sumup_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "sumup_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.s64 = npyv_sumup_s64( arg.data.vs64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_s64 }; return simd_arg_to_obj(&ret); } | |||
| 5985 | #endif // sumup_sup | |||
| 5986 | ||||
| 5987 | /*************************** | |||
| 5988 | * Math | |||
| 5989 | ***************************/ | |||
| 5990 | #if 0 | |||
| 5991 | #line 386 | |||
| 5992 | SIMD_IMPL_INTRIN_1(sqrt_s64, vs64, vs64)static PyObject *simd__intrin_sqrt_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_s64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vs64 = npyv_sqrt_s64( arg.data.vs64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5993 | ||||
| 5994 | #line 386 | |||
| 5995 | SIMD_IMPL_INTRIN_1(recip_s64, vs64, vs64)static PyObject *simd__intrin_recip_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "recip_s64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vs64 = npyv_recip_s64( arg.data.vs64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5996 | ||||
| 5997 | #line 386 | |||
| 5998 | SIMD_IMPL_INTRIN_1(abs_s64, vs64, vs64)static PyObject *simd__intrin_abs_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "abs_s64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vs64 = npyv_abs_s64( arg.data.vs64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 5999 | ||||
| 6000 | #line 386 | |||
| 6001 | SIMD_IMPL_INTRIN_1(square_s64, vs64, vs64)static PyObject *simd__intrin_square_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&:" "square_s64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vs64 = npyv_square_s64( arg.data.vs64 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 6002 | ||||
| 6003 | #endif | |||
| 6004 | ||||
| 6005 | #line 393 | |||
| 6006 | SIMD_IMPL_INTRIN_2(max_s64, vs64, vs64, vs64)static PyObject *simd__intrin_max_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""max_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_max_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 6007 | ||||
| 6008 | #line 393 | |||
| 6009 | SIMD_IMPL_INTRIN_2(min_s64, vs64, vs64, vs64)static PyObject *simd__intrin_min_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""min_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_min_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 6010 | ||||
| 6011 | ||||
| 6012 | #if 0 | |||
| 6013 | #line 400 | |||
| 6014 | SIMD_IMPL_INTRIN_2(maxp_s64, vs64, vs64, vs64)static PyObject *simd__intrin_maxp_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_maxp_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 6015 | ||||
| 6016 | #line 400 | |||
| 6017 | SIMD_IMPL_INTRIN_2(minp_s64, vs64, vs64, vs64)static PyObject *simd__intrin_minp_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vs64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_s64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vs64 = npyv_minp_s64( arg1.data.vs64, arg2 .data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(&ret); } | |||
| 6018 | ||||
| 6019 | #endif | |||
| 6020 | ||||
| 6021 | /*************************** | |||
| 6022 | * Mask operations | |||
| 6023 | ***************************/ | |||
| 6024 | #line 410 | |||
| 6025 | SIMD_IMPL_INTRIN_4(ifadd_s64, vs64, vb64, vs64, vs64, vs64)static PyObject *simd__intrin_ifadd_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; simd_arg arg4 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_s64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs64 = npyv_ifadd_s64 ( arg1.data.vb64, arg2.data.vs64, arg3.data.vs64, arg4.data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(& ret); } | |||
| 6026 | ||||
| 6027 | #line 410 | |||
| 6028 | SIMD_IMPL_INTRIN_4(ifsub_s64, vs64, vb64, vs64, vs64, vs64)static PyObject *simd__intrin_ifsub_s64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vs64 }; simd_arg arg3 = {.dtype = simd_data_vs64}; simd_arg arg4 = {.dtype = simd_data_vs64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_s64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vs64 = npyv_ifsub_s64 ( arg1.data.vb64, arg2.data.vs64, arg3.data.vs64, arg4.data.vs64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj(& ret); } | |||
| 6029 | ||||
| 6030 | ||||
| 6031 | #endif // simd_sup | |||
| 6032 | ||||
| 6033 | #line 34 | |||
| 6034 | #if 1 | |||
| 6035 | /*************************** | |||
| 6036 | * Memory | |||
| 6037 | ***************************/ | |||
| 6038 | #line 41 | |||
| 6039 | SIMD_IMPL_INTRIN_1(load_f32, vf32, qf32)static PyObject *simd__intrin_load_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf32}; if (!PyArg_ParseTuple( args, "O&:" "load_f32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vf32 = _mm_loadu_ps( arg.data.qf32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6040 | ||||
| 6041 | #line 41 | |||
| 6042 | SIMD_IMPL_INTRIN_1(loada_f32, vf32, qf32)static PyObject *simd__intrin_loada_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf32}; if (!PyArg_ParseTuple( args, "O&:" "loada_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf32 = _mm_load_ps( arg.data.qf32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6043 | ||||
| 6044 | #line 41 | |||
| 6045 | SIMD_IMPL_INTRIN_1(loads_f32, vf32, qf32)static PyObject *simd__intrin_loads_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf32}; if (!PyArg_ParseTuple( args, "O&:" "loads_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf32 = _mm_castsi128_ps(_mm_load_si128( (const __m128i *)(arg.data.qf32)))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6046 | ||||
| 6047 | #line 41 | |||
| 6048 | SIMD_IMPL_INTRIN_1(loadl_f32, vf32, qf32)static PyObject *simd__intrin_loadl_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf32}; if (!PyArg_ParseTuple( args, "O&:" "loadl_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf32 = _mm_castsi128_ps(npyv_loadl_u32( (const npy_uint32*)(arg.data.qf32)))}; simd_arg_free(&arg ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6049 | ||||
| 6050 | #line 46 | |||
| 6051 | // special definition due to the nature of store | |||
| 6052 | static PyObject * | |||
| 6053 | simd__intrin_store_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6054 | { | |||
| 6055 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6056 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6057 | if (!PyArg_ParseTuple( | |||
| 6058 | args, "O&O&:store_f32", | |||
| 6059 | simd_arg_converter, &seq_arg, | |||
| 6060 | simd_arg_converter, &vec_arg | |||
| 6061 | )) { | |||
| 6062 | return NULL((void*)0); | |||
| 6063 | } | |||
| 6064 | npyv_store_f32_mm_storeu_ps(seq_arg.data.qf32, vec_arg.data.vf32); | |||
| 6065 | // write-back | |||
| 6066 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6067 | simd_arg_free(&seq_arg); | |||
| 6068 | return NULL((void*)0); | |||
| 6069 | } | |||
| 6070 | simd_arg_free(&seq_arg); | |||
| 6071 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6072 | } | |||
| 6073 | ||||
| 6074 | #line 46 | |||
| 6075 | // special definition due to the nature of storea | |||
| 6076 | static PyObject * | |||
| 6077 | simd__intrin_storea_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6078 | { | |||
| 6079 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6080 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6081 | if (!PyArg_ParseTuple( | |||
| 6082 | args, "O&O&:storea_f32", | |||
| 6083 | simd_arg_converter, &seq_arg, | |||
| 6084 | simd_arg_converter, &vec_arg | |||
| 6085 | )) { | |||
| 6086 | return NULL((void*)0); | |||
| 6087 | } | |||
| 6088 | npyv_storea_f32_mm_store_ps(seq_arg.data.qf32, vec_arg.data.vf32); | |||
| 6089 | // write-back | |||
| 6090 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6091 | simd_arg_free(&seq_arg); | |||
| 6092 | return NULL((void*)0); | |||
| 6093 | } | |||
| 6094 | simd_arg_free(&seq_arg); | |||
| 6095 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6096 | } | |||
| 6097 | ||||
| 6098 | #line 46 | |||
| 6099 | // special definition due to the nature of stores | |||
| 6100 | static PyObject * | |||
| 6101 | simd__intrin_stores_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6102 | { | |||
| 6103 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6104 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6105 | if (!PyArg_ParseTuple( | |||
| 6106 | args, "O&O&:stores_f32", | |||
| 6107 | simd_arg_converter, &seq_arg, | |||
| 6108 | simd_arg_converter, &vec_arg | |||
| 6109 | )) { | |||
| 6110 | return NULL((void*)0); | |||
| 6111 | } | |||
| 6112 | npyv_stores_f32_mm_stream_ps(seq_arg.data.qf32, vec_arg.data.vf32); | |||
| 6113 | // write-back | |||
| 6114 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6115 | simd_arg_free(&seq_arg); | |||
| 6116 | return NULL((void*)0); | |||
| 6117 | } | |||
| 6118 | simd_arg_free(&seq_arg); | |||
| 6119 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6120 | } | |||
| 6121 | ||||
| 6122 | #line 46 | |||
| 6123 | // special definition due to the nature of storel | |||
| 6124 | static PyObject * | |||
| 6125 | simd__intrin_storel_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6126 | { | |||
| 6127 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6128 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6129 | if (!PyArg_ParseTuple( | |||
| 6130 | args, "O&O&:storel_f32", | |||
| 6131 | simd_arg_converter, &seq_arg, | |||
| 6132 | simd_arg_converter, &vec_arg | |||
| 6133 | )) { | |||
| 6134 | return NULL((void*)0); | |||
| 6135 | } | |||
| 6136 | npyv_storel_f32(seq_arg.data.qf32, vec_arg.data.vf32)_mm_storel_epi64((__m128i*)(seq_arg.data.qf32), _mm_castps_si128 (vec_arg.data.vf32));; | |||
| 6137 | // write-back | |||
| 6138 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6139 | simd_arg_free(&seq_arg); | |||
| 6140 | return NULL((void*)0); | |||
| 6141 | } | |||
| 6142 | simd_arg_free(&seq_arg); | |||
| 6143 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6144 | } | |||
| 6145 | ||||
| 6146 | #line 46 | |||
| 6147 | // special definition due to the nature of storeh | |||
| 6148 | static PyObject * | |||
| 6149 | simd__intrin_storeh_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6150 | { | |||
| 6151 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6152 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6153 | if (!PyArg_ParseTuple( | |||
| 6154 | args, "O&O&:storeh_f32", | |||
| 6155 | simd_arg_converter, &seq_arg, | |||
| 6156 | simd_arg_converter, &vec_arg | |||
| 6157 | )) { | |||
| 6158 | return NULL((void*)0); | |||
| 6159 | } | |||
| 6160 | npyv_storeh_f32(seq_arg.data.qf32, vec_arg.data.vf32)npyv_storeh_u32((npy_uint32*)(seq_arg.data.qf32), _mm_castps_si128 (vec_arg.data.vf32)); | |||
| 6161 | // write-back | |||
| 6162 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6163 | simd_arg_free(&seq_arg); | |||
| 6164 | return NULL((void*)0); | |||
| 6165 | } | |||
| 6166 | simd_arg_free(&seq_arg); | |||
| 6167 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6168 | } | |||
| 6169 | ||||
| 6170 | ||||
| 6171 | /**************************************** | |||
| 6172 | * Non-contiguous/Partial Memory access | |||
| 6173 | ****************************************/ | |||
| 6174 | #if 1 | |||
| 6175 | // Partial Load | |||
| 6176 | SIMD_IMPL_INTRIN_3(load_till_f32, vf32, qf32, u32, f32)static PyObject *simd__intrin_load_till_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qf32}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_f32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_load_till_f32 ( arg1.data.qf32, arg2.data.u32, arg3.data.f32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6177 | SIMD_IMPL_INTRIN_2(load_tillz_f32, vf32, qf32, u32)static PyObject *simd__intrin_load_tillz_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qf32}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_f32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf32 = npyv_load_tillz_f32 ( arg1.data.qf32, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6178 | ||||
| 6179 | // Partial Store | |||
| 6180 | static PyObject * | |||
| 6181 | simd__intrin_store_till_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6182 | { | |||
| 6183 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6184 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6185 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6186 | if (!PyArg_ParseTuple( | |||
| 6187 | args, "O&O&O&:store_till_f32", | |||
| 6188 | simd_arg_converter, &seq_arg, | |||
| 6189 | simd_arg_converter, &nlane_arg, | |||
| 6190 | simd_arg_converter, &vec_arg | |||
| 6191 | )) { | |||
| 6192 | return NULL((void*)0); | |||
| 6193 | } | |||
| 6194 | npyv_store_till_f32( | |||
| 6195 | seq_arg.data.qf32, nlane_arg.data.u32, vec_arg.data.vf32 | |||
| 6196 | ); | |||
| 6197 | // write-back | |||
| 6198 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6199 | simd_arg_free(&seq_arg); | |||
| 6200 | return NULL((void*)0); | |||
| 6201 | } | |||
| 6202 | simd_arg_free(&seq_arg); | |||
| 6203 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6204 | } | |||
| 6205 | ||||
| 6206 | // Non-contiguous Load | |||
| 6207 | #line 112 | |||
| 6208 | static PyObject * | |||
| 6209 | simd__intrin_loadn_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6210 | { | |||
| 6211 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6212 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6213 | #if 0 | |||
| 6214 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6215 | #endif // till | |||
| 6216 | #if 0 | |||
| 6217 | simd_arg fill_arg = {.dtype = simd_data_f32}; | |||
| 6218 | #endif | |||
| 6219 | if (!PyArg_ParseTuple( | |||
| 6220 | args, "O&O&:loadn_f32", | |||
| 6221 | simd_arg_converter, &seq_arg, | |||
| 6222 | simd_arg_converter, &stride_arg | |||
| 6223 | #if 0 | |||
| 6224 | ,simd_arg_converter, &nlane_arg | |||
| 6225 | #endif | |||
| 6226 | #if 0 | |||
| 6227 | ,simd_arg_converter, &fill_arg | |||
| 6228 | #endif | |||
| 6229 | )) { | |||
| 6230 | return NULL((void*)0); | |||
| 6231 | } | |||
| 6232 | npyv_lanetype_f32 *seq_ptr = seq_arg.data.qf32; | |||
| 6233 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6234 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6235 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f324; | |||
| 6236 | if (stride < 0) { | |||
| 6237 | seq_ptr += cur_seq_len -1; | |||
| 6238 | min_seq_len = -min_seq_len; | |||
| 6239 | } | |||
| 6240 | if (cur_seq_len < min_seq_len) { | |||
| 6241 | PyErr_Format(PyExc_ValueError, | |||
| 6242 | "loadn_f32(), according to provided stride %d, the " | |||
| 6243 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6244 | stride, min_seq_len, cur_seq_len | |||
| 6245 | ); | |||
| 6246 | goto err; | |||
| 6247 | } | |||
| 6248 | npyv_f32 rvec = npyv_loadn_f32( | |||
| 6249 | seq_ptr, stride | |||
| 6250 | #if 0 | |||
| 6251 | , nlane_arg.data.u32 | |||
| 6252 | #endif | |||
| 6253 | #if 0 | |||
| 6254 | , fill_arg.data.f32 | |||
| 6255 | #endif | |||
| 6256 | ); | |||
| 6257 | simd_arg ret = { | |||
| 6258 | .dtype = simd_data_vf32, .data = {.vf32=rvec} | |||
| 6259 | }; | |||
| 6260 | simd_arg_free(&seq_arg); | |||
| 6261 | return simd_arg_to_obj(&ret); | |||
| 6262 | err: | |||
| 6263 | simd_arg_free(&seq_arg); | |||
| 6264 | return NULL((void*)0); | |||
| 6265 | } | |||
| 6266 | ||||
| 6267 | #line 112 | |||
| 6268 | static PyObject * | |||
| 6269 | simd__intrin_loadn_till_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6270 | { | |||
| 6271 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6272 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6273 | #if 1 | |||
| 6274 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6275 | #endif // till | |||
| 6276 | #if 1 | |||
| 6277 | simd_arg fill_arg = {.dtype = simd_data_f32}; | |||
| 6278 | #endif | |||
| 6279 | if (!PyArg_ParseTuple( | |||
| 6280 | args, "O&O&O&O&:loadn_till_f32", | |||
| 6281 | simd_arg_converter, &seq_arg, | |||
| 6282 | simd_arg_converter, &stride_arg | |||
| 6283 | #if 1 | |||
| 6284 | ,simd_arg_converter, &nlane_arg | |||
| 6285 | #endif | |||
| 6286 | #if 1 | |||
| 6287 | ,simd_arg_converter, &fill_arg | |||
| 6288 | #endif | |||
| 6289 | )) { | |||
| 6290 | return NULL((void*)0); | |||
| 6291 | } | |||
| 6292 | npyv_lanetype_f32 *seq_ptr = seq_arg.data.qf32; | |||
| 6293 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6294 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6295 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f324; | |||
| 6296 | if (stride < 0) { | |||
| 6297 | seq_ptr += cur_seq_len -1; | |||
| 6298 | min_seq_len = -min_seq_len; | |||
| 6299 | } | |||
| 6300 | if (cur_seq_len < min_seq_len) { | |||
| 6301 | PyErr_Format(PyExc_ValueError, | |||
| 6302 | "loadn_till_f32(), according to provided stride %d, the " | |||
| 6303 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6304 | stride, min_seq_len, cur_seq_len | |||
| 6305 | ); | |||
| 6306 | goto err; | |||
| 6307 | } | |||
| 6308 | npyv_f32 rvec = npyv_loadn_till_f32( | |||
| 6309 | seq_ptr, stride | |||
| 6310 | #if 1 | |||
| 6311 | , nlane_arg.data.u32 | |||
| 6312 | #endif | |||
| 6313 | #if 1 | |||
| 6314 | , fill_arg.data.f32 | |||
| 6315 | #endif | |||
| 6316 | ); | |||
| 6317 | simd_arg ret = { | |||
| 6318 | .dtype = simd_data_vf32, .data = {.vf32=rvec} | |||
| 6319 | }; | |||
| 6320 | simd_arg_free(&seq_arg); | |||
| 6321 | return simd_arg_to_obj(&ret); | |||
| 6322 | err: | |||
| 6323 | simd_arg_free(&seq_arg); | |||
| 6324 | return NULL((void*)0); | |||
| 6325 | } | |||
| 6326 | ||||
| 6327 | #line 112 | |||
| 6328 | static PyObject * | |||
| 6329 | simd__intrin_loadn_tillz_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6330 | { | |||
| 6331 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6332 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6333 | #if 1 | |||
| 6334 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6335 | #endif // till | |||
| 6336 | #if 0 | |||
| 6337 | simd_arg fill_arg = {.dtype = simd_data_f32}; | |||
| 6338 | #endif | |||
| 6339 | if (!PyArg_ParseTuple( | |||
| 6340 | args, "O&O&O&:loadn_tillz_f32", | |||
| 6341 | simd_arg_converter, &seq_arg, | |||
| 6342 | simd_arg_converter, &stride_arg | |||
| 6343 | #if 1 | |||
| 6344 | ,simd_arg_converter, &nlane_arg | |||
| 6345 | #endif | |||
| 6346 | #if 0 | |||
| 6347 | ,simd_arg_converter, &fill_arg | |||
| 6348 | #endif | |||
| 6349 | )) { | |||
| 6350 | return NULL((void*)0); | |||
| 6351 | } | |||
| 6352 | npyv_lanetype_f32 *seq_ptr = seq_arg.data.qf32; | |||
| 6353 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6354 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6355 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f324; | |||
| 6356 | if (stride < 0) { | |||
| 6357 | seq_ptr += cur_seq_len -1; | |||
| 6358 | min_seq_len = -min_seq_len; | |||
| 6359 | } | |||
| 6360 | if (cur_seq_len < min_seq_len) { | |||
| 6361 | PyErr_Format(PyExc_ValueError, | |||
| 6362 | "loadn_tillz_f32(), according to provided stride %d, the " | |||
| 6363 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6364 | stride, min_seq_len, cur_seq_len | |||
| 6365 | ); | |||
| 6366 | goto err; | |||
| 6367 | } | |||
| 6368 | npyv_f32 rvec = npyv_loadn_tillz_f32( | |||
| 6369 | seq_ptr, stride | |||
| 6370 | #if 1 | |||
| 6371 | , nlane_arg.data.u32 | |||
| 6372 | #endif | |||
| 6373 | #if 0 | |||
| 6374 | , fill_arg.data.f32 | |||
| 6375 | #endif | |||
| 6376 | ); | |||
| 6377 | simd_arg ret = { | |||
| 6378 | .dtype = simd_data_vf32, .data = {.vf32=rvec} | |||
| 6379 | }; | |||
| 6380 | simd_arg_free(&seq_arg); | |||
| 6381 | return simd_arg_to_obj(&ret); | |||
| 6382 | err: | |||
| 6383 | simd_arg_free(&seq_arg); | |||
| 6384 | return NULL((void*)0); | |||
| 6385 | } | |||
| 6386 | ||||
| 6387 | ||||
| 6388 | // Non-contiguous Store | |||
| 6389 | #line 178 | |||
| 6390 | static PyObject * | |||
| 6391 | simd__intrin_storen_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6392 | { | |||
| 6393 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6394 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6395 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6396 | #if 0 | |||
| 6397 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6398 | #endif | |||
| 6399 | if (!PyArg_ParseTuple( | |||
| 6400 | args, "O&O&O&:storen_f32", | |||
| 6401 | simd_arg_converter, &seq_arg, | |||
| 6402 | simd_arg_converter, &stride_arg | |||
| 6403 | #if 0 | |||
| 6404 | ,simd_arg_converter, &nlane_arg | |||
| 6405 | #endif | |||
| 6406 | ,simd_arg_converter, &vec_arg | |||
| 6407 | )) { | |||
| 6408 | return NULL((void*)0); | |||
| 6409 | } | |||
| 6410 | npyv_lanetype_f32 *seq_ptr = seq_arg.data.qf32; | |||
| 6411 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6412 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6413 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f324; | |||
| 6414 | if (stride < 0) { | |||
| 6415 | seq_ptr += cur_seq_len -1; | |||
| 6416 | min_seq_len = -min_seq_len; | |||
| 6417 | } | |||
| 6418 | // overflow guard | |||
| 6419 | if (cur_seq_len < min_seq_len) { | |||
| 6420 | PyErr_Format(PyExc_ValueError, | |||
| 6421 | "storen_f32(), according to provided stride %d, the" | |||
| 6422 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6423 | stride, min_seq_len, cur_seq_len | |||
| 6424 | ); | |||
| 6425 | goto err; | |||
| 6426 | } | |||
| 6427 | npyv_storen_f32( | |||
| 6428 | seq_ptr, stride | |||
| 6429 | #if 0 | |||
| 6430 | ,nlane_arg.data.u32 | |||
| 6431 | #endif | |||
| 6432 | ,vec_arg.data.vf32 | |||
| 6433 | ); | |||
| 6434 | // write-back | |||
| 6435 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6436 | goto err; | |||
| 6437 | } | |||
| 6438 | simd_arg_free(&seq_arg); | |||
| 6439 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6440 | err: | |||
| 6441 | simd_arg_free(&seq_arg); | |||
| 6442 | return NULL((void*)0); | |||
| 6443 | } | |||
| 6444 | ||||
| 6445 | #line 178 | |||
| 6446 | static PyObject * | |||
| 6447 | simd__intrin_storen_till_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6448 | { | |||
| 6449 | simd_arg seq_arg = {.dtype = simd_data_qf32}; | |||
| 6450 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6451 | simd_arg vec_arg = {.dtype = simd_data_vf32}; | |||
| 6452 | #if 1 | |||
| 6453 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6454 | #endif | |||
| 6455 | if (!PyArg_ParseTuple( | |||
| 6456 | args, "O&O&O&O&:storen_f32", | |||
| 6457 | simd_arg_converter, &seq_arg, | |||
| 6458 | simd_arg_converter, &stride_arg | |||
| 6459 | #if 1 | |||
| 6460 | ,simd_arg_converter, &nlane_arg | |||
| 6461 | #endif | |||
| 6462 | ,simd_arg_converter, &vec_arg | |||
| 6463 | )) { | |||
| 6464 | return NULL((void*)0); | |||
| 6465 | } | |||
| 6466 | npyv_lanetype_f32 *seq_ptr = seq_arg.data.qf32; | |||
| 6467 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6468 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6469 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f324; | |||
| 6470 | if (stride < 0) { | |||
| 6471 | seq_ptr += cur_seq_len -1; | |||
| 6472 | min_seq_len = -min_seq_len; | |||
| 6473 | } | |||
| 6474 | // overflow guard | |||
| 6475 | if (cur_seq_len < min_seq_len) { | |||
| 6476 | PyErr_Format(PyExc_ValueError, | |||
| 6477 | "storen_till_f32(), according to provided stride %d, the" | |||
| 6478 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6479 | stride, min_seq_len, cur_seq_len | |||
| 6480 | ); | |||
| 6481 | goto err; | |||
| 6482 | } | |||
| 6483 | npyv_storen_till_f32( | |||
| 6484 | seq_ptr, stride | |||
| 6485 | #if 1 | |||
| 6486 | ,nlane_arg.data.u32 | |||
| 6487 | #endif | |||
| 6488 | ,vec_arg.data.vf32 | |||
| 6489 | ); | |||
| 6490 | // write-back | |||
| 6491 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf32, simd_data_qf32)) { | |||
| 6492 | goto err; | |||
| 6493 | } | |||
| 6494 | simd_arg_free(&seq_arg); | |||
| 6495 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6496 | err: | |||
| 6497 | simd_arg_free(&seq_arg); | |||
| 6498 | return NULL((void*)0); | |||
| 6499 | } | |||
| 6500 | ||||
| 6501 | #endif // 1 | |||
| 6502 | ||||
| 6503 | /*************************** | |||
| 6504 | * Misc | |||
| 6505 | ***************************/ | |||
| 6506 | SIMD_IMPL_INTRIN_0(zero_f32, vf32)static PyObject *simd__intrin_zero_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_f32") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vf32, .data = {.vf32 = _mm_setzero_ps()}, } ; return simd_arg_to_obj(&a); } | |||
| 6507 | SIMD_IMPL_INTRIN_1(setall_f32, vf32, f32)static PyObject *simd__intrin_setall_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_f32}; if (!PyArg_ParseTuple( args, "O&:" "setall_f32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vf32 = _mm_set1_ps( arg.data.f32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6508 | SIMD_IMPL_INTRIN_3(select_f32, vf32, vb32, vf32, vf32)static PyObject *simd__intrin_select_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_select_f32 ( arg1.data.vb32, arg2.data.vf32, arg3.data.vf32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6509 | ||||
| 6510 | #line 246 | |||
| 6511 | #if 1 | |||
| 6512 | SIMD_IMPL_INTRIN_1(reinterpret_u8_f32, vu8, vf32)static PyObject *simd__intrin_reinterpret_u8_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj (&ret); } | |||
| 6513 | #endif // simd_sup2 | |||
| 6514 | ||||
| 6515 | #line 246 | |||
| 6516 | #if 1 | |||
| 6517 | SIMD_IMPL_INTRIN_1(reinterpret_s8_f32, vs8, vf32)static PyObject *simd__intrin_reinterpret_s8_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj (&ret); } | |||
| 6518 | #endif // simd_sup2 | |||
| 6519 | ||||
| 6520 | #line 246 | |||
| 6521 | #if 1 | |||
| 6522 | SIMD_IMPL_INTRIN_1(reinterpret_u16_f32, vu16, vf32)static PyObject *simd__intrin_reinterpret_u16_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj (&ret); } | |||
| 6523 | #endif // simd_sup2 | |||
| 6524 | ||||
| 6525 | #line 246 | |||
| 6526 | #if 1 | |||
| 6527 | SIMD_IMPL_INTRIN_1(reinterpret_s16_f32, vs16, vf32)static PyObject *simd__intrin_reinterpret_s16_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj (&ret); } | |||
| 6528 | #endif // simd_sup2 | |||
| 6529 | ||||
| 6530 | #line 246 | |||
| 6531 | #if 1 | |||
| 6532 | SIMD_IMPL_INTRIN_1(reinterpret_u32_f32, vu32, vf32)static PyObject *simd__intrin_reinterpret_u32_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj (&ret); } | |||
| 6533 | #endif // simd_sup2 | |||
| 6534 | ||||
| 6535 | #line 246 | |||
| 6536 | #if 1 | |||
| 6537 | SIMD_IMPL_INTRIN_1(reinterpret_s32_f32, vs32, vf32)static PyObject *simd__intrin_reinterpret_s32_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj (&ret); } | |||
| 6538 | #endif // simd_sup2 | |||
| 6539 | ||||
| 6540 | #line 246 | |||
| 6541 | #if 1 | |||
| 6542 | SIMD_IMPL_INTRIN_1(reinterpret_u64_f32, vu64, vf32)static PyObject *simd__intrin_reinterpret_u64_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj (&ret); } | |||
| 6543 | #endif // simd_sup2 | |||
| 6544 | ||||
| 6545 | #line 246 | |||
| 6546 | #if 1 | |||
| 6547 | SIMD_IMPL_INTRIN_1(reinterpret_s64_f32, vs64, vf32)static PyObject *simd__intrin_reinterpret_s64_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = _mm_castps_si128 ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj (&ret); } | |||
| 6548 | #endif // simd_sup2 | |||
| 6549 | ||||
| 6550 | #line 246 | |||
| 6551 | #if 1 | |||
| 6552 | SIMD_IMPL_INTRIN_1(reinterpret_f32_f32, vf32, vf32)static PyObject *simd__intrin_reinterpret_f32_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = arg.data. vf32}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret) ; } | |||
| 6553 | #endif // simd_sup2 | |||
| 6554 | ||||
| 6555 | #line 246 | |||
| 6556 | #if NPY_SIMD_F641 | |||
| 6557 | SIMD_IMPL_INTRIN_1(reinterpret_f64_f32, vf64, vf32)static PyObject *simd__intrin_reinterpret_f64_f32 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_f32", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = _mm_castps_pd ( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 6558 | #endif // simd_sup2 | |||
| 6559 | ||||
| 6560 | ||||
| 6561 | /** | |||
| 6562 | * special definition due to the nature of intrinsics | |||
| 6563 | * npyv_setf_f32 and npy_set_f32. | |||
| 6564 | */ | |||
| 6565 | #line 258 | |||
| 6566 | static PyObject * | |||
| 6567 | simd__intrin_setf_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6568 | { | |||
| 6569 | npyv_lanetype_f32 *data = simd_sequence_from_iterable(args, simd_data_qf32, npyv_nlanes_f324); | |||
| 6570 | if (data == NULL((void*)0)) { | |||
| 6571 | return NULL((void*)0); | |||
| 6572 | } | |||
| 6573 | simd_data r = {.vf32 = npyv_setf_f32(npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6574 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6575 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6576 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6577 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6578 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6579 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6580 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6581 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6582 | data[64] // for setfnpyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4])) | |||
| 6583 | )npyv__setr_ps((float)(data[1]), (float)(data[2]), (float)(data [3]), (float)(data[4]))}; | |||
| 6584 | simd_sequence_free(data); | |||
| 6585 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vf32); | |||
| 6586 | } | |||
| 6587 | ||||
| 6588 | #line 258 | |||
| 6589 | static PyObject * | |||
| 6590 | simd__intrin_set_f32(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6591 | { | |||
| 6592 | npyv_lanetype_f32 *data = simd_sequence_from_iterable(args, simd_data_qf32, npyv_nlanes_f324); | |||
| 6593 | if (data == NULL((void*)0)) { | |||
| 6594 | return NULL((void*)0); | |||
| 6595 | } | |||
| 6596 | simd_data r = {.vf32 = npyv_set_f32(npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6597 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6598 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6599 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6600 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6601 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6602 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6603 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6604 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6605 | data[64] // for setfnpyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3])) | |||
| 6606 | )npyv__setr_ps((float)(data[0]), (float)(data[1]), (float)(data [2]), (float)(data[3]))}; | |||
| 6607 | simd_sequence_free(data); | |||
| 6608 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vf32); | |||
| 6609 | } | |||
| 6610 | ||||
| 6611 | ||||
| 6612 | /*************************** | |||
| 6613 | * Reorder | |||
| 6614 | ***************************/ | |||
| 6615 | #line 287 | |||
| 6616 | SIMD_IMPL_INTRIN_2(combinel_f32, vf32, vf32, vf32)static PyObject *simd__intrin_combinel_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_f32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps (_mm_unpacklo_epi64(_mm_castps_si128(arg1.data.vf32), _mm_castps_si128 (arg2.data.vf32)))}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6617 | ||||
| 6618 | #line 287 | |||
| 6619 | SIMD_IMPL_INTRIN_2(combineh_f32, vf32, vf32, vf32)static PyObject *simd__intrin_combineh_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_f32" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf32 = _mm_castsi128_ps (_mm_unpackhi_epi64(_mm_castps_si128(arg1.data.vf32), _mm_castps_si128 (arg2.data.vf32)))}; simd_arg_free(&arg1); simd_arg_free( &arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6620 | ||||
| 6621 | ||||
| 6622 | #line 293 | |||
| 6623 | SIMD_IMPL_INTRIN_2(combine_f32, vf32x2, vf32, vf32)static PyObject *simd__intrin_combine_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_f32", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf32x2 = npyv_combine_f32 ( arg1.data.vf32, arg2.data.vf32 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32x2 }; return simd_arg_to_obj(&ret); } | |||
| 6624 | ||||
| 6625 | #line 293 | |||
| 6626 | SIMD_IMPL_INTRIN_2(zip_f32, vf32x2, vf32, vf32)static PyObject *simd__intrin_zip_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32x2 = npyv_zip_f32( arg1.data.vf32, arg2.data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32x2 }; return simd_arg_to_obj(&ret); } | |||
| 6627 | ||||
| 6628 | ||||
| 6629 | #if 1 | |||
| 6630 | SIMD_IMPL_INTRIN_1(rev64_f32, vf32, vf32)static PyObject *simd__intrin_rev64_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "rev64_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf32 = npyv_rev64_f32( arg.data.vf32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6631 | #endif | |||
| 6632 | ||||
| 6633 | /*************************** | |||
| 6634 | * Operators | |||
| 6635 | ***************************/ | |||
| 6636 | #if 0 > 0 | |||
| 6637 | SIMD_IMPL_INTRIN_2(shl_f32, vf32, vf32, u8)static PyObject *simd__intrin_shl_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_shl_f32( arg1.data.vf32, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6638 | SIMD_IMPL_INTRIN_2(shr_f32, vf32, vf32, u8)static PyObject *simd__intrin_shr_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_shr_f32( arg1.data.vf32, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6639 | // immediate constant | |||
| 6640 | SIMD_IMPL_INTRIN_2IMM(shli_f32, vf32, vf32, 0)static PyObject *simd__intrin_shli_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vf32 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shli_f32, vf32 ) data.vf32; simd_arg_free (&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6641 | SIMD_IMPL_INTRIN_2IMM(shri_f32, vf32, vf32, 0)static PyObject *simd__intrin_shri_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vf32 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shri_f32, vf32 ) data.vf32; simd_arg_free (&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6642 | #endif // shl_imm | |||
| 6643 | ||||
| 6644 | #line 314 | |||
| 6645 | SIMD_IMPL_INTRIN_2(and_f32, vf32, vf32, vf32)static PyObject *simd__intrin_and_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""and_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_and_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6646 | ||||
| 6647 | #line 314 | |||
| 6648 | SIMD_IMPL_INTRIN_2(or_f32, vf32, vf32, vf32)static PyObject *simd__intrin_or_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""or_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_or_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6649 | ||||
| 6650 | #line 314 | |||
| 6651 | SIMD_IMPL_INTRIN_2(xor_f32, vf32, vf32, vf32)static PyObject *simd__intrin_xor_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_xor_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6652 | ||||
| 6653 | ||||
| 6654 | SIMD_IMPL_INTRIN_1(not_f32, vf32, vf32)static PyObject *simd__intrin_not_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "not_f32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vf32 = _mm_xor_ps(arg.data.vf32, _mm_castsi128_ps (_mm_set1_epi32(-1)))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 6655 | ||||
| 6656 | #line 322 | |||
| 6657 | SIMD_IMPL_INTRIN_2(cmpeq_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmpeq_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmpeq_ps(arg1 .data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6658 | ||||
| 6659 | #line 322 | |||
| 6660 | SIMD_IMPL_INTRIN_2(cmpneq_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmpneq_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmpneq_ps( arg1.data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6661 | ||||
| 6662 | #line 322 | |||
| 6663 | SIMD_IMPL_INTRIN_2(cmpgt_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmpgt_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmpgt_ps(arg1 .data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6664 | ||||
| 6665 | #line 322 | |||
| 6666 | SIMD_IMPL_INTRIN_2(cmpge_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmpge_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmpge_ps(arg1 .data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6667 | ||||
| 6668 | #line 322 | |||
| 6669 | SIMD_IMPL_INTRIN_2(cmplt_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmplt_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmplt_ps(arg1 .data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6670 | ||||
| 6671 | #line 322 | |||
| 6672 | SIMD_IMPL_INTRIN_2(cmple_f32, vb32, vf32, vf32)static PyObject *simd__intrin_cmple_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_castps_si128(_mm_cmple_ps(arg1 .data.vf32, arg2.data.vf32))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6673 | ||||
| 6674 | ||||
| 6675 | /*************************** | |||
| 6676 | * Conversion | |||
| 6677 | ***************************/ | |||
| 6678 | SIMD_IMPL_INTRIN_1(cvt_f32_b32, vf32, vb32)static PyObject *simd__intrin_cvt_f32_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_f32_b32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vf32 = _mm_castsi128_ps( arg.data.vb32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6679 | SIMD_IMPL_INTRIN_1(cvt_b32_f32, vb32, vf32)static PyObject *simd__intrin_cvt_b32_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b32_f32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb32 = _mm_castps_si128( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 6680 | #if 0 | |||
| 6681 | SIMD_IMPL_INTRIN_1(expand_f32_f32, vf32x2, vf32)static PyObject *simd__intrin_expand_f32_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "expand_f32_f32", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vf32x2 = npyv_expand_f32_f32( arg.data .vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vf32x2 }; return simd_arg_to_obj(&ret ); } | |||
| 6682 | #endif // expand_sup | |||
| 6683 | /*************************** | |||
| 6684 | * Arithmetic | |||
| 6685 | ***************************/ | |||
| 6686 | #line 339 | |||
| 6687 | SIMD_IMPL_INTRIN_2(add_f32, vf32, vf32, vf32)static PyObject *simd__intrin_add_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""add_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_add_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6688 | ||||
| 6689 | #line 339 | |||
| 6690 | SIMD_IMPL_INTRIN_2(sub_f32, vf32, vf32, vf32)static PyObject *simd__intrin_sub_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_sub_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6691 | ||||
| 6692 | ||||
| 6693 | #if 0 | |||
| 6694 | #line 346 | |||
| 6695 | SIMD_IMPL_INTRIN_2(adds_f32, vf32, vf32, vf32)static PyObject *simd__intrin_adds_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_adds_f32( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6696 | ||||
| 6697 | #line 346 | |||
| 6698 | SIMD_IMPL_INTRIN_2(subs_f32, vf32, vf32, vf32)static PyObject *simd__intrin_subs_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_subs_f32( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6699 | ||||
| 6700 | #endif // sat_sup | |||
| 6701 | ||||
| 6702 | #if 1 | |||
| 6703 | SIMD_IMPL_INTRIN_2(mul_f32, vf32, vf32, vf32)static PyObject *simd__intrin_mul_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_mul_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6704 | #endif // mul_sup | |||
| 6705 | ||||
| 6706 | #if 1 | |||
| 6707 | SIMD_IMPL_INTRIN_2(div_f32, vf32, vf32, vf32)static PyObject *simd__intrin_div_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""div_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_div_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6708 | #endif // div_sup | |||
| 6709 | ||||
| 6710 | #if 0 | |||
| 6711 | SIMD_IMPL_INTRIN_1(divisor_f32, vf32x3, f32)static PyObject *simd__intrin_divisor_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_f32}; if (!PyArg_ParseTuple( args, "O&:" "divisor_f32", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vf32x3 = npyv_divisor_f32( arg.data.f32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vf32x3 }; return simd_arg_to_obj(&ret); } | |||
| 6712 | SIMD_IMPL_INTRIN_2(divc_f32, vf32, vf32, vf32x3)static PyObject *simd__intrin_divc_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_divc_f32( arg1.data.vf32, arg2 .data.vf32x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6713 | #endif // intdiv_sup | |||
| 6714 | ||||
| 6715 | #if 1 | |||
| 6716 | #line 367 | |||
| 6717 | SIMD_IMPL_INTRIN_3(muladd_f32, vf32, vf32, vf32, vf32)static PyObject *simd__intrin_muladd_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_muladd_f32 ( arg1.data.vf32, arg2.data.vf32, arg3.data.vf32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6718 | ||||
| 6719 | #line 367 | |||
| 6720 | SIMD_IMPL_INTRIN_3(mulsub_f32, vf32, vf32, vf32, vf32)static PyObject *simd__intrin_mulsub_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_mulsub_f32 ( arg1.data.vf32, arg2.data.vf32, arg3.data.vf32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6721 | ||||
| 6722 | #line 367 | |||
| 6723 | SIMD_IMPL_INTRIN_3(nmuladd_f32, vf32, vf32, vf32, vf32)static PyObject *simd__intrin_nmuladd_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_nmuladd_f32 ( arg1.data.vf32, arg2.data.vf32, arg3.data.vf32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6724 | ||||
| 6725 | #line 367 | |||
| 6726 | SIMD_IMPL_INTRIN_3(nmulsub_f32, vf32, vf32, vf32, vf32)static PyObject *simd__intrin_nmulsub_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf32 = npyv_nmulsub_f32 ( arg1.data.vf32, arg2.data.vf32, arg3.data.vf32 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6727 | ||||
| 6728 | #endif // fused_sup | |||
| 6729 | ||||
| 6730 | #if 1 | |||
| 6731 | SIMD_IMPL_INTRIN_1(sum_f32, f32, vf32)static PyObject *simd__intrin_sum_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "sum_f32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.f32 = npyv_sum_f32( arg.data.vf32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_f32 }; return simd_arg_to_obj(&ret); } | |||
| 6732 | #endif // sum_sup | |||
| 6733 | ||||
| 6734 | #if 0 | |||
| 6735 | SIMD_IMPL_INTRIN_1(sumup_f32, f32, vf32)static PyObject *simd__intrin_sumup_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "sumup_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.f32 = npyv_sumup_f32( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_f32 }; return simd_arg_to_obj(&ret); } | |||
| 6736 | #endif // sumup_sup | |||
| 6737 | ||||
| 6738 | /*************************** | |||
| 6739 | * Math | |||
| 6740 | ***************************/ | |||
| 6741 | #if 1 | |||
| 6742 | #line 386 | |||
| 6743 | SIMD_IMPL_INTRIN_1(sqrt_f32, vf32, vf32)static PyObject *simd__intrin_sqrt_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_f32", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vf32 = _mm_sqrt_ps( arg.data.vf32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6744 | ||||
| 6745 | #line 386 | |||
| 6746 | SIMD_IMPL_INTRIN_1(recip_f32, vf32, vf32)static PyObject *simd__intrin_recip_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "recip_f32", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf32 = npyv_recip_f32( arg.data.vf32 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6747 | ||||
| 6748 | #line 386 | |||
| 6749 | SIMD_IMPL_INTRIN_1(abs_f32, vf32, vf32)static PyObject *simd__intrin_abs_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "abs_f32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vf32 = npyv_abs_f32( arg.data.vf32 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6750 | ||||
| 6751 | #line 386 | |||
| 6752 | SIMD_IMPL_INTRIN_1(square_f32, vf32, vf32)static PyObject *simd__intrin_square_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "square_f32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vf32 = npyv_square_f32( arg.data.vf32 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6753 | ||||
| 6754 | #endif | |||
| 6755 | ||||
| 6756 | #line 393 | |||
| 6757 | SIMD_IMPL_INTRIN_2(max_f32, vf32, vf32, vf32)static PyObject *simd__intrin_max_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""max_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_max_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6758 | ||||
| 6759 | #line 393 | |||
| 6760 | SIMD_IMPL_INTRIN_2(min_f32, vf32, vf32, vf32)static PyObject *simd__intrin_min_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""min_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = _mm_min_ps( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6761 | ||||
| 6762 | ||||
| 6763 | #if 1 | |||
| 6764 | #line 400 | |||
| 6765 | SIMD_IMPL_INTRIN_2(maxp_f32, vf32, vf32, vf32)static PyObject *simd__intrin_maxp_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_maxp_f32( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6766 | ||||
| 6767 | #line 400 | |||
| 6768 | SIMD_IMPL_INTRIN_2(minp_f32, vf32, vf32, vf32)static PyObject *simd__intrin_minp_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_f32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf32 = npyv_minp_f32( arg1.data.vf32, arg2 .data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(&ret); } | |||
| 6769 | ||||
| 6770 | #endif | |||
| 6771 | ||||
| 6772 | /*************************** | |||
| 6773 | * Mask operations | |||
| 6774 | ***************************/ | |||
| 6775 | #line 410 | |||
| 6776 | SIMD_IMPL_INTRIN_4(ifadd_f32, vf32, vb32, vf32, vf32, vf32)static PyObject *simd__intrin_ifadd_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; simd_arg arg4 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_f32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vf32 = npyv_ifadd_f32 ( arg1.data.vb32, arg2.data.vf32, arg3.data.vf32, arg4.data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(& ret); } | |||
| 6777 | ||||
| 6778 | #line 410 | |||
| 6779 | SIMD_IMPL_INTRIN_4(ifsub_f32, vf32, vb32, vf32, vf32, vf32)static PyObject *simd__intrin_ifsub_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vf32 }; simd_arg arg3 = {.dtype = simd_data_vf32}; simd_arg arg4 = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_f32", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vf32 = npyv_ifsub_f32 ( arg1.data.vb32, arg2.data.vf32, arg3.data.vf32, arg4.data.vf32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj(& ret); } | |||
| 6780 | ||||
| 6781 | ||||
| 6782 | #endif // simd_sup | |||
| 6783 | ||||
| 6784 | #line 34 | |||
| 6785 | #if NPY_SIMD_F641 | |||
| 6786 | /*************************** | |||
| 6787 | * Memory | |||
| 6788 | ***************************/ | |||
| 6789 | #line 41 | |||
| 6790 | SIMD_IMPL_INTRIN_1(load_f64, vf64, qf64)static PyObject *simd__intrin_load_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf64}; if (!PyArg_ParseTuple( args, "O&:" "load_f64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vf64 = _mm_loadu_pd( arg.data.qf64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6791 | ||||
| 6792 | #line 41 | |||
| 6793 | SIMD_IMPL_INTRIN_1(loada_f64, vf64, qf64)static PyObject *simd__intrin_loada_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf64}; if (!PyArg_ParseTuple( args, "O&:" "loada_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf64 = _mm_load_pd( arg.data.qf64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6794 | ||||
| 6795 | #line 41 | |||
| 6796 | SIMD_IMPL_INTRIN_1(loads_f64, vf64, qf64)static PyObject *simd__intrin_loads_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf64}; if (!PyArg_ParseTuple( args, "O&:" "loads_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf64 = _mm_castsi128_pd(_mm_load_si128( (const __m128i *)(arg.data.qf64)))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6797 | ||||
| 6798 | #line 41 | |||
| 6799 | SIMD_IMPL_INTRIN_1(loadl_f64, vf64, qf64)static PyObject *simd__intrin_loadl_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_qf64}; if (!PyArg_ParseTuple( args, "O&:" "loadl_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf64 = _mm_castsi128_pd(npyv_loadl_u32( (const npy_uint32*)(arg.data.qf64)))}; simd_arg_free(&arg ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6800 | ||||
| 6801 | #line 46 | |||
| 6802 | // special definition due to the nature of store | |||
| 6803 | static PyObject * | |||
| 6804 | simd__intrin_store_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6805 | { | |||
| 6806 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6807 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6808 | if (!PyArg_ParseTuple( | |||
| 6809 | args, "O&O&:store_f64", | |||
| 6810 | simd_arg_converter, &seq_arg, | |||
| 6811 | simd_arg_converter, &vec_arg | |||
| 6812 | )) { | |||
| 6813 | return NULL((void*)0); | |||
| 6814 | } | |||
| 6815 | npyv_store_f64_mm_storeu_pd(seq_arg.data.qf64, vec_arg.data.vf64); | |||
| 6816 | // write-back | |||
| 6817 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6818 | simd_arg_free(&seq_arg); | |||
| 6819 | return NULL((void*)0); | |||
| 6820 | } | |||
| 6821 | simd_arg_free(&seq_arg); | |||
| 6822 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6823 | } | |||
| 6824 | ||||
| 6825 | #line 46 | |||
| 6826 | // special definition due to the nature of storea | |||
| 6827 | static PyObject * | |||
| 6828 | simd__intrin_storea_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6829 | { | |||
| 6830 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6831 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6832 | if (!PyArg_ParseTuple( | |||
| 6833 | args, "O&O&:storea_f64", | |||
| 6834 | simd_arg_converter, &seq_arg, | |||
| 6835 | simd_arg_converter, &vec_arg | |||
| 6836 | )) { | |||
| 6837 | return NULL((void*)0); | |||
| 6838 | } | |||
| 6839 | npyv_storea_f64_mm_store_pd(seq_arg.data.qf64, vec_arg.data.vf64); | |||
| 6840 | // write-back | |||
| 6841 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6842 | simd_arg_free(&seq_arg); | |||
| 6843 | return NULL((void*)0); | |||
| 6844 | } | |||
| 6845 | simd_arg_free(&seq_arg); | |||
| 6846 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6847 | } | |||
| 6848 | ||||
| 6849 | #line 46 | |||
| 6850 | // special definition due to the nature of stores | |||
| 6851 | static PyObject * | |||
| 6852 | simd__intrin_stores_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6853 | { | |||
| 6854 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6855 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6856 | if (!PyArg_ParseTuple( | |||
| 6857 | args, "O&O&:stores_f64", | |||
| 6858 | simd_arg_converter, &seq_arg, | |||
| 6859 | simd_arg_converter, &vec_arg | |||
| 6860 | )) { | |||
| 6861 | return NULL((void*)0); | |||
| 6862 | } | |||
| 6863 | npyv_stores_f64_mm_stream_pd(seq_arg.data.qf64, vec_arg.data.vf64); | |||
| 6864 | // write-back | |||
| 6865 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6866 | simd_arg_free(&seq_arg); | |||
| 6867 | return NULL((void*)0); | |||
| 6868 | } | |||
| 6869 | simd_arg_free(&seq_arg); | |||
| 6870 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6871 | } | |||
| 6872 | ||||
| 6873 | #line 46 | |||
| 6874 | // special definition due to the nature of storel | |||
| 6875 | static PyObject * | |||
| 6876 | simd__intrin_storel_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6877 | { | |||
| 6878 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6879 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6880 | if (!PyArg_ParseTuple( | |||
| 6881 | args, "O&O&:storel_f64", | |||
| 6882 | simd_arg_converter, &seq_arg, | |||
| 6883 | simd_arg_converter, &vec_arg | |||
| 6884 | )) { | |||
| 6885 | return NULL((void*)0); | |||
| 6886 | } | |||
| 6887 | npyv_storel_f64(seq_arg.data.qf64, vec_arg.data.vf64)_mm_storel_epi64((__m128i*)(seq_arg.data.qf64), _mm_castpd_si128 (vec_arg.data.vf64));; | |||
| 6888 | // write-back | |||
| 6889 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6890 | simd_arg_free(&seq_arg); | |||
| 6891 | return NULL((void*)0); | |||
| 6892 | } | |||
| 6893 | simd_arg_free(&seq_arg); | |||
| 6894 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6895 | } | |||
| 6896 | ||||
| 6897 | #line 46 | |||
| 6898 | // special definition due to the nature of storeh | |||
| 6899 | static PyObject * | |||
| 6900 | simd__intrin_storeh_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6901 | { | |||
| 6902 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6903 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6904 | if (!PyArg_ParseTuple( | |||
| 6905 | args, "O&O&:storeh_f64", | |||
| 6906 | simd_arg_converter, &seq_arg, | |||
| 6907 | simd_arg_converter, &vec_arg | |||
| 6908 | )) { | |||
| 6909 | return NULL((void*)0); | |||
| 6910 | } | |||
| 6911 | npyv_storeh_f64(seq_arg.data.qf64, vec_arg.data.vf64)npyv_storeh_u32((npy_uint32*)(seq_arg.data.qf64), _mm_castpd_si128 (vec_arg.data.vf64)); | |||
| 6912 | // write-back | |||
| 6913 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6914 | simd_arg_free(&seq_arg); | |||
| 6915 | return NULL((void*)0); | |||
| 6916 | } | |||
| 6917 | simd_arg_free(&seq_arg); | |||
| 6918 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6919 | } | |||
| 6920 | ||||
| 6921 | ||||
| 6922 | /**************************************** | |||
| 6923 | * Non-contiguous/Partial Memory access | |||
| 6924 | ****************************************/ | |||
| 6925 | #if 1 | |||
| 6926 | // Partial Load | |||
| 6927 | SIMD_IMPL_INTRIN_3(load_till_f64, vf64, qf64, u32, f64)static PyObject *simd__intrin_load_till_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qf64}; simd_arg arg2 = {.dtype = simd_data_u32 }; simd_arg arg3 = {.dtype = simd_data_f64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""load_till_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_load_till_f64 ( arg1.data.qf64, arg2.data.u32, arg3.data.f64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6928 | SIMD_IMPL_INTRIN_2(load_tillz_f64, vf64, qf64, u32)static PyObject *simd__intrin_load_tillz_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_qf64}; simd_arg arg2 = {.dtype = simd_data_u32 }; if (!PyArg_ParseTuple( args, "O&O&:""load_tillz_f64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf64 = npyv_load_tillz_f64 ( arg1.data.qf64, arg2.data.u32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 6929 | ||||
| 6930 | // Partial Store | |||
| 6931 | static PyObject * | |||
| 6932 | simd__intrin_store_till_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6933 | { | |||
| 6934 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6935 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6936 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 6937 | if (!PyArg_ParseTuple( | |||
| 6938 | args, "O&O&O&:store_till_f64", | |||
| 6939 | simd_arg_converter, &seq_arg, | |||
| 6940 | simd_arg_converter, &nlane_arg, | |||
| 6941 | simd_arg_converter, &vec_arg | |||
| 6942 | )) { | |||
| 6943 | return NULL((void*)0); | |||
| 6944 | } | |||
| 6945 | npyv_store_till_f64( | |||
| 6946 | seq_arg.data.qf64, nlane_arg.data.u32, vec_arg.data.vf64 | |||
| 6947 | ); | |||
| 6948 | // write-back | |||
| 6949 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 6950 | simd_arg_free(&seq_arg); | |||
| 6951 | return NULL((void*)0); | |||
| 6952 | } | |||
| 6953 | simd_arg_free(&seq_arg); | |||
| 6954 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 6955 | } | |||
| 6956 | ||||
| 6957 | // Non-contiguous Load | |||
| 6958 | #line 112 | |||
| 6959 | static PyObject * | |||
| 6960 | simd__intrin_loadn_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 6961 | { | |||
| 6962 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 6963 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 6964 | #if 0 | |||
| 6965 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 6966 | #endif // till | |||
| 6967 | #if 0 | |||
| 6968 | simd_arg fill_arg = {.dtype = simd_data_f64}; | |||
| 6969 | #endif | |||
| 6970 | if (!PyArg_ParseTuple( | |||
| 6971 | args, "O&O&:loadn_f64", | |||
| 6972 | simd_arg_converter, &seq_arg, | |||
| 6973 | simd_arg_converter, &stride_arg | |||
| 6974 | #if 0 | |||
| 6975 | ,simd_arg_converter, &nlane_arg | |||
| 6976 | #endif | |||
| 6977 | #if 0 | |||
| 6978 | ,simd_arg_converter, &fill_arg | |||
| 6979 | #endif | |||
| 6980 | )) { | |||
| 6981 | return NULL((void*)0); | |||
| 6982 | } | |||
| 6983 | npyv_lanetype_f64 *seq_ptr = seq_arg.data.qf64; | |||
| 6984 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 6985 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 6986 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f642; | |||
| 6987 | if (stride < 0) { | |||
| 6988 | seq_ptr += cur_seq_len -1; | |||
| 6989 | min_seq_len = -min_seq_len; | |||
| 6990 | } | |||
| 6991 | if (cur_seq_len < min_seq_len) { | |||
| 6992 | PyErr_Format(PyExc_ValueError, | |||
| 6993 | "loadn_f64(), according to provided stride %d, the " | |||
| 6994 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 6995 | stride, min_seq_len, cur_seq_len | |||
| 6996 | ); | |||
| 6997 | goto err; | |||
| 6998 | } | |||
| 6999 | npyv_f64 rvec = npyv_loadn_f64( | |||
| 7000 | seq_ptr, stride | |||
| 7001 | #if 0 | |||
| 7002 | , nlane_arg.data.u32 | |||
| 7003 | #endif | |||
| 7004 | #if 0 | |||
| 7005 | , fill_arg.data.f64 | |||
| 7006 | #endif | |||
| 7007 | ); | |||
| 7008 | simd_arg ret = { | |||
| 7009 | .dtype = simd_data_vf64, .data = {.vf64=rvec} | |||
| 7010 | }; | |||
| 7011 | simd_arg_free(&seq_arg); | |||
| 7012 | return simd_arg_to_obj(&ret); | |||
| 7013 | err: | |||
| 7014 | simd_arg_free(&seq_arg); | |||
| 7015 | return NULL((void*)0); | |||
| 7016 | } | |||
| 7017 | ||||
| 7018 | #line 112 | |||
| 7019 | static PyObject * | |||
| 7020 | simd__intrin_loadn_till_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7021 | { | |||
| 7022 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 7023 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 7024 | #if 1 | |||
| 7025 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 7026 | #endif // till | |||
| 7027 | #if 1 | |||
| 7028 | simd_arg fill_arg = {.dtype = simd_data_f64}; | |||
| 7029 | #endif | |||
| 7030 | if (!PyArg_ParseTuple( | |||
| 7031 | args, "O&O&O&O&:loadn_till_f64", | |||
| 7032 | simd_arg_converter, &seq_arg, | |||
| 7033 | simd_arg_converter, &stride_arg | |||
| 7034 | #if 1 | |||
| 7035 | ,simd_arg_converter, &nlane_arg | |||
| 7036 | #endif | |||
| 7037 | #if 1 | |||
| 7038 | ,simd_arg_converter, &fill_arg | |||
| 7039 | #endif | |||
| 7040 | )) { | |||
| 7041 | return NULL((void*)0); | |||
| 7042 | } | |||
| 7043 | npyv_lanetype_f64 *seq_ptr = seq_arg.data.qf64; | |||
| 7044 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 7045 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 7046 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f642; | |||
| 7047 | if (stride < 0) { | |||
| 7048 | seq_ptr += cur_seq_len -1; | |||
| 7049 | min_seq_len = -min_seq_len; | |||
| 7050 | } | |||
| 7051 | if (cur_seq_len < min_seq_len) { | |||
| 7052 | PyErr_Format(PyExc_ValueError, | |||
| 7053 | "loadn_till_f64(), according to provided stride %d, the " | |||
| 7054 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 7055 | stride, min_seq_len, cur_seq_len | |||
| 7056 | ); | |||
| 7057 | goto err; | |||
| 7058 | } | |||
| 7059 | npyv_f64 rvec = npyv_loadn_till_f64( | |||
| 7060 | seq_ptr, stride | |||
| 7061 | #if 1 | |||
| 7062 | , nlane_arg.data.u32 | |||
| 7063 | #endif | |||
| 7064 | #if 1 | |||
| 7065 | , fill_arg.data.f64 | |||
| 7066 | #endif | |||
| 7067 | ); | |||
| 7068 | simd_arg ret = { | |||
| 7069 | .dtype = simd_data_vf64, .data = {.vf64=rvec} | |||
| 7070 | }; | |||
| 7071 | simd_arg_free(&seq_arg); | |||
| 7072 | return simd_arg_to_obj(&ret); | |||
| 7073 | err: | |||
| 7074 | simd_arg_free(&seq_arg); | |||
| 7075 | return NULL((void*)0); | |||
| 7076 | } | |||
| 7077 | ||||
| 7078 | #line 112 | |||
| 7079 | static PyObject * | |||
| 7080 | simd__intrin_loadn_tillz_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7081 | { | |||
| 7082 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 7083 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 7084 | #if 1 | |||
| 7085 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 7086 | #endif // till | |||
| 7087 | #if 0 | |||
| 7088 | simd_arg fill_arg = {.dtype = simd_data_f64}; | |||
| 7089 | #endif | |||
| 7090 | if (!PyArg_ParseTuple( | |||
| 7091 | args, "O&O&O&:loadn_tillz_f64", | |||
| 7092 | simd_arg_converter, &seq_arg, | |||
| 7093 | simd_arg_converter, &stride_arg | |||
| 7094 | #if 1 | |||
| 7095 | ,simd_arg_converter, &nlane_arg | |||
| 7096 | #endif | |||
| 7097 | #if 0 | |||
| 7098 | ,simd_arg_converter, &fill_arg | |||
| 7099 | #endif | |||
| 7100 | )) { | |||
| 7101 | return NULL((void*)0); | |||
| 7102 | } | |||
| 7103 | npyv_lanetype_f64 *seq_ptr = seq_arg.data.qf64; | |||
| 7104 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 7105 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 7106 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f642; | |||
| 7107 | if (stride < 0) { | |||
| 7108 | seq_ptr += cur_seq_len -1; | |||
| 7109 | min_seq_len = -min_seq_len; | |||
| 7110 | } | |||
| 7111 | if (cur_seq_len < min_seq_len) { | |||
| 7112 | PyErr_Format(PyExc_ValueError, | |||
| 7113 | "loadn_tillz_f64(), according to provided stride %d, the " | |||
| 7114 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 7115 | stride, min_seq_len, cur_seq_len | |||
| 7116 | ); | |||
| 7117 | goto err; | |||
| 7118 | } | |||
| 7119 | npyv_f64 rvec = npyv_loadn_tillz_f64( | |||
| 7120 | seq_ptr, stride | |||
| 7121 | #if 1 | |||
| 7122 | , nlane_arg.data.u32 | |||
| 7123 | #endif | |||
| 7124 | #if 0 | |||
| 7125 | , fill_arg.data.f64 | |||
| 7126 | #endif | |||
| 7127 | ); | |||
| 7128 | simd_arg ret = { | |||
| 7129 | .dtype = simd_data_vf64, .data = {.vf64=rvec} | |||
| 7130 | }; | |||
| 7131 | simd_arg_free(&seq_arg); | |||
| 7132 | return simd_arg_to_obj(&ret); | |||
| 7133 | err: | |||
| 7134 | simd_arg_free(&seq_arg); | |||
| 7135 | return NULL((void*)0); | |||
| 7136 | } | |||
| 7137 | ||||
| 7138 | ||||
| 7139 | // Non-contiguous Store | |||
| 7140 | #line 178 | |||
| 7141 | static PyObject * | |||
| 7142 | simd__intrin_storen_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7143 | { | |||
| 7144 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 7145 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 7146 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 7147 | #if 0 | |||
| 7148 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 7149 | #endif | |||
| 7150 | if (!PyArg_ParseTuple( | |||
| 7151 | args, "O&O&O&:storen_f64", | |||
| 7152 | simd_arg_converter, &seq_arg, | |||
| 7153 | simd_arg_converter, &stride_arg | |||
| 7154 | #if 0 | |||
| 7155 | ,simd_arg_converter, &nlane_arg | |||
| 7156 | #endif | |||
| 7157 | ,simd_arg_converter, &vec_arg | |||
| 7158 | )) { | |||
| 7159 | return NULL((void*)0); | |||
| 7160 | } | |||
| 7161 | npyv_lanetype_f64 *seq_ptr = seq_arg.data.qf64; | |||
| 7162 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 7163 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 7164 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f642; | |||
| 7165 | if (stride < 0) { | |||
| 7166 | seq_ptr += cur_seq_len -1; | |||
| 7167 | min_seq_len = -min_seq_len; | |||
| 7168 | } | |||
| 7169 | // overflow guard | |||
| 7170 | if (cur_seq_len < min_seq_len) { | |||
| 7171 | PyErr_Format(PyExc_ValueError, | |||
| 7172 | "storen_f64(), according to provided stride %d, the" | |||
| 7173 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 7174 | stride, min_seq_len, cur_seq_len | |||
| 7175 | ); | |||
| 7176 | goto err; | |||
| 7177 | } | |||
| 7178 | npyv_storen_f64( | |||
| 7179 | seq_ptr, stride | |||
| 7180 | #if 0 | |||
| 7181 | ,nlane_arg.data.u32 | |||
| 7182 | #endif | |||
| 7183 | ,vec_arg.data.vf64 | |||
| 7184 | ); | |||
| 7185 | // write-back | |||
| 7186 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 7187 | goto err; | |||
| 7188 | } | |||
| 7189 | simd_arg_free(&seq_arg); | |||
| 7190 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 7191 | err: | |||
| 7192 | simd_arg_free(&seq_arg); | |||
| 7193 | return NULL((void*)0); | |||
| 7194 | } | |||
| 7195 | ||||
| 7196 | #line 178 | |||
| 7197 | static PyObject * | |||
| 7198 | simd__intrin_storen_till_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7199 | { | |||
| 7200 | simd_arg seq_arg = {.dtype = simd_data_qf64}; | |||
| 7201 | simd_arg stride_arg = {.dtype = simd_data_s64}; | |||
| 7202 | simd_arg vec_arg = {.dtype = simd_data_vf64}; | |||
| 7203 | #if 1 | |||
| 7204 | simd_arg nlane_arg = {.dtype = simd_data_u32}; | |||
| 7205 | #endif | |||
| 7206 | if (!PyArg_ParseTuple( | |||
| 7207 | args, "O&O&O&O&:storen_f64", | |||
| 7208 | simd_arg_converter, &seq_arg, | |||
| 7209 | simd_arg_converter, &stride_arg | |||
| 7210 | #if 1 | |||
| 7211 | ,simd_arg_converter, &nlane_arg | |||
| 7212 | #endif | |||
| 7213 | ,simd_arg_converter, &vec_arg | |||
| 7214 | )) { | |||
| 7215 | return NULL((void*)0); | |||
| 7216 | } | |||
| 7217 | npyv_lanetype_f64 *seq_ptr = seq_arg.data.qf64; | |||
| 7218 | npy_intp stride = (npy_intp)stride_arg.data.s64; | |||
| 7219 | Py_ssize_t cur_seq_len = simd_sequence_len(seq_ptr); | |||
| 7220 | Py_ssize_t min_seq_len = stride * npyv_nlanes_f642; | |||
| 7221 | if (stride < 0) { | |||
| 7222 | seq_ptr += cur_seq_len -1; | |||
| 7223 | min_seq_len = -min_seq_len; | |||
| 7224 | } | |||
| 7225 | // overflow guard | |||
| 7226 | if (cur_seq_len < min_seq_len) { | |||
| 7227 | PyErr_Format(PyExc_ValueError, | |||
| 7228 | "storen_till_f64(), according to provided stride %d, the" | |||
| 7229 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 7230 | stride, min_seq_len, cur_seq_len | |||
| 7231 | ); | |||
| 7232 | goto err; | |||
| 7233 | } | |||
| 7234 | npyv_storen_till_f64( | |||
| 7235 | seq_ptr, stride | |||
| 7236 | #if 1 | |||
| 7237 | ,nlane_arg.data.u32 | |||
| 7238 | #endif | |||
| 7239 | ,vec_arg.data.vf64 | |||
| 7240 | ); | |||
| 7241 | // write-back | |||
| 7242 | if (simd_sequence_fill_iterable(seq_arg.obj, seq_arg.data.qf64, simd_data_qf64)) { | |||
| 7243 | goto err; | |||
| 7244 | } | |||
| 7245 | simd_arg_free(&seq_arg); | |||
| 7246 | Py_RETURN_NONEreturn _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (& _Py_NoneStruct); | |||
| 7247 | err: | |||
| 7248 | simd_arg_free(&seq_arg); | |||
| 7249 | return NULL((void*)0); | |||
| 7250 | } | |||
| 7251 | ||||
| 7252 | #endif // 1 | |||
| 7253 | ||||
| 7254 | /*************************** | |||
| 7255 | * Misc | |||
| 7256 | ***************************/ | |||
| 7257 | SIMD_IMPL_INTRIN_0(zero_f64, vf64)static PyObject *simd__intrin_zero_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "zero_f64") ) return ((void*)0); simd_arg a = { . dtype = simd_data_vf64, .data = {.vf64 = _mm_setzero_pd()}, } ; return simd_arg_to_obj(&a); } | |||
| 7258 | SIMD_IMPL_INTRIN_1(setall_f64, vf64, f64)static PyObject *simd__intrin_setall_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_f64}; if (!PyArg_ParseTuple( args, "O&:" "setall_f64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vf64 = _mm_set1_pd( arg.data.f64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7259 | SIMD_IMPL_INTRIN_3(select_f64, vf64, vb64, vf64, vf64)static PyObject *simd__intrin_select_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""select_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_select_f64 ( arg1.data.vb64, arg2.data.vf64, arg3.data.vf64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7260 | ||||
| 7261 | #line 246 | |||
| 7262 | #if 1 | |||
| 7263 | SIMD_IMPL_INTRIN_1(reinterpret_u8_f64, vu8, vf64)static PyObject *simd__intrin_reinterpret_u8_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u8_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu8 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu8 }; return simd_arg_to_obj (&ret); } | |||
| 7264 | #endif // simd_sup2 | |||
| 7265 | ||||
| 7266 | #line 246 | |||
| 7267 | #if 1 | |||
| 7268 | SIMD_IMPL_INTRIN_1(reinterpret_s8_f64, vs8, vf64)static PyObject *simd__intrin_reinterpret_s8_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s8_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs8 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs8 }; return simd_arg_to_obj (&ret); } | |||
| 7269 | #endif // simd_sup2 | |||
| 7270 | ||||
| 7271 | #line 246 | |||
| 7272 | #if 1 | |||
| 7273 | SIMD_IMPL_INTRIN_1(reinterpret_u16_f64, vu16, vf64)static PyObject *simd__intrin_reinterpret_u16_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u16_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu16 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu16 }; return simd_arg_to_obj (&ret); } | |||
| 7274 | #endif // simd_sup2 | |||
| 7275 | ||||
| 7276 | #line 246 | |||
| 7277 | #if 1 | |||
| 7278 | SIMD_IMPL_INTRIN_1(reinterpret_s16_f64, vs16, vf64)static PyObject *simd__intrin_reinterpret_s16_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s16_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs16 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs16 }; return simd_arg_to_obj (&ret); } | |||
| 7279 | #endif // simd_sup2 | |||
| 7280 | ||||
| 7281 | #line 246 | |||
| 7282 | #if 1 | |||
| 7283 | SIMD_IMPL_INTRIN_1(reinterpret_u32_f64, vu32, vf64)static PyObject *simd__intrin_reinterpret_u32_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u32_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu32 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu32 }; return simd_arg_to_obj (&ret); } | |||
| 7284 | #endif // simd_sup2 | |||
| 7285 | ||||
| 7286 | #line 246 | |||
| 7287 | #if 1 | |||
| 7288 | SIMD_IMPL_INTRIN_1(reinterpret_s32_f64, vs32, vf64)static PyObject *simd__intrin_reinterpret_s32_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s32_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs32 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj (&ret); } | |||
| 7289 | #endif // simd_sup2 | |||
| 7290 | ||||
| 7291 | #line 246 | |||
| 7292 | #if 1 | |||
| 7293 | SIMD_IMPL_INTRIN_1(reinterpret_u64_f64, vu64, vf64)static PyObject *simd__intrin_reinterpret_u64_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_u64_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vu64 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vu64 }; return simd_arg_to_obj (&ret); } | |||
| 7294 | #endif // simd_sup2 | |||
| 7295 | ||||
| 7296 | #line 246 | |||
| 7297 | #if 1 | |||
| 7298 | SIMD_IMPL_INTRIN_1(reinterpret_s64_f64, vs64, vf64)static PyObject *simd__intrin_reinterpret_s64_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_s64_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vs64 = _mm_castpd_si128 ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vs64 }; return simd_arg_to_obj (&ret); } | |||
| 7299 | #endif // simd_sup2 | |||
| 7300 | ||||
| 7301 | #line 246 | |||
| 7302 | #if 1 | |||
| 7303 | SIMD_IMPL_INTRIN_1(reinterpret_f32_f64, vf32, vf64)static PyObject *simd__intrin_reinterpret_f32_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f32_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf32 = _mm_castpd_ps ( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf32 }; return simd_arg_to_obj (&ret); } | |||
| 7304 | #endif // simd_sup2 | |||
| 7305 | ||||
| 7306 | #line 246 | |||
| 7307 | #if NPY_SIMD_F641 | |||
| 7308 | SIMD_IMPL_INTRIN_1(reinterpret_f64_f64, vf64, vf64)static PyObject *simd__intrin_reinterpret_f64_f64 (PyObject* ( __NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&:""reinterpret_f64_f64", simd_arg_converter, & arg )) return ((void*)0); simd_data data = {.vf64 = arg.data. vf64}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret) ; } | |||
| 7309 | #endif // simd_sup2 | |||
| 7310 | ||||
| 7311 | ||||
| 7312 | /** | |||
| 7313 | * special definition due to the nature of intrinsics | |||
| 7314 | * npyv_setf_f64 and npy_set_f64. | |||
| 7315 | */ | |||
| 7316 | #line 258 | |||
| 7317 | static PyObject * | |||
| 7318 | simd__intrin_setf_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7319 | { | |||
| 7320 | npyv_lanetype_f64 *data = simd_sequence_from_iterable(args, simd_data_qf64, npyv_nlanes_f642); | |||
| 7321 | if (data == NULL((void*)0)) { | |||
| 7322 | return NULL((void*)0); | |||
| 7323 | } | |||
| 7324 | simd_data r = {.vf64 = npyv_setf_f64(npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7325 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7326 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7327 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7328 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7329 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7330 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7331 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7332 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7333 | data[64] // for setfnpyv__setr_pd((double)(data[1]), (double)(data[2])) | |||
| 7334 | )npyv__setr_pd((double)(data[1]), (double)(data[2]))}; | |||
| 7335 | simd_sequence_free(data); | |||
| 7336 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vf64); | |||
| 7337 | } | |||
| 7338 | ||||
| 7339 | #line 258 | |||
| 7340 | static PyObject * | |||
| 7341 | simd__intrin_set_f64(PyObject* NPY_UNUSED(self)(__NPY_UNUSED_TAGGEDself) __attribute__ ((__unused__)), PyObject *args) | |||
| 7342 | { | |||
| 7343 | npyv_lanetype_f64 *data = simd_sequence_from_iterable(args, simd_data_qf64, npyv_nlanes_f642); | |||
| ||||
| 7344 | if (data == NULL((void*)0)) { | |||
| 7345 | return NULL((void*)0); | |||
| 7346 | } | |||
| 7347 | simd_data r = {.vf64 = npyv_set_f64(npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7348 | data[0], data[1], data[2], data[3], data[4], data[5], data[6], data[7],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7349 | data[8], data[9], data[10], data[11], data[12], data[13], data[14], data[15],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7350 | data[16], data[17], data[18], data[19], data[20], data[21], data[22], data[23],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7351 | data[24], data[25], data[26], data[27], data[28], data[29], data[30], data[31],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7352 | data[32], data[33], data[34], data[35], data[36], data[37], data[38], data[39],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7353 | data[40], data[41], data[42], data[43], data[44], data[45], data[46], data[47],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7354 | data[48], data[49], data[50], data[51], data[52], data[53], data[54], data[55],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7355 | data[56], data[57], data[58], data[59], data[60], data[61], data[62], data[63],npyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7356 | data[64] // for setfnpyv__setr_pd((double)(data[0]), (double)(data[1])) | |||
| 7357 | )npyv__setr_pd((double)(data[0]), (double)(data[1]))}; | |||
| 7358 | simd_sequence_free(data); | |||
| 7359 | return (PyObject*)PySIMDVector_FromData(r, simd_data_vf64); | |||
| 7360 | } | |||
| 7361 | ||||
| 7362 | ||||
| 7363 | /*************************** | |||
| 7364 | * Reorder | |||
| 7365 | ***************************/ | |||
| 7366 | #line 287 | |||
| 7367 | SIMD_IMPL_INTRIN_2(combinel_f64, vf64, vf64, vf64)static PyObject *simd__intrin_combinel_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""combinel_f64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf64 = _mm_unpacklo_pd ( arg1.data.vf64, arg2.data.vf64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7368 | ||||
| 7369 | #line 287 | |||
| 7370 | SIMD_IMPL_INTRIN_2(combineh_f64, vf64, vf64, vf64)static PyObject *simd__intrin_combineh_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""combineh_f64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf64 = _mm_unpackhi_pd ( arg1.data.vf64, arg2.data.vf64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7371 | ||||
| 7372 | ||||
| 7373 | #line 293 | |||
| 7374 | SIMD_IMPL_INTRIN_2(combine_f64, vf64x2, vf64, vf64)static PyObject *simd__intrin_combine_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""combine_f64", simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vf64x2 = npyv_combine_f64 ( arg1.data.vf64, arg2.data.vf64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64x2 }; return simd_arg_to_obj(&ret); } | |||
| 7375 | ||||
| 7376 | #line 293 | |||
| 7377 | SIMD_IMPL_INTRIN_2(zip_f64, vf64x2, vf64, vf64)static PyObject *simd__intrin_zip_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""zip_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64x2 = npyv_zip_f64( arg1.data.vf64, arg2.data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64x2 }; return simd_arg_to_obj(&ret); } | |||
| 7378 | ||||
| 7379 | ||||
| 7380 | #if 0 | |||
| 7381 | SIMD_IMPL_INTRIN_1(rev64_f64, vf64, vf64)static PyObject *simd__intrin_rev64_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "rev64_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf64 = npyv_rev64_f64( arg.data.vf64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7382 | #endif | |||
| 7383 | ||||
| 7384 | /*************************** | |||
| 7385 | * Operators | |||
| 7386 | ***************************/ | |||
| 7387 | #if 0 > 0 | |||
| 7388 | SIMD_IMPL_INTRIN_2(shl_f64, vf64, vf64, u8)static PyObject *simd__intrin_shl_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shl_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_shl_f64( arg1.data.vf64, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7389 | SIMD_IMPL_INTRIN_2(shr_f64, vf64, vf64, u8)static PyObject *simd__intrin_shr_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shr_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_shr_f64( arg1.data.vf64, arg2 .data.u8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7390 | // immediate constant | |||
| 7391 | SIMD_IMPL_INTRIN_2IMM(shli_f64, vf64, vf64, 0)static PyObject *simd__intrin_shli_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shli_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vf64 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shli_f64, vf64 ) data.vf64; simd_arg_free (&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7392 | SIMD_IMPL_INTRIN_2IMM(shri_f64, vf64, vf64, 0)static PyObject *simd__intrin_shri_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_u8 }; if (!PyArg_ParseTuple( args, "O&O&:""shri_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.u64 = 0}; data.vf64 = SIMD__IMPL_COUNT_0 ( SIMD__REPEAT_2IMM, shri_f64, vf64 ) data.vf64; simd_arg_free (&arg1); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7393 | #endif // shl_imm | |||
| 7394 | ||||
| 7395 | #line 314 | |||
| 7396 | SIMD_IMPL_INTRIN_2(and_f64, vf64, vf64, vf64)static PyObject *simd__intrin_and_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""and_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_and_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7397 | ||||
| 7398 | #line 314 | |||
| 7399 | SIMD_IMPL_INTRIN_2(or_f64, vf64, vf64, vf64)static PyObject *simd__intrin_or_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""or_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_or_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7400 | ||||
| 7401 | #line 314 | |||
| 7402 | SIMD_IMPL_INTRIN_2(xor_f64, vf64, vf64, vf64)static PyObject *simd__intrin_xor_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_xor_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7403 | ||||
| 7404 | ||||
| 7405 | SIMD_IMPL_INTRIN_1(not_f64, vf64, vf64)static PyObject *simd__intrin_not_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "not_f64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vf64 = _mm_xor_pd(arg.data.vf64, _mm_castsi128_pd (_mm_set1_epi32(-1)))}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj (&ret); } | |||
| 7406 | ||||
| 7407 | #line 322 | |||
| 7408 | SIMD_IMPL_INTRIN_2(cmpeq_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmpeq_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpeq_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmpeq_pd(arg1 .data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7409 | ||||
| 7410 | #line 322 | |||
| 7411 | SIMD_IMPL_INTRIN_2(cmpneq_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmpneq_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpneq_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmpneq_pd( arg1.data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7412 | ||||
| 7413 | #line 322 | |||
| 7414 | SIMD_IMPL_INTRIN_2(cmpgt_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmpgt_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpgt_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmpgt_pd(arg1 .data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7415 | ||||
| 7416 | #line 322 | |||
| 7417 | SIMD_IMPL_INTRIN_2(cmpge_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmpge_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmpge_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmpge_pd(arg1 .data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7418 | ||||
| 7419 | #line 322 | |||
| 7420 | SIMD_IMPL_INTRIN_2(cmplt_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmplt_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmplt_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmplt_pd(arg1 .data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7421 | ||||
| 7422 | #line 322 | |||
| 7423 | SIMD_IMPL_INTRIN_2(cmple_f64, vb64, vf64, vf64)static PyObject *simd__intrin_cmple_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""cmple_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_castpd_si128(_mm_cmple_pd(arg1 .data.vf64, arg2.data.vf64))}; simd_arg_free(&arg1); simd_arg_free (&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7424 | ||||
| 7425 | ||||
| 7426 | /*************************** | |||
| 7427 | * Conversion | |||
| 7428 | ***************************/ | |||
| 7429 | SIMD_IMPL_INTRIN_1(cvt_f64_b64, vf64, vb64)static PyObject *simd__intrin_cvt_f64_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_f64_b64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vf64 = _mm_castsi128_pd( arg.data.vb64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7430 | SIMD_IMPL_INTRIN_1(cvt_b64_f64, vb64, vf64)static PyObject *simd__intrin_cvt_b64_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "cvt_b64_f64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vb64 = _mm_castpd_si128( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7431 | #if 0 | |||
| 7432 | SIMD_IMPL_INTRIN_1(expand_f64_f64, vf64x2, vf64)static PyObject *simd__intrin_expand_f64_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "expand_f64_f64", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vf64x2 = npyv_expand_f64_f64( arg.data .vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vf64x2 }; return simd_arg_to_obj(&ret ); } | |||
| 7433 | #endif // expand_sup | |||
| 7434 | /*************************** | |||
| 7435 | * Arithmetic | |||
| 7436 | ***************************/ | |||
| 7437 | #line 339 | |||
| 7438 | SIMD_IMPL_INTRIN_2(add_f64, vf64, vf64, vf64)static PyObject *simd__intrin_add_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""add_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_add_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7439 | ||||
| 7440 | #line 339 | |||
| 7441 | SIMD_IMPL_INTRIN_2(sub_f64, vf64, vf64, vf64)static PyObject *simd__intrin_sub_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""sub_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_sub_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7442 | ||||
| 7443 | ||||
| 7444 | #if 0 | |||
| 7445 | #line 346 | |||
| 7446 | SIMD_IMPL_INTRIN_2(adds_f64, vf64, vf64, vf64)static PyObject *simd__intrin_adds_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""adds_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_adds_f64( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7447 | ||||
| 7448 | #line 346 | |||
| 7449 | SIMD_IMPL_INTRIN_2(subs_f64, vf64, vf64, vf64)static PyObject *simd__intrin_subs_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""subs_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_subs_f64( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7450 | ||||
| 7451 | #endif // sat_sup | |||
| 7452 | ||||
| 7453 | #if 1 | |||
| 7454 | SIMD_IMPL_INTRIN_2(mul_f64, vf64, vf64, vf64)static PyObject *simd__intrin_mul_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""mul_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_mul_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7455 | #endif // mul_sup | |||
| 7456 | ||||
| 7457 | #if 1 | |||
| 7458 | SIMD_IMPL_INTRIN_2(div_f64, vf64, vf64, vf64)static PyObject *simd__intrin_div_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""div_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_div_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7459 | #endif // div_sup | |||
| 7460 | ||||
| 7461 | #if 0 | |||
| 7462 | SIMD_IMPL_INTRIN_1(divisor_f64, vf64x3, f64)static PyObject *simd__intrin_divisor_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_f64}; if (!PyArg_ParseTuple( args, "O&:" "divisor_f64", simd_arg_converter, &arg )) return ((void* )0); simd_data data = {.vf64x3 = npyv_divisor_f64( arg.data.f64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vf64x3 }; return simd_arg_to_obj(&ret); } | |||
| 7463 | SIMD_IMPL_INTRIN_2(divc_f64, vf64, vf64, vf64x3)static PyObject *simd__intrin_divc_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64x3 }; if (!PyArg_ParseTuple( args, "O&O&:""divc_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_divc_f64( arg1.data.vf64, arg2 .data.vf64x3 )}; simd_arg_free(&arg1); simd_arg_free(& arg2); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7464 | #endif // intdiv_sup | |||
| 7465 | ||||
| 7466 | #if 1 | |||
| 7467 | #line 367 | |||
| 7468 | SIMD_IMPL_INTRIN_3(muladd_f64, vf64, vf64, vf64, vf64)static PyObject *simd__intrin_muladd_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""muladd_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_muladd_f64 ( arg1.data.vf64, arg2.data.vf64, arg3.data.vf64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7469 | ||||
| 7470 | #line 367 | |||
| 7471 | SIMD_IMPL_INTRIN_3(mulsub_f64, vf64, vf64, vf64, vf64)static PyObject *simd__intrin_mulsub_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""mulsub_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_mulsub_f64 ( arg1.data.vf64, arg2.data.vf64, arg3.data.vf64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7472 | ||||
| 7473 | #line 367 | |||
| 7474 | SIMD_IMPL_INTRIN_3(nmuladd_f64, vf64, vf64, vf64, vf64)static PyObject *simd__intrin_nmuladd_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmuladd_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_nmuladd_f64 ( arg1.data.vf64, arg2.data.vf64, arg3.data.vf64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7475 | ||||
| 7476 | #line 367 | |||
| 7477 | SIMD_IMPL_INTRIN_3(nmulsub_f64, vf64, vf64, vf64, vf64)static PyObject *simd__intrin_nmulsub_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple ( args, "O&O&O&:""nmulsub_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2, simd_arg_converter , &arg3 )) return ((void*)0); simd_data data = {.vf64 = npyv_nmulsub_f64 ( arg1.data.vf64, arg2.data.vf64, arg3.data.vf64 )}; simd_arg_free (&arg1); simd_arg_free(&arg2); simd_arg_free(&arg3 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7478 | ||||
| 7479 | #endif // fused_sup | |||
| 7480 | ||||
| 7481 | #if 1 | |||
| 7482 | SIMD_IMPL_INTRIN_1(sum_f64, f64, vf64)static PyObject *simd__intrin_sum_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "sum_f64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.f64 = npyv_sum_f64( arg.data.vf64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_f64 }; return simd_arg_to_obj(&ret); } | |||
| 7483 | #endif // sum_sup | |||
| 7484 | ||||
| 7485 | #if 0 | |||
| 7486 | SIMD_IMPL_INTRIN_1(sumup_f64, f64, vf64)static PyObject *simd__intrin_sumup_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "sumup_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.f64 = npyv_sumup_f64( arg.data.vf64 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_f64 }; return simd_arg_to_obj(&ret); } | |||
| 7487 | #endif // sumup_sup | |||
| 7488 | ||||
| 7489 | /*************************** | |||
| 7490 | * Math | |||
| 7491 | ***************************/ | |||
| 7492 | #if 1 | |||
| 7493 | #line 386 | |||
| 7494 | SIMD_IMPL_INTRIN_1(sqrt_f64, vf64, vf64)static PyObject *simd__intrin_sqrt_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "sqrt_f64", simd_arg_converter, &arg )) return ((void*)0) ; simd_data data = {.vf64 = _mm_sqrt_pd( arg.data.vf64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7495 | ||||
| 7496 | #line 386 | |||
| 7497 | SIMD_IMPL_INTRIN_1(recip_f64, vf64, vf64)static PyObject *simd__intrin_recip_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "recip_f64", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.vf64 = npyv_recip_f64( arg.data.vf64 )} ; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7498 | ||||
| 7499 | #line 386 | |||
| 7500 | SIMD_IMPL_INTRIN_1(abs_f64, vf64, vf64)static PyObject *simd__intrin_abs_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "abs_f64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vf64 = npyv_abs_f64( arg.data.vf64 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7501 | ||||
| 7502 | #line 386 | |||
| 7503 | SIMD_IMPL_INTRIN_1(square_f64, vf64, vf64)static PyObject *simd__intrin_square_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "square_f64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vf64 = npyv_square_f64( arg.data.vf64 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7504 | ||||
| 7505 | #endif | |||
| 7506 | ||||
| 7507 | #line 393 | |||
| 7508 | SIMD_IMPL_INTRIN_2(max_f64, vf64, vf64, vf64)static PyObject *simd__intrin_max_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""max_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_max_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7509 | ||||
| 7510 | #line 393 | |||
| 7511 | SIMD_IMPL_INTRIN_2(min_f64, vf64, vf64, vf64)static PyObject *simd__intrin_min_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""min_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = _mm_min_pd( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7512 | ||||
| 7513 | ||||
| 7514 | #if 1 | |||
| 7515 | #line 400 | |||
| 7516 | SIMD_IMPL_INTRIN_2(maxp_f64, vf64, vf64, vf64)static PyObject *simd__intrin_maxp_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""maxp_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_maxp_f64( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7517 | ||||
| 7518 | #line 400 | |||
| 7519 | SIMD_IMPL_INTRIN_2(minp_f64, vf64, vf64, vf64)static PyObject *simd__intrin_minp_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""minp_f64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vf64 = npyv_minp_f64( arg1.data.vf64, arg2 .data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(&ret); } | |||
| 7520 | ||||
| 7521 | #endif | |||
| 7522 | ||||
| 7523 | /*************************** | |||
| 7524 | * Mask operations | |||
| 7525 | ***************************/ | |||
| 7526 | #line 410 | |||
| 7527 | SIMD_IMPL_INTRIN_4(ifadd_f64, vf64, vb64, vf64, vf64, vf64)static PyObject *simd__intrin_ifadd_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; simd_arg arg4 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifadd_f64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vf64 = npyv_ifadd_f64 ( arg1.data.vb64, arg2.data.vf64, arg3.data.vf64, arg4.data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(& ret); } | |||
| 7528 | ||||
| 7529 | #line 410 | |||
| 7530 | SIMD_IMPL_INTRIN_4(ifsub_f64, vf64, vb64, vf64, vf64, vf64)static PyObject *simd__intrin_ifsub_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; simd_arg arg3 = {.dtype = simd_data_vf64}; simd_arg arg4 = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&O&O&O&:" "ifsub_f64", simd_arg_converter, &arg1, simd_arg_converter , &arg2, simd_arg_converter, &arg3, simd_arg_converter , &arg4 )) return ((void*)0); simd_data data = {.vf64 = npyv_ifsub_f64 ( arg1.data.vb64, arg2.data.vf64, arg3.data.vf64, arg4.data.vf64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2); simd_arg_free (&arg3); simd_arg_free(&arg4); simd_arg ret = { .data = data, .dtype = simd_data_vf64 }; return simd_arg_to_obj(& ret); } | |||
| 7531 | ||||
| 7532 | ||||
| 7533 | #endif // simd_sup | |||
| 7534 | ||||
| 7535 | /************************************************************************* | |||
| 7536 | * Variant | |||
| 7537 | ************************************************************************/ | |||
| 7538 | SIMD_IMPL_INTRIN_0N(cleanup)static PyObject *simd__intrin_cleanup (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { if (!PyArg_ParseTuple ( args, ":" "cleanup") ) return ((void*)0); ((void)0); return _Py_INCREF(((PyObject*)((&_Py_NoneStruct)))), (&_Py_NoneStruct ); } | |||
| 7539 | ||||
| 7540 | /************************************************************************* | |||
| 7541 | * A special section for f32/f64 intrinsics outside the main repeater | |||
| 7542 | ************************************************************************/ | |||
| 7543 | /*************************** | |||
| 7544 | * Operators | |||
| 7545 | ***************************/ | |||
| 7546 | // check special cases | |||
| 7547 | SIMD_IMPL_INTRIN_1(notnan_f32, vb32, vf32)static PyObject *simd__intrin_notnan_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "notnan_f32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vb32 = npyv_notnan_f32( arg.data.vf32 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 7548 | #if NPY_SIMD_F641 | |||
| 7549 | SIMD_IMPL_INTRIN_1(notnan_f64, vb64, vf64)static PyObject *simd__intrin_notnan_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf64}; if (!PyArg_ParseTuple( args, "O&:" "notnan_f64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.vb64 = npyv_notnan_f64( arg.data.vf64 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7550 | #endif | |||
| 7551 | /*************************** | |||
| 7552 | * Conversions | |||
| 7553 | ***************************/ | |||
| 7554 | // round to nearest integer (assume even) | |||
| 7555 | SIMD_IMPL_INTRIN_1(round_s32_f32, vs32, vf32)static PyObject *simd__intrin_round_s32_f32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vf32}; if (!PyArg_ParseTuple( args, "O&:" "round_s32_f32", simd_arg_converter, &arg )) return ((void *)0); simd_data data = {.vs32 = _mm_cvtps_epi32( arg.data.vf32 )}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 7556 | #if NPY_SIMD_F641 | |||
| 7557 | SIMD_IMPL_INTRIN_2(round_s32_f64, vs32, vf64, vf64)static PyObject *simd__intrin_round_s32_f64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vf64}; simd_arg arg2 = {.dtype = simd_data_vf64 }; if (!PyArg_ParseTuple( args, "O&O&:""round_s32_f64" , simd_arg_converter, &arg1, simd_arg_converter, &arg2 )) return ((void*)0); simd_data data = {.vs32 = npyv_round_s32_f64 ( arg1.data.vf64, arg2.data.vf64 )}; simd_arg_free(&arg1) ; simd_arg_free(&arg2); simd_arg ret = { .data = data, .dtype = simd_data_vs32 }; return simd_arg_to_obj(&ret); } | |||
| 7558 | #endif | |||
| 7559 | ||||
| 7560 | /************************************************************************* | |||
| 7561 | * A special section for boolean intrinsics outside the main repeater | |||
| 7562 | ************************************************************************/ | |||
| 7563 | /*************************** | |||
| 7564 | * Operators | |||
| 7565 | ***************************/ | |||
| 7566 | // Logical | |||
| 7567 | #line 450 | |||
| 7568 | SIMD_IMPL_INTRIN_2(and_b8, vb8, vb8, vb8)static PyObject *simd__intrin_and_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vb8 }; if (!PyArg_ParseTuple( args, "O&O&:""and_b8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_and_si128( arg1.data.vb8, arg2 .data.vb8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 7569 | SIMD_IMPL_INTRIN_2(or_b8, vb8, vb8, vb8)static PyObject *simd__intrin_or_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vb8 }; if (!PyArg_ParseTuple( args, "O&O&:""or_b8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_or_si128( arg1.data.vb8, arg2 .data.vb8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 7570 | SIMD_IMPL_INTRIN_2(xor_b8, vb8, vb8, vb8)static PyObject *simd__intrin_xor_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb8}; simd_arg arg2 = {.dtype = simd_data_vb8 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_b8", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb8 = _mm_xor_si128( arg1.data.vb8, arg2 .data.vb8 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 7571 | SIMD_IMPL_INTRIN_1(not_b8, vb8, vb8)static PyObject *simd__intrin_not_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb8}; if (!PyArg_ParseTuple( args, "O&:" "not_b8", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vb8 = _mm_xor_si128(arg.data.vb8, _mm_set1_epi32(-1 ))}; simd_arg_free(&arg); simd_arg ret = { .data = data, . dtype = simd_data_vb8 }; return simd_arg_to_obj(&ret); } | |||
| 7572 | ||||
| 7573 | #line 450 | |||
| 7574 | SIMD_IMPL_INTRIN_2(and_b16, vb16, vb16, vb16)static PyObject *simd__intrin_and_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vb16 }; if (!PyArg_ParseTuple( args, "O&O&:""and_b16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_and_si128( arg1.data.vb16, arg2 .data.vb16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 7575 | SIMD_IMPL_INTRIN_2(or_b16, vb16, vb16, vb16)static PyObject *simd__intrin_or_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vb16 }; if (!PyArg_ParseTuple( args, "O&O&:""or_b16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_or_si128( arg1.data.vb16, arg2 .data.vb16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 7576 | SIMD_IMPL_INTRIN_2(xor_b16, vb16, vb16, vb16)static PyObject *simd__intrin_xor_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb16}; simd_arg arg2 = {.dtype = simd_data_vb16 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_b16", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb16 = _mm_xor_si128( arg1.data.vb16, arg2 .data.vb16 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret); } | |||
| 7577 | SIMD_IMPL_INTRIN_1(not_b16, vb16, vb16)static PyObject *simd__intrin_not_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb16}; if (!PyArg_ParseTuple( args, "O&:" "not_b16", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vb16 = _mm_xor_si128(arg.data.vb16, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vb16 }; return simd_arg_to_obj(&ret) ; } | |||
| 7578 | ||||
| 7579 | #line 450 | |||
| 7580 | SIMD_IMPL_INTRIN_2(and_b32, vb32, vb32, vb32)static PyObject *simd__intrin_and_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vb32 }; if (!PyArg_ParseTuple( args, "O&O&:""and_b32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_and_si128( arg1.data.vb32, arg2 .data.vb32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 7581 | SIMD_IMPL_INTRIN_2(or_b32, vb32, vb32, vb32)static PyObject *simd__intrin_or_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vb32 }; if (!PyArg_ParseTuple( args, "O&O&:""or_b32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_or_si128( arg1.data.vb32, arg2 .data.vb32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 7582 | SIMD_IMPL_INTRIN_2(xor_b32, vb32, vb32, vb32)static PyObject *simd__intrin_xor_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb32}; simd_arg arg2 = {.dtype = simd_data_vb32 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_b32", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb32 = _mm_xor_si128( arg1.data.vb32, arg2 .data.vb32 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret); } | |||
| 7583 | SIMD_IMPL_INTRIN_1(not_b32, vb32, vb32)static PyObject *simd__intrin_not_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb32}; if (!PyArg_ParseTuple( args, "O&:" "not_b32", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vb32 = _mm_xor_si128(arg.data.vb32, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vb32 }; return simd_arg_to_obj(&ret) ; } | |||
| 7584 | ||||
| 7585 | #line 450 | |||
| 7586 | SIMD_IMPL_INTRIN_2(and_b64, vb64, vb64, vb64)static PyObject *simd__intrin_and_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vb64 }; if (!PyArg_ParseTuple( args, "O&O&:""and_b64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_and_si128( arg1.data.vb64, arg2 .data.vb64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7587 | SIMD_IMPL_INTRIN_2(or_b64, vb64, vb64, vb64)static PyObject *simd__intrin_or_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vb64 }; if (!PyArg_ParseTuple( args, "O&O&:""or_b64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_or_si128( arg1.data.vb64, arg2 .data.vb64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7588 | SIMD_IMPL_INTRIN_2(xor_b64, vb64, vb64, vb64)static PyObject *simd__intrin_xor_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg1 = {.dtype = simd_data_vb64}; simd_arg arg2 = {.dtype = simd_data_vb64 }; if (!PyArg_ParseTuple( args, "O&O&:""xor_b64", simd_arg_converter , &arg1, simd_arg_converter, &arg2 )) return ((void*) 0); simd_data data = {.vb64 = _mm_xor_si128( arg1.data.vb64, arg2 .data.vb64 )}; simd_arg_free(&arg1); simd_arg_free(&arg2 ); simd_arg ret = { .data = data, .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret); } | |||
| 7589 | SIMD_IMPL_INTRIN_1(not_b64, vb64, vb64)static PyObject *simd__intrin_not_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb64}; if (!PyArg_ParseTuple( args, "O&:" "not_b64", simd_arg_converter, &arg )) return ((void*)0); simd_data data = {.vb64 = _mm_xor_si128(arg.data.vb64, _mm_set1_epi32 (-1))}; simd_arg_free(&arg); simd_arg ret = { .data = data , .dtype = simd_data_vb64 }; return simd_arg_to_obj(&ret) ; } | |||
| 7590 | ||||
| 7591 | /*************************** | |||
| 7592 | * Conversions | |||
| 7593 | ***************************/ | |||
| 7594 | // Convert mask vector to integer bitfield | |||
| 7595 | #line 462 | |||
| 7596 | SIMD_IMPL_INTRIN_1(tobits_b8, u64, vb8)static PyObject *simd__intrin_tobits_b8 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb8}; if (!PyArg_ParseTuple( args, "O&:" "tobits_b8", simd_arg_converter, &arg )) return ((void*)0 ); simd_data data = {.u64 = npyv_tobits_b8( arg.data.vb8 )}; simd_arg_free (&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 7597 | ||||
| 7598 | #line 462 | |||
| 7599 | SIMD_IMPL_INTRIN_1(tobits_b16, u64, vb16)static PyObject *simd__intrin_tobits_b16 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb16}; if (!PyArg_ParseTuple( args, "O&:" "tobits_b16", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.u64 = npyv_tobits_b16( arg.data.vb16 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 7600 | ||||
| 7601 | #line 462 | |||
| 7602 | SIMD_IMPL_INTRIN_1(tobits_b32, u64, vb32)static PyObject *simd__intrin_tobits_b32 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb32}; if (!PyArg_ParseTuple( args, "O&:" "tobits_b32", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.u64 = npyv_tobits_b32( arg.data.vb32 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 7603 | ||||
| 7604 | #line 462 | |||
| 7605 | SIMD_IMPL_INTRIN_1(tobits_b64, u64, vb64)static PyObject *simd__intrin_tobits_b64 (PyObject* (__NPY_UNUSED_TAGGEDself ) __attribute__ ((__unused__)), PyObject *args) { simd_arg arg = {.dtype = simd_data_vb64}; if (!PyArg_ParseTuple( args, "O&:" "tobits_b64", simd_arg_converter, &arg )) return ((void*) 0); simd_data data = {.u64 = npyv_tobits_b64( arg.data.vb64 ) }; simd_arg_free(&arg); simd_arg ret = { .data = data, .dtype = simd_data_u64 }; return simd_arg_to_obj(&ret); } | |||
| 7606 | ||||
| 7607 | ||||
| 7608 | ||||
| 7609 | //######################################################################### | |||
| 7610 | //## Attach module functions | |||
| 7611 | //######################################################################### | |||
| 7612 | static PyMethodDef simd__intrinsics_methods[] = { | |||
| 7613 | #line 489 | |||
| 7614 | #if 1 | |||
| 7615 | ||||
| 7616 | /*************************** | |||
| 7617 | * Memory | |||
| 7618 | ***************************/ | |||
| 7619 | #line 497 | |||
| 7620 | SIMD_INTRIN_DEF(load_u8){ "load_u8", simd__intrin_load_u8, 0x0001, ((void*)0) } , | |||
| 7621 | ||||
| 7622 | #line 497 | |||
| 7623 | SIMD_INTRIN_DEF(loada_u8){ "loada_u8", simd__intrin_loada_u8, 0x0001, ((void*)0) } , | |||
| 7624 | ||||
| 7625 | #line 497 | |||
| 7626 | SIMD_INTRIN_DEF(loads_u8){ "loads_u8", simd__intrin_loads_u8, 0x0001, ((void*)0) } , | |||
| 7627 | ||||
| 7628 | #line 497 | |||
| 7629 | SIMD_INTRIN_DEF(loadl_u8){ "loadl_u8", simd__intrin_loadl_u8, 0x0001, ((void*)0) } , | |||
| 7630 | ||||
| 7631 | #line 497 | |||
| 7632 | SIMD_INTRIN_DEF(store_u8){ "store_u8", simd__intrin_store_u8, 0x0001, ((void*)0) } , | |||
| 7633 | ||||
| 7634 | #line 497 | |||
| 7635 | SIMD_INTRIN_DEF(storea_u8){ "storea_u8", simd__intrin_storea_u8, 0x0001, ((void*)0) } , | |||
| 7636 | ||||
| 7637 | #line 497 | |||
| 7638 | SIMD_INTRIN_DEF(stores_u8){ "stores_u8", simd__intrin_stores_u8, 0x0001, ((void*)0) } , | |||
| 7639 | ||||
| 7640 | #line 497 | |||
| 7641 | SIMD_INTRIN_DEF(storel_u8){ "storel_u8", simd__intrin_storel_u8, 0x0001, ((void*)0) } , | |||
| 7642 | ||||
| 7643 | #line 497 | |||
| 7644 | SIMD_INTRIN_DEF(storeh_u8){ "storeh_u8", simd__intrin_storeh_u8, 0x0001, ((void*)0) } , | |||
| 7645 | ||||
| 7646 | ||||
| 7647 | /**************************************** | |||
| 7648 | * Non-contiguous/Partial Memory access | |||
| 7649 | ****************************************/ | |||
| 7650 | #if 0 | |||
| 7651 | #line 508 | |||
| 7652 | SIMD_INTRIN_DEF(load_till_u8){ "load_till_u8", simd__intrin_load_till_u8, 0x0001, ((void*) 0) } , | |||
| 7653 | ||||
| 7654 | #line 508 | |||
| 7655 | SIMD_INTRIN_DEF(load_tillz_u8){ "load_tillz_u8", simd__intrin_load_tillz_u8, 0x0001, ((void *)0) } , | |||
| 7656 | ||||
| 7657 | #line 508 | |||
| 7658 | SIMD_INTRIN_DEF(loadn_u8){ "loadn_u8", simd__intrin_loadn_u8, 0x0001, ((void*)0) } , | |||
| 7659 | ||||
| 7660 | #line 508 | |||
| 7661 | SIMD_INTRIN_DEF(loadn_till_u8){ "loadn_till_u8", simd__intrin_loadn_till_u8, 0x0001, ((void *)0) } , | |||
| 7662 | ||||
| 7663 | #line 508 | |||
| 7664 | SIMD_INTRIN_DEF(loadn_tillz_u8){ "loadn_tillz_u8", simd__intrin_loadn_tillz_u8, 0x0001, ((void *)0) } , | |||
| 7665 | ||||
| 7666 | #line 508 | |||
| 7667 | SIMD_INTRIN_DEF(store_till_u8){ "store_till_u8", simd__intrin_store_till_u8, 0x0001, ((void *)0) } , | |||
| 7668 | ||||
| 7669 | #line 508 | |||
| 7670 | SIMD_INTRIN_DEF(storen_u8){ "storen_u8", simd__intrin_storen_u8, 0x0001, ((void*)0) } , | |||
| 7671 | ||||
| 7672 | #line 508 | |||
| 7673 | SIMD_INTRIN_DEF(storen_till_u8){ "storen_till_u8", simd__intrin_storen_till_u8, 0x0001, ((void *)0) } , | |||
| 7674 | ||||
| 7675 | #endif // ncont_sup | |||
| 7676 | ||||
| 7677 | /*************************** | |||
| 7678 | * Misc | |||
| 7679 | ***************************/ | |||
| 7680 | #line 519 | |||
| 7681 | #if 1 | |||
| 7682 | SIMD_INTRIN_DEF(reinterpret_u8_u8){ "reinterpret_u8_u8", simd__intrin_reinterpret_u8_u8, 0x0001 , ((void*)0) } , | |||
| 7683 | #endif // simd_sup2 | |||
| 7684 | ||||
| 7685 | #line 519 | |||
| 7686 | #if 1 | |||
| 7687 | SIMD_INTRIN_DEF(reinterpret_s8_u8){ "reinterpret_s8_u8", simd__intrin_reinterpret_s8_u8, 0x0001 , ((void*)0) } , | |||
| 7688 | #endif // simd_sup2 | |||
| 7689 | ||||
| 7690 | #line 519 | |||
| 7691 | #if 1 | |||
| 7692 | SIMD_INTRIN_DEF(reinterpret_u16_u8){ "reinterpret_u16_u8", simd__intrin_reinterpret_u16_u8, 0x0001 , ((void*)0) } , | |||
| 7693 | #endif // simd_sup2 | |||
| 7694 | ||||
| 7695 | #line 519 | |||
| 7696 | #if 1 | |||
| 7697 | SIMD_INTRIN_DEF(reinterpret_s16_u8){ "reinterpret_s16_u8", simd__intrin_reinterpret_s16_u8, 0x0001 , ((void*)0) } , | |||
| 7698 | #endif // simd_sup2 | |||
| 7699 | ||||
| 7700 | #line 519 | |||
| 7701 | #if 1 | |||
| 7702 | SIMD_INTRIN_DEF(reinterpret_u32_u8){ "reinterpret_u32_u8", simd__intrin_reinterpret_u32_u8, 0x0001 , ((void*)0) } , | |||
| 7703 | #endif // simd_sup2 | |||
| 7704 | ||||
| 7705 | #line 519 | |||
| 7706 | #if 1 | |||
| 7707 | SIMD_INTRIN_DEF(reinterpret_s32_u8){ "reinterpret_s32_u8", simd__intrin_reinterpret_s32_u8, 0x0001 , ((void*)0) } , | |||
| 7708 | #endif // simd_sup2 | |||
| 7709 | ||||
| 7710 | #line 519 | |||
| 7711 | #if 1 | |||
| 7712 | SIMD_INTRIN_DEF(reinterpret_u64_u8){ "reinterpret_u64_u8", simd__intrin_reinterpret_u64_u8, 0x0001 , ((void*)0) } , | |||
| 7713 | #endif // simd_sup2 | |||
| 7714 | ||||
| 7715 | #line 519 | |||
| 7716 | #if 1 | |||
| 7717 | SIMD_INTRIN_DEF(reinterpret_s64_u8){ "reinterpret_s64_u8", simd__intrin_reinterpret_s64_u8, 0x0001 , ((void*)0) } , | |||
| 7718 | #endif // simd_sup2 | |||
| 7719 | ||||
| 7720 | #line 519 | |||
| 7721 | #if 1 | |||
| 7722 | SIMD_INTRIN_DEF(reinterpret_f32_u8){ "reinterpret_f32_u8", simd__intrin_reinterpret_f32_u8, 0x0001 , ((void*)0) } , | |||
| 7723 | #endif // simd_sup2 | |||
| 7724 | ||||
| 7725 | #line 519 | |||
| 7726 | #if NPY_SIMD_F641 | |||
| 7727 | SIMD_INTRIN_DEF(reinterpret_f64_u8){ "reinterpret_f64_u8", simd__intrin_reinterpret_f64_u8, 0x0001 , ((void*)0) } , | |||
| 7728 | #endif // simd_sup2 | |||
| 7729 | ||||
| 7730 | ||||
| 7731 | #line 527 | |||
| 7732 | SIMD_INTRIN_DEF(set_u8){ "set_u8", simd__intrin_set_u8, 0x0001, ((void*)0) } , | |||
| 7733 | ||||
| 7734 | #line 527 | |||
| 7735 | SIMD_INTRIN_DEF(setf_u8){ "setf_u8", simd__intrin_setf_u8, 0x0001, ((void*)0) } , | |||
| 7736 | ||||
| 7737 | #line 527 | |||
| 7738 | SIMD_INTRIN_DEF(setall_u8){ "setall_u8", simd__intrin_setall_u8, 0x0001, ((void*)0) } , | |||
| 7739 | ||||
| 7740 | #line 527 | |||
| 7741 | SIMD_INTRIN_DEF(zero_u8){ "zero_u8", simd__intrin_zero_u8, 0x0001, ((void*)0) } , | |||
| 7742 | ||||
| 7743 | #line 527 | |||
| 7744 | SIMD_INTRIN_DEF(select_u8){ "select_u8", simd__intrin_select_u8, 0x0001, ((void*)0) } , | |||
| 7745 | ||||
| 7746 | ||||
| 7747 | /*************************** | |||
| 7748 | * Reorder | |||
| 7749 | ***************************/ | |||
| 7750 | #line 536 | |||
| 7751 | SIMD_INTRIN_DEF(combinel_u8){ "combinel_u8", simd__intrin_combinel_u8, 0x0001, ((void*)0) } , | |||
| 7752 | ||||
| 7753 | #line 536 | |||
| 7754 | SIMD_INTRIN_DEF(combineh_u8){ "combineh_u8", simd__intrin_combineh_u8, 0x0001, ((void*)0) } , | |||
| 7755 | ||||
| 7756 | #line 536 | |||
| 7757 | SIMD_INTRIN_DEF(combine_u8){ "combine_u8", simd__intrin_combine_u8, 0x0001, ((void*)0) } , | |||
| 7758 | ||||
| 7759 | #line 536 | |||
| 7760 | SIMD_INTRIN_DEF(zip_u8){ "zip_u8", simd__intrin_zip_u8, 0x0001, ((void*)0) } , | |||
| 7761 | ||||
| 7762 | ||||
| 7763 | #if 1 | |||
| 7764 | SIMD_INTRIN_DEF(rev64_u8){ "rev64_u8", simd__intrin_rev64_u8, 0x0001, ((void*)0) } , | |||
| 7765 | #endif | |||
| 7766 | ||||
| 7767 | /*************************** | |||
| 7768 | * Operators | |||
| 7769 | ***************************/ | |||
| 7770 | #if 0 > 0 | |||
| 7771 | #line 550 | |||
| 7772 | SIMD_INTRIN_DEF(shl_u8){ "shl_u8", simd__intrin_shl_u8, 0x0001, ((void*)0) } , | |||
| 7773 | ||||
| 7774 | #line 550 | |||
| 7775 | SIMD_INTRIN_DEF(shr_u8){ "shr_u8", simd__intrin_shr_u8, 0x0001, ((void*)0) } , | |||
| 7776 | ||||
| 7777 | #line 550 | |||
| 7778 | SIMD_INTRIN_DEF(shli_u8){ "shli_u8", simd__intrin_shli_u8, 0x0001, ((void*)0) } , | |||
| 7779 | ||||
| 7780 | #line 550 | |||
| 7781 | SIMD_INTRIN_DEF(shri_u8){ "shri_u8", simd__intrin_shri_u8, 0x0001, ((void*)0) } , | |||
| 7782 | ||||
| 7783 | #endif // shl_imm | |||
| 7784 | ||||
| 7785 | #line 557 | |||
| 7786 | SIMD_INTRIN_DEF(and_u8){ "and_u8", simd__intrin_and_u8, 0x0001, ((void*)0) } , | |||
| 7787 | ||||
| 7788 | #line 557 | |||
| 7789 | SIMD_INTRIN_DEF(or_u8){ "or_u8", simd__intrin_or_u8, 0x0001, ((void*)0) } , | |||
| 7790 | ||||
| 7791 | #line 557 | |||
| 7792 | SIMD_INTRIN_DEF(xor_u8){ "xor_u8", simd__intrin_xor_u8, 0x0001, ((void*)0) } , | |||
| 7793 | ||||
| 7794 | #line 557 | |||
| 7795 | SIMD_INTRIN_DEF(not_u8){ "not_u8", simd__intrin_not_u8, 0x0001, ((void*)0) } , | |||
| 7796 | ||||
| 7797 | #line 557 | |||
| 7798 | SIMD_INTRIN_DEF(cmpeq_u8){ "cmpeq_u8", simd__intrin_cmpeq_u8, 0x0001, ((void*)0) } , | |||
| 7799 | ||||
| 7800 | #line 557 | |||
| 7801 | SIMD_INTRIN_DEF(cmpneq_u8){ "cmpneq_u8", simd__intrin_cmpneq_u8, 0x0001, ((void*)0) } , | |||
| 7802 | ||||
| 7803 | #line 557 | |||
| 7804 | SIMD_INTRIN_DEF(cmpgt_u8){ "cmpgt_u8", simd__intrin_cmpgt_u8, 0x0001, ((void*)0) } , | |||
| 7805 | ||||
| 7806 | #line 557 | |||
| 7807 | SIMD_INTRIN_DEF(cmpge_u8){ "cmpge_u8", simd__intrin_cmpge_u8, 0x0001, ((void*)0) } , | |||
| 7808 | ||||
| 7809 | #line 557 | |||
| 7810 | SIMD_INTRIN_DEF(cmplt_u8){ "cmplt_u8", simd__intrin_cmplt_u8, 0x0001, ((void*)0) } , | |||
| 7811 | ||||
| 7812 | #line 557 | |||
| 7813 | SIMD_INTRIN_DEF(cmple_u8){ "cmple_u8", simd__intrin_cmple_u8, 0x0001, ((void*)0) } , | |||
| 7814 | ||||
| 7815 | ||||
| 7816 | /*************************** | |||
| 7817 | * Conversion | |||
| 7818 | ***************************/ | |||
| 7819 | SIMD_INTRIN_DEF(cvt_u8_b8){ "cvt_u8_b8", simd__intrin_cvt_u8_b8, 0x0001, ((void*)0) } , | |||
| 7820 | SIMD_INTRIN_DEF(cvt_b8_u8){ "cvt_b8_u8", simd__intrin_cvt_b8_u8, 0x0001, ((void*)0) } , | |||
| 7821 | #if 1 | |||
| 7822 | SIMD_INTRIN_DEF(expand_u16_u8){ "expand_u16_u8", simd__intrin_expand_u16_u8, 0x0001, ((void *)0) } , | |||
| 7823 | #endif // expand_sup | |||
| 7824 | /*************************** | |||
| 7825 | * Arithmetic | |||
| 7826 | ***************************/ | |||
| 7827 | #line 574 | |||
| 7828 | SIMD_INTRIN_DEF(add_u8){ "add_u8", simd__intrin_add_u8, 0x0001, ((void*)0) } , | |||
| 7829 | ||||
| 7830 | #line 574 | |||
| 7831 | SIMD_INTRIN_DEF(sub_u8){ "sub_u8", simd__intrin_sub_u8, 0x0001, ((void*)0) } , | |||
| 7832 | ||||
| 7833 | ||||
| 7834 | #if 1 | |||
| 7835 | #line 581 | |||
| 7836 | SIMD_INTRIN_DEF(adds_u8){ "adds_u8", simd__intrin_adds_u8, 0x0001, ((void*)0) } , | |||
| 7837 | ||||
| 7838 | #line 581 | |||
| 7839 | SIMD_INTRIN_DEF(subs_u8){ "subs_u8", simd__intrin_subs_u8, 0x0001, ((void*)0) } , | |||
| 7840 | ||||
| 7841 | #endif // sat_sup | |||
| 7842 | ||||
| 7843 | #if 1 | |||
| 7844 | SIMD_INTRIN_DEF(mul_u8){ "mul_u8", simd__intrin_mul_u8, 0x0001, ((void*)0) } , | |||
| 7845 | #endif // mul_sup | |||
| 7846 | ||||
| 7847 | #if 0 | |||
| 7848 | SIMD_INTRIN_DEF(div_u8){ "div_u8", simd__intrin_div_u8, 0x0001, ((void*)0) } , | |||
| 7849 | #endif // div_sup | |||
| 7850 | ||||
| 7851 | #if 1 | |||
| 7852 | SIMD_INTRIN_DEF(divisor_u8){ "divisor_u8", simd__intrin_divisor_u8, 0x0001, ((void*)0) } , | |||
| 7853 | SIMD_INTRIN_DEF(divc_u8){ "divc_u8", simd__intrin_divc_u8, 0x0001, ((void*)0) } , | |||
| 7854 | #endif // intdiv_sup | |||
| 7855 | ||||
| 7856 | #if 0 | |||
| 7857 | #line 602 | |||
| 7858 | SIMD_INTRIN_DEF(muladd_u8){ "muladd_u8", simd__intrin_muladd_u8, 0x0001, ((void*)0) } , | |||
| 7859 | ||||
| 7860 | #line 602 | |||
| 7861 | SIMD_INTRIN_DEF(mulsub_u8){ "mulsub_u8", simd__intrin_mulsub_u8, 0x0001, ((void*)0) } , | |||
| 7862 | ||||
| 7863 | #line 602 | |||
| 7864 | SIMD_INTRIN_DEF(nmuladd_u8){ "nmuladd_u8", simd__intrin_nmuladd_u8, 0x0001, ((void*)0) } , | |||
| 7865 | ||||
| 7866 | #line 602 | |||
| 7867 | SIMD_INTRIN_DEF(nmulsub_u8){ "nmulsub_u8", simd__intrin_nmulsub_u8, 0x0001, ((void*)0) } , | |||
| 7868 | ||||
| 7869 | #endif // fused_sup | |||
| 7870 | ||||
| 7871 | #if 0 | |||
| 7872 | SIMD_INTRIN_DEF(sum_u8){ "sum_u8", simd__intrin_sum_u8, 0x0001, ((void*)0) } , | |||
| 7873 | #endif // sum_sup | |||
| 7874 | ||||
| 7875 | #if 1 | |||
| 7876 | SIMD_INTRIN_DEF(sumup_u8){ "sumup_u8", simd__intrin_sumup_u8, 0x0001, ((void*)0) } , | |||
| 7877 | #endif // sumup_sup | |||
| 7878 | /*************************** | |||
| 7879 | * Math | |||
| 7880 | ***************************/ | |||
| 7881 | #if 0 | |||
| 7882 | #line 620 | |||
| 7883 | SIMD_INTRIN_DEF(sqrt_u8){ "sqrt_u8", simd__intrin_sqrt_u8, 0x0001, ((void*)0) } , | |||
| 7884 | ||||
| 7885 | #line 620 | |||
| 7886 | SIMD_INTRIN_DEF(recip_u8){ "recip_u8", simd__intrin_recip_u8, 0x0001, ((void*)0) } , | |||
| 7887 | ||||
| 7888 | #line 620 | |||
| 7889 | SIMD_INTRIN_DEF(abs_u8){ "abs_u8", simd__intrin_abs_u8, 0x0001, ((void*)0) } , | |||
| 7890 | ||||
| 7891 | #line 620 | |||
| 7892 | SIMD_INTRIN_DEF(square_u8){ "square_u8", simd__intrin_square_u8, 0x0001, ((void*)0) } , | |||
| 7893 | ||||
| 7894 | #endif | |||
| 7895 | ||||
| 7896 | #line 627 | |||
| 7897 | SIMD_INTRIN_DEF(max_u8){ "max_u8", simd__intrin_max_u8, 0x0001, ((void*)0) } , | |||
| 7898 | ||||
| 7899 | #line 627 | |||
| 7900 | SIMD_INTRIN_DEF(min_u8){ "min_u8", simd__intrin_min_u8, 0x0001, ((void*)0) } , | |||
| 7901 | ||||
| 7902 | ||||
| 7903 | #if 0 | |||
| 7904 | #line 634 | |||
| 7905 | SIMD_INTRIN_DEF(maxp_u8){ "maxp_u8", simd__intrin_maxp_u8, 0x0001, ((void*)0) } , | |||
| 7906 | ||||
| 7907 | #line 634 | |||
| 7908 | SIMD_INTRIN_DEF(minp_u8){ "minp_u8", simd__intrin_minp_u8, 0x0001, ((void*)0) } , | |||
| 7909 | ||||
| 7910 | #endif | |||
| 7911 | ||||
| 7912 | /*************************** | |||
| 7913 | * Mask operations | |||
| 7914 | ***************************/ | |||
| 7915 | #line 644 | |||
| 7916 | SIMD_INTRIN_DEF(ifadd_u8){ "ifadd_u8", simd__intrin_ifadd_u8, 0x0001, ((void*)0) } , | |||
| 7917 | ||||
| 7918 | #line 644 | |||
| 7919 | SIMD_INTRIN_DEF(ifsub_u8){ "ifsub_u8", simd__intrin_ifsub_u8, 0x0001, ((void*)0) } , | |||
| 7920 | ||||
| 7921 | ||||
| 7922 | #endif // simd_sup | |||
| 7923 | ||||
| 7924 | #line 489 | |||
| 7925 | #if 1 | |||
| 7926 | ||||
| 7927 | /*************************** | |||
| 7928 | * Memory | |||
| 7929 | ***************************/ | |||
| 7930 | #line 497 | |||
| 7931 | SIMD_INTRIN_DEF(load_s8){ "load_s8", simd__intrin_load_s8, 0x0001, ((void*)0) } , | |||
| 7932 | ||||
| 7933 | #line 497 | |||
| 7934 | SIMD_INTRIN_DEF(loada_s8){ "loada_s8", simd__intrin_loada_s8, 0x0001, ((void*)0) } , | |||
| 7935 | ||||
| 7936 | #line 497 | |||
| 7937 | SIMD_INTRIN_DEF(loads_s8){ "loads_s8", simd__intrin_loads_s8, 0x0001, ((void*)0) } , | |||
| 7938 | ||||
| 7939 | #line 497 | |||
| 7940 | SIMD_INTRIN_DEF(loadl_s8){ "loadl_s8", simd__intrin_loadl_s8, 0x0001, ((void*)0) } , | |||
| 7941 | ||||
| 7942 | #line 497 | |||
| 7943 | SIMD_INTRIN_DEF(store_s8){ "store_s8", simd__intrin_store_s8, 0x0001, ((void*)0) } , | |||
| 7944 | ||||
| 7945 | #line 497 | |||
| 7946 | SIMD_INTRIN_DEF(storea_s8){ "storea_s8", simd__intrin_storea_s8, 0x0001, ((void*)0) } , | |||
| 7947 | ||||
| 7948 | #line 497 | |||
| 7949 | SIMD_INTRIN_DEF(stores_s8){ "stores_s8", simd__intrin_stores_s8, 0x0001, ((void*)0) } , | |||
| 7950 | ||||
| 7951 | #line 497 | |||
| 7952 | SIMD_INTRIN_DEF(storel_s8){ "storel_s8", simd__intrin_storel_s8, 0x0001, ((void*)0) } , | |||
| 7953 | ||||
| 7954 | #line 497 | |||
| 7955 | SIMD_INTRIN_DEF(storeh_s8){ "storeh_s8", simd__intrin_storeh_s8, 0x0001, ((void*)0) } , | |||
| 7956 | ||||
| 7957 | ||||
| 7958 | /**************************************** | |||
| 7959 | * Non-contiguous/Partial Memory access | |||
| 7960 | ****************************************/ | |||
| 7961 | #if 0 | |||
| 7962 | #line 508 | |||
| 7963 | SIMD_INTRIN_DEF(load_till_s8){ "load_till_s8", simd__intrin_load_till_s8, 0x0001, ((void*) 0) } , | |||
| 7964 | ||||
| 7965 | #line 508 | |||
| 7966 | SIMD_INTRIN_DEF(load_tillz_s8){ "load_tillz_s8", simd__intrin_load_tillz_s8, 0x0001, ((void *)0) } , | |||
| 7967 | ||||
| 7968 | #line 508 | |||
| 7969 | SIMD_INTRIN_DEF(loadn_s8){ "loadn_s8", simd__intrin_loadn_s8, 0x0001, ((void*)0) } , | |||
| 7970 | ||||
| 7971 | #line 508 | |||
| 7972 | SIMD_INTRIN_DEF(loadn_till_s8){ "loadn_till_s8", simd__intrin_loadn_till_s8, 0x0001, ((void *)0) } , | |||
| 7973 | ||||
| 7974 | #line 508 | |||
| 7975 | SIMD_INTRIN_DEF(loadn_tillz_s8){ "loadn_tillz_s8", simd__intrin_loadn_tillz_s8, 0x0001, ((void *)0) } , | |||
| 7976 | ||||
| 7977 | #line 508 | |||
| 7978 | SIMD_INTRIN_DEF(store_till_s8){ "store_till_s8", simd__intrin_store_till_s8, 0x0001, ((void *)0) } , | |||
| 7979 | ||||
| 7980 | #line 508 | |||
| 7981 | SIMD_INTRIN_DEF(storen_s8){ "storen_s8", simd__intrin_storen_s8, 0x0001, ((void*)0) } , | |||
| 7982 | ||||
| 7983 | #line 508 | |||
| 7984 | SIMD_INTRIN_DEF(storen_till_s8){ "storen_till_s8", simd__intrin_storen_till_s8, 0x0001, ((void *)0) } , | |||
| 7985 | ||||
| 7986 | #endif // ncont_sup | |||
| 7987 | ||||
| 7988 | /*************************** | |||
| 7989 | * Misc | |||
| 7990 | ***************************/ | |||
| 7991 | #line 519 | |||
| 7992 | #if 1 | |||
| 7993 | SIMD_INTRIN_DEF(reinterpret_u8_s8){ "reinterpret_u8_s8", simd__intrin_reinterpret_u8_s8, 0x0001 , ((void*)0) } , | |||
| 7994 | #endif // simd_sup2 | |||
| 7995 | ||||
| 7996 | #line 519 | |||
| 7997 | #if 1 | |||
| 7998 | SIMD_INTRIN_DEF(reinterpret_s8_s8){ "reinterpret_s8_s8", simd__intrin_reinterpret_s8_s8, 0x0001 , ((void*)0) } , | |||
| 7999 | #endif // simd_sup2 | |||
| 8000 | ||||
| 8001 | #line 519 | |||
| 8002 | #if 1 | |||
| 8003 | SIMD_INTRIN_DEF(reinterpret_u16_s8){ "reinterpret_u16_s8", simd__intrin_reinterpret_u16_s8, 0x0001 , ((void*)0) } , | |||
| 8004 | #endif // simd_sup2 | |||
| 8005 | ||||
| 8006 | #line 519 | |||
| 8007 | #if 1 | |||
| 8008 | SIMD_INTRIN_DEF(reinterpret_s16_s8){ "reinterpret_s16_s8", simd__intrin_reinterpret_s16_s8, 0x0001 , ((void*)0) } , | |||
| 8009 | #endif // simd_sup2 | |||
| 8010 | ||||
| 8011 | #line 519 | |||
| 8012 | #if 1 | |||
| 8013 | SIMD_INTRIN_DEF(reinterpret_u32_s8){ "reinterpret_u32_s8", simd__intrin_reinterpret_u32_s8, 0x0001 , ((void*)0) } , | |||
| 8014 | #endif // simd_sup2 | |||
| 8015 | ||||
| 8016 | #line 519 | |||
| 8017 | #if 1 | |||
| 8018 | SIMD_INTRIN_DEF(reinterpret_s32_s8){ "reinterpret_s32_s8", simd__intrin_reinterpret_s32_s8, 0x0001 , ((void*)0) } , | |||
| 8019 | #endif // simd_sup2 | |||
| 8020 | ||||
| 8021 | #line 519 | |||
| 8022 | #if 1 | |||
| 8023 | SIMD_INTRIN_DEF(reinterpret_u64_s8){ "reinterpret_u64_s8", simd__intrin_reinterpret_u64_s8, 0x0001 , ((void*)0) } , | |||
| 8024 | #endif // simd_sup2 | |||
| 8025 | ||||
| 8026 | #line 519 | |||
| 8027 | #if 1 | |||
| 8028 | SIMD_INTRIN_DEF(reinterpret_s64_s8){ "reinterpret_s64_s8", simd__intrin_reinterpret_s64_s8, 0x0001 , ((void*)0) } , | |||
| 8029 | #endif // simd_sup2 | |||
| 8030 | ||||
| 8031 | #line 519 | |||
| 8032 | #if 1 | |||
| 8033 | SIMD_INTRIN_DEF(reinterpret_f32_s8){ "reinterpret_f32_s8", simd__intrin_reinterpret_f32_s8, 0x0001 , ((void*)0) } , | |||
| 8034 | #endif // simd_sup2 | |||
| 8035 | ||||
| 8036 | #line 519 | |||
| 8037 | #if NPY_SIMD_F641 | |||
| 8038 | SIMD_INTRIN_DEF(reinterpret_f64_s8){ "reinterpret_f64_s8", simd__intrin_reinterpret_f64_s8, 0x0001 , ((void*)0) } , | |||
| 8039 | #endif // simd_sup2 | |||
| 8040 | ||||
| 8041 | ||||
| 8042 | #line 527 | |||
| 8043 | SIMD_INTRIN_DEF(set_s8){ "set_s8", simd__intrin_set_s8, 0x0001, ((void*)0) } , | |||
| 8044 | ||||
| 8045 | #line 527 | |||
| 8046 | SIMD_INTRIN_DEF(setf_s8){ "setf_s8", simd__intrin_setf_s8, 0x0001, ((void*)0) } , | |||
| 8047 | ||||
| 8048 | #line 527 | |||
| 8049 | SIMD_INTRIN_DEF(setall_s8){ "setall_s8", simd__intrin_setall_s8, 0x0001, ((void*)0) } , | |||
| 8050 | ||||
| 8051 | #line 527 | |||
| 8052 | SIMD_INTRIN_DEF(zero_s8){ "zero_s8", simd__intrin_zero_s8, 0x0001, ((void*)0) } , | |||
| 8053 | ||||
| 8054 | #line 527 | |||
| 8055 | SIMD_INTRIN_DEF(select_s8){ "select_s8", simd__intrin_select_s8, 0x0001, ((void*)0) } , | |||
| 8056 | ||||
| 8057 | ||||
| 8058 | /*************************** | |||
| 8059 | * Reorder | |||
| 8060 | ***************************/ | |||
| 8061 | #line 536 | |||
| 8062 | SIMD_INTRIN_DEF(combinel_s8){ "combinel_s8", simd__intrin_combinel_s8, 0x0001, ((void*)0) } , | |||
| 8063 | ||||
| 8064 | #line 536 | |||
| 8065 | SIMD_INTRIN_DEF(combineh_s8){ "combineh_s8", simd__intrin_combineh_s8, 0x0001, ((void*)0) } , | |||
| 8066 | ||||
| 8067 | #line 536 | |||
| 8068 | SIMD_INTRIN_DEF(combine_s8){ "combine_s8", simd__intrin_combine_s8, 0x0001, ((void*)0) } , | |||
| 8069 | ||||
| 8070 | #line 536 | |||
| 8071 | SIMD_INTRIN_DEF(zip_s8){ "zip_s8", simd__intrin_zip_s8, 0x0001, ((void*)0) } , | |||
| 8072 | ||||
| 8073 | ||||
| 8074 | #if 1 | |||
| 8075 | SIMD_INTRIN_DEF(rev64_s8){ "rev64_s8", simd__intrin_rev64_s8, 0x0001, ((void*)0) } , | |||
| 8076 | #endif | |||
| 8077 | ||||
| 8078 | /*************************** | |||
| 8079 | * Operators | |||
| 8080 | ***************************/ | |||
| 8081 | #if 0 > 0 | |||
| 8082 | #line 550 | |||
| 8083 | SIMD_INTRIN_DEF(shl_s8){ "shl_s8", simd__intrin_shl_s8, 0x0001, ((void*)0) } , | |||
| 8084 | ||||
| 8085 | #line 550 | |||
| 8086 | SIMD_INTRIN_DEF(shr_s8){ "shr_s8", simd__intrin_shr_s8, 0x0001, ((void*)0) } , | |||
| 8087 | ||||
| 8088 | #line 550 | |||
| 8089 | SIMD_INTRIN_DEF(shli_s8){ "shli_s8", simd__intrin_shli_s8, 0x0001, ((void*)0) } , | |||
| 8090 | ||||
| 8091 | #line 550 | |||
| 8092 | SIMD_INTRIN_DEF(shri_s8){ "shri_s8", simd__intrin_shri_s8, 0x0001, ((void*)0) } , | |||
| 8093 | ||||
| 8094 | #endif // shl_imm | |||
| 8095 | ||||
| 8096 | #line 557 | |||
| 8097 | SIMD_INTRIN_DEF(and_s8){ "and_s8", simd__intrin_and_s8, 0x0001, ((void*)0) } , | |||
| 8098 | ||||
| 8099 | #line 557 | |||
| 8100 | SIMD_INTRIN_DEF(or_s8){ "or_s8", simd__intrin_or_s8, 0x0001, ((void*)0) } , | |||
| 8101 | ||||
| 8102 | #line 557 | |||
| 8103 | SIMD_INTRIN_DEF(xor_s8){ "xor_s8", simd__intrin_xor_s8, 0x0001, ((void*)0) } , | |||
| 8104 | ||||
| 8105 | #line 557 | |||
| 8106 | SIMD_INTRIN_DEF(not_s8){ "not_s8", simd__intrin_not_s8, 0x0001, ((void*)0) } , | |||
| 8107 | ||||
| 8108 | #line 557 | |||
| 8109 | SIMD_INTRIN_DEF(cmpeq_s8){ "cmpeq_s8", simd__intrin_cmpeq_s8, 0x0001, ((void*)0) } , | |||
| 8110 | ||||
| 8111 | #line 557 | |||
| 8112 | SIMD_INTRIN_DEF(cmpneq_s8){ "cmpneq_s8", simd__intrin_cmpneq_s8, 0x0001, ((void*)0) } , | |||
| 8113 | ||||
| 8114 | #line 557 | |||
| 8115 | SIMD_INTRIN_DEF(cmpgt_s8){ "cmpgt_s8", simd__intrin_cmpgt_s8, 0x0001, ((void*)0) } , | |||
| 8116 | ||||
| 8117 | #line 557 | |||
| 8118 | SIMD_INTRIN_DEF(cmpge_s8){ "cmpge_s8", simd__intrin_cmpge_s8, 0x0001, ((void*)0) } , | |||
| 8119 | ||||
| 8120 | #line 557 | |||
| 8121 | SIMD_INTRIN_DEF(cmplt_s8){ "cmplt_s8", simd__intrin_cmplt_s8, 0x0001, ((void*)0) } , | |||
| 8122 | ||||
| 8123 | #line 557 | |||
| 8124 | SIMD_INTRIN_DEF(cmple_s8){ "cmple_s8", simd__intrin_cmple_s8, 0x0001, ((void*)0) } , | |||
| 8125 | ||||
| 8126 | ||||
| 8127 | /*************************** | |||
| 8128 | * Conversion | |||
| 8129 | ***************************/ | |||
| 8130 | SIMD_INTRIN_DEF(cvt_s8_b8){ "cvt_s8_b8", simd__intrin_cvt_s8_b8, 0x0001, ((void*)0) } , | |||
| 8131 | SIMD_INTRIN_DEF(cvt_b8_s8){ "cvt_b8_s8", simd__intrin_cvt_b8_s8, 0x0001, ((void*)0) } , | |||
| 8132 | #if 0 | |||
| 8133 | SIMD_INTRIN_DEF(expand_s8_s8){ "expand_s8_s8", simd__intrin_expand_s8_s8, 0x0001, ((void*) 0) } , | |||
| 8134 | #endif // expand_sup | |||
| 8135 | /*************************** | |||
| 8136 | * Arithmetic | |||
| 8137 | ***************************/ | |||
| 8138 | #line 574 | |||
| 8139 | SIMD_INTRIN_DEF(add_s8){ "add_s8", simd__intrin_add_s8, 0x0001, ((void*)0) } , | |||
| 8140 | ||||
| 8141 | #line 574 | |||
| 8142 | SIMD_INTRIN_DEF(sub_s8){ "sub_s8", simd__intrin_sub_s8, 0x0001, ((void*)0) } , | |||
| 8143 | ||||
| 8144 | ||||
| 8145 | #if 1 | |||
| 8146 | #line 581 | |||
| 8147 | SIMD_INTRIN_DEF(adds_s8){ "adds_s8", simd__intrin_adds_s8, 0x0001, ((void*)0) } , | |||
| 8148 | ||||
| 8149 | #line 581 | |||
| 8150 | SIMD_INTRIN_DEF(subs_s8){ "subs_s8", simd__intrin_subs_s8, 0x0001, ((void*)0) } , | |||
| 8151 | ||||
| 8152 | #endif // sat_sup | |||
| 8153 | ||||
| 8154 | #if 1 | |||
| 8155 | SIMD_INTRIN_DEF(mul_s8){ "mul_s8", simd__intrin_mul_s8, 0x0001, ((void*)0) } , | |||
| 8156 | #endif // mul_sup | |||
| 8157 | ||||
| 8158 | #if 0 | |||
| 8159 | SIMD_INTRIN_DEF(div_s8){ "div_s8", simd__intrin_div_s8, 0x0001, ((void*)0) } , | |||
| 8160 | #endif // div_sup | |||
| 8161 | ||||
| 8162 | #if 1 | |||
| 8163 | SIMD_INTRIN_DEF(divisor_s8){ "divisor_s8", simd__intrin_divisor_s8, 0x0001, ((void*)0) } , | |||
| 8164 | SIMD_INTRIN_DEF(divc_s8){ "divc_s8", simd__intrin_divc_s8, 0x0001, ((void*)0) } , | |||
| 8165 | #endif // intdiv_sup | |||
| 8166 | ||||
| 8167 | #if 0 | |||
| 8168 | #line 602 | |||
| 8169 | SIMD_INTRIN_DEF(muladd_s8){ "muladd_s8", simd__intrin_muladd_s8, 0x0001, ((void*)0) } , | |||
| 8170 | ||||
| 8171 | #line 602 | |||
| 8172 | SIMD_INTRIN_DEF(mulsub_s8){ "mulsub_s8", simd__intrin_mulsub_s8, 0x0001, ((void*)0) } , | |||
| 8173 | ||||
| 8174 | #line 602 | |||
| 8175 | SIMD_INTRIN_DEF(nmuladd_s8){ "nmuladd_s8", simd__intrin_nmuladd_s8, 0x0001, ((void*)0) } , | |||
| 8176 | ||||
| 8177 | #line 602 | |||
| 8178 | SIMD_INTRIN_DEF(nmulsub_s8){ "nmulsub_s8", simd__intrin_nmulsub_s8, 0x0001, ((void*)0) } , | |||
| 8179 | ||||
| 8180 | #endif // fused_sup | |||
| 8181 | ||||
| 8182 | #if 0 | |||
| 8183 | SIMD_INTRIN_DEF(sum_s8){ "sum_s8", simd__intrin_sum_s8, 0x0001, ((void*)0) } , | |||
| 8184 | #endif // sum_sup | |||
| 8185 | ||||
| 8186 | #if 0 | |||
| 8187 | SIMD_INTRIN_DEF(sumup_s8){ "sumup_s8", simd__intrin_sumup_s8, 0x0001, ((void*)0) } , | |||
| 8188 | #endif // sumup_sup | |||
| 8189 | /*************************** | |||
| 8190 | * Math | |||
| 8191 | ***************************/ | |||
| 8192 | #if 0 | |||
| 8193 | #line 620 | |||
| 8194 | SIMD_INTRIN_DEF(sqrt_s8){ "sqrt_s8", simd__intrin_sqrt_s8, 0x0001, ((void*)0) } , | |||
| 8195 | ||||
| 8196 | #line 620 | |||
| 8197 | SIMD_INTRIN_DEF(recip_s8){ "recip_s8", simd__intrin_recip_s8, 0x0001, ((void*)0) } , | |||
| 8198 | ||||
| 8199 | #line 620 | |||
| 8200 | SIMD_INTRIN_DEF(abs_s8){ "abs_s8", simd__intrin_abs_s8, 0x0001, ((void*)0) } , | |||
| 8201 | ||||
| 8202 | #line 620 | |||
| 8203 | SIMD_INTRIN_DEF(square_s8){ "square_s8", simd__intrin_square_s8, 0x0001, ((void*)0) } , | |||
| 8204 | ||||
| 8205 | #endif | |||
| 8206 | ||||
| 8207 | #line 627 | |||
| 8208 | SIMD_INTRIN_DEF(max_s8){ "max_s8", simd__intrin_max_s8, 0x0001, ((void*)0) } , | |||
| 8209 | ||||
| 8210 | #line 627 | |||
| 8211 | SIMD_INTRIN_DEF(min_s8){ "min_s8", simd__intrin_min_s8, 0x0001, ((void*)0) } , | |||
| 8212 | ||||
| 8213 | ||||
| 8214 | #if 0 | |||
| 8215 | #line 634 | |||
| 8216 | SIMD_INTRIN_DEF(maxp_s8){ "maxp_s8", simd__intrin_maxp_s8, 0x0001, ((void*)0) } , | |||
| 8217 | ||||
| 8218 | #line 634 | |||
| 8219 | SIMD_INTRIN_DEF(minp_s8){ "minp_s8", simd__intrin_minp_s8, 0x0001, ((void*)0) } , | |||
| 8220 | ||||
| 8221 | #endif | |||
| 8222 | ||||
| 8223 | /*************************** | |||
| 8224 | * Mask operations | |||
| 8225 | ***************************/ | |||
| 8226 | #line 644 | |||
| 8227 | SIMD_INTRIN_DEF(ifadd_s8){ "ifadd_s8", simd__intrin_ifadd_s8, 0x0001, ((void*)0) } , | |||
| 8228 | ||||
| 8229 | #line 644 | |||
| 8230 | SIMD_INTRIN_DEF(ifsub_s8){ "ifsub_s8", simd__intrin_ifsub_s8, 0x0001, ((void*)0) } , | |||
| 8231 | ||||
| 8232 | ||||
| 8233 | #endif // simd_sup | |||
| 8234 | ||||
| 8235 | #line 489 | |||
| 8236 | #if 1 | |||
| 8237 | ||||
| 8238 | /*************************** | |||
| 8239 | * Memory | |||
| 8240 | ***************************/ | |||
| 8241 | #line 497 | |||
| 8242 | SIMD_INTRIN_DEF(load_u16){ "load_u16", simd__intrin_load_u16, 0x0001, ((void*)0) } , | |||
| 8243 | ||||
| 8244 | #line 497 | |||
| 8245 | SIMD_INTRIN_DEF(loada_u16){ "loada_u16", simd__intrin_loada_u16, 0x0001, ((void*)0) } , | |||
| 8246 | ||||
| 8247 | #line 497 | |||
| 8248 | SIMD_INTRIN_DEF(loads_u16){ "loads_u16", simd__intrin_loads_u16, 0x0001, ((void*)0) } , | |||
| 8249 | ||||
| 8250 | #line 497 | |||
| 8251 | SIMD_INTRIN_DEF(loadl_u16){ "loadl_u16", simd__intrin_loadl_u16, 0x0001, ((void*)0) } , | |||
| 8252 | ||||
| 8253 | #line 497 | |||
| 8254 | SIMD_INTRIN_DEF(store_u16){ "store_u16", simd__intrin_store_u16, 0x0001, ((void*)0) } , | |||
| 8255 | ||||
| 8256 | #line 497 | |||
| 8257 | SIMD_INTRIN_DEF(storea_u16){ "storea_u16", simd__intrin_storea_u16, 0x0001, ((void*)0) } , | |||
| 8258 | ||||
| 8259 | #line 497 | |||
| 8260 | SIMD_INTRIN_DEF(stores_u16){ "stores_u16", simd__intrin_stores_u16, 0x0001, ((void*)0) } , | |||
| 8261 | ||||
| 8262 | #line 497 | |||
| 8263 | SIMD_INTRIN_DEF(storel_u16){ "storel_u16", simd__intrin_storel_u16, 0x0001, ((void*)0) } , | |||
| 8264 | ||||
| 8265 | #line 497 | |||
| 8266 | SIMD_INTRIN_DEF(storeh_u16){ "storeh_u16", simd__intrin_storeh_u16, 0x0001, ((void*)0) } , | |||
| 8267 | ||||
| 8268 | ||||
| 8269 | /**************************************** | |||
| 8270 | * Non-contiguous/Partial Memory access | |||
| 8271 | ****************************************/ | |||
| 8272 | #if 0 | |||
| 8273 | #line 508 | |||
| 8274 | SIMD_INTRIN_DEF(load_till_u16){ "load_till_u16", simd__intrin_load_till_u16, 0x0001, ((void *)0) } , | |||
| 8275 | ||||
| 8276 | #line 508 | |||
| 8277 | SIMD_INTRIN_DEF(load_tillz_u16){ "load_tillz_u16", simd__intrin_load_tillz_u16, 0x0001, ((void *)0) } , | |||
| 8278 | ||||
| 8279 | #line 508 | |||
| 8280 | SIMD_INTRIN_DEF(loadn_u16){ "loadn_u16", simd__intrin_loadn_u16, 0x0001, ((void*)0) } , | |||
| 8281 | ||||
| 8282 | #line 508 | |||
| 8283 | SIMD_INTRIN_DEF(loadn_till_u16){ "loadn_till_u16", simd__intrin_loadn_till_u16, 0x0001, ((void *)0) } , | |||
| 8284 | ||||
| 8285 | #line 508 | |||
| 8286 | SIMD_INTRIN_DEF(loadn_tillz_u16){ "loadn_tillz_u16", simd__intrin_loadn_tillz_u16, 0x0001, (( void*)0) } , | |||
| 8287 | ||||
| 8288 | #line 508 | |||
| 8289 | SIMD_INTRIN_DEF(store_till_u16){ "store_till_u16", simd__intrin_store_till_u16, 0x0001, ((void *)0) } , | |||
| 8290 | ||||
| 8291 | #line 508 | |||
| 8292 | SIMD_INTRIN_DEF(storen_u16){ "storen_u16", simd__intrin_storen_u16, 0x0001, ((void*)0) } , | |||
| 8293 | ||||
| 8294 | #line 508 | |||
| 8295 | SIMD_INTRIN_DEF(storen_till_u16){ "storen_till_u16", simd__intrin_storen_till_u16, 0x0001, (( void*)0) } , | |||
| 8296 | ||||
| 8297 | #endif // ncont_sup | |||
| 8298 | ||||
| 8299 | /*************************** | |||
| 8300 | * Misc | |||
| 8301 | ***************************/ | |||
| 8302 | #line 519 | |||
| 8303 | #if 1 | |||
| 8304 | SIMD_INTRIN_DEF(reinterpret_u8_u16){ "reinterpret_u8_u16", simd__intrin_reinterpret_u8_u16, 0x0001 , ((void*)0) } , | |||
| 8305 | #endif // simd_sup2 | |||
| 8306 | ||||
| 8307 | #line 519 | |||
| 8308 | #if 1 | |||
| 8309 | SIMD_INTRIN_DEF(reinterpret_s8_u16){ "reinterpret_s8_u16", simd__intrin_reinterpret_s8_u16, 0x0001 , ((void*)0) } , | |||
| 8310 | #endif // simd_sup2 | |||
| 8311 | ||||
| 8312 | #line 519 | |||
| 8313 | #if 1 | |||
| 8314 | SIMD_INTRIN_DEF(reinterpret_u16_u16){ "reinterpret_u16_u16", simd__intrin_reinterpret_u16_u16, 0x0001 , ((void*)0) } , | |||
| 8315 | #endif // simd_sup2 | |||
| 8316 | ||||
| 8317 | #line 519 | |||
| 8318 | #if 1 | |||
| 8319 | SIMD_INTRIN_DEF(reinterpret_s16_u16){ "reinterpret_s16_u16", simd__intrin_reinterpret_s16_u16, 0x0001 , ((void*)0) } , | |||
| 8320 | #endif // simd_sup2 | |||
| 8321 | ||||
| 8322 | #line 519 | |||
| 8323 | #if 1 | |||
| 8324 | SIMD_INTRIN_DEF(reinterpret_u32_u16){ "reinterpret_u32_u16", simd__intrin_reinterpret_u32_u16, 0x0001 , ((void*)0) } , | |||
| 8325 | #endif // simd_sup2 | |||
| 8326 | ||||
| 8327 | #line 519 | |||
| 8328 | #if 1 | |||
| 8329 | SIMD_INTRIN_DEF(reinterpret_s32_u16){ "reinterpret_s32_u16", simd__intrin_reinterpret_s32_u16, 0x0001 , ((void*)0) } , | |||
| 8330 | #endif // simd_sup2 | |||
| 8331 | ||||
| 8332 | #line 519 | |||
| 8333 | #if 1 | |||
| 8334 | SIMD_INTRIN_DEF(reinterpret_u64_u16){ "reinterpret_u64_u16", simd__intrin_reinterpret_u64_u16, 0x0001 , ((void*)0) } , | |||
| 8335 | #endif // simd_sup2 | |||
| 8336 | ||||
| 8337 | #line 519 | |||
| 8338 | #if 1 | |||
| 8339 | SIMD_INTRIN_DEF(reinterpret_s64_u16){ "reinterpret_s64_u16", simd__intrin_reinterpret_s64_u16, 0x0001 , ((void*)0) } , | |||
| 8340 | #endif // simd_sup2 | |||
| 8341 | ||||
| 8342 | #line 519 | |||
| 8343 | #if 1 | |||
| 8344 | SIMD_INTRIN_DEF(reinterpret_f32_u16){ "reinterpret_f32_u16", simd__intrin_reinterpret_f32_u16, 0x0001 , ((void*)0) } , | |||
| 8345 | #endif // simd_sup2 | |||
| 8346 | ||||
| 8347 | #line 519 | |||
| 8348 | #if NPY_SIMD_F641 | |||
| 8349 | SIMD_INTRIN_DEF(reinterpret_f64_u16){ "reinterpret_f64_u16", simd__intrin_reinterpret_f64_u16, 0x0001 , ((void*)0) } , | |||
| 8350 | #endif // simd_sup2 | |||
| 8351 | ||||
| 8352 | ||||
| 8353 | #line 527 | |||
| 8354 | SIMD_INTRIN_DEF(set_u16){ "set_u16", simd__intrin_set_u16, 0x0001, ((void*)0) } , | |||
| 8355 | ||||
| 8356 | #line 527 | |||
| 8357 | SIMD_INTRIN_DEF(setf_u16){ "setf_u16", simd__intrin_setf_u16, 0x0001, ((void*)0) } , | |||
| 8358 | ||||
| 8359 | #line 527 | |||
| 8360 | SIMD_INTRIN_DEF(setall_u16){ "setall_u16", simd__intrin_setall_u16, 0x0001, ((void*)0) } , | |||
| 8361 | ||||
| 8362 | #line 527 | |||
| 8363 | SIMD_INTRIN_DEF(zero_u16){ "zero_u16", simd__intrin_zero_u16, 0x0001, ((void*)0) } , | |||
| 8364 | ||||
| 8365 | #line 527 | |||
| 8366 | SIMD_INTRIN_DEF(select_u16){ "select_u16", simd__intrin_select_u16, 0x0001, ((void*)0) } , | |||
| 8367 | ||||
| 8368 | ||||
| 8369 | /*************************** | |||
| 8370 | * Reorder | |||
| 8371 | ***************************/ | |||
| 8372 | #line 536 | |||
| 8373 | SIMD_INTRIN_DEF(combinel_u16){ "combinel_u16", simd__intrin_combinel_u16, 0x0001, ((void*) 0) } , | |||
| 8374 | ||||
| 8375 | #line 536 | |||
| 8376 | SIMD_INTRIN_DEF(combineh_u16){ "combineh_u16", simd__intrin_combineh_u16, 0x0001, ((void*) 0) } , | |||
| 8377 | ||||
| 8378 | #line 536 | |||
| 8379 | SIMD_INTRIN_DEF(combine_u16){ "combine_u16", simd__intrin_combine_u16, 0x0001, ((void*)0) } , | |||
| 8380 | ||||
| 8381 | #line 536 | |||
| 8382 | SIMD_INTRIN_DEF(zip_u16){ "zip_u16", simd__intrin_zip_u16, 0x0001, ((void*)0) } , | |||
| 8383 | ||||
| 8384 | ||||
| 8385 | #if 1 | |||
| 8386 | SIMD_INTRIN_DEF(rev64_u16){ "rev64_u16", simd__intrin_rev64_u16, 0x0001, ((void*)0) } , | |||
| 8387 | #endif | |||
| 8388 | ||||
| 8389 | /*************************** | |||
| 8390 | * Operators | |||
| 8391 | ***************************/ | |||
| 8392 | #if 15 > 0 | |||
| 8393 | #line 550 | |||
| 8394 | SIMD_INTRIN_DEF(shl_u16){ "shl_u16", simd__intrin_shl_u16, 0x0001, ((void*)0) } , | |||
| 8395 | ||||
| 8396 | #line 550 | |||
| 8397 | SIMD_INTRIN_DEF(shr_u16){ "shr_u16", simd__intrin_shr_u16, 0x0001, ((void*)0) } , | |||
| 8398 | ||||
| 8399 | #line 550 | |||
| 8400 | SIMD_INTRIN_DEF(shli_u16){ "shli_u16", simd__intrin_shli_u16, 0x0001, ((void*)0) } , | |||
| 8401 | ||||
| 8402 | #line 550 | |||
| 8403 | SIMD_INTRIN_DEF(shri_u16){ "shri_u16", simd__intrin_shri_u16, 0x0001, ((void*)0) } , | |||
| 8404 | ||||
| 8405 | #endif // shl_imm | |||
| 8406 | ||||
| 8407 | #line 557 | |||
| 8408 | SIMD_INTRIN_DEF(and_u16){ "and_u16", simd__intrin_and_u16, 0x0001, ((void*)0) } , | |||
| 8409 | ||||
| 8410 | #line 557 | |||
| 8411 | SIMD_INTRIN_DEF(or_u16){ "or_u16", simd__intrin_or_u16, 0x0001, ((void*)0) } , | |||
| 8412 | ||||
| 8413 | #line 557 | |||
| 8414 | SIMD_INTRIN_DEF(xor_u16){ "xor_u16", simd__intrin_xor_u16, 0x0001, ((void*)0) } , | |||
| 8415 | ||||
| 8416 | #line 557 | |||
| 8417 | SIMD_INTRIN_DEF(not_u16){ "not_u16", simd__intrin_not_u16, 0x0001, ((void*)0) } , | |||
| 8418 | ||||
| 8419 | #line 557 | |||
| 8420 | SIMD_INTRIN_DEF(cmpeq_u16){ "cmpeq_u16", simd__intrin_cmpeq_u16, 0x0001, ((void*)0) } , | |||
| 8421 | ||||
| 8422 | #line 557 | |||
| 8423 | SIMD_INTRIN_DEF(cmpneq_u16){ "cmpneq_u16", simd__intrin_cmpneq_u16, 0x0001, ((void*)0) } , | |||
| 8424 | ||||
| 8425 | #line 557 | |||
| 8426 | SIMD_INTRIN_DEF(cmpgt_u16){ "cmpgt_u16", simd__intrin_cmpgt_u16, 0x0001, ((void*)0) } , | |||
| 8427 | ||||
| 8428 | #line 557 | |||
| 8429 | SIMD_INTRIN_DEF(cmpge_u16){ "cmpge_u16", simd__intrin_cmpge_u16, 0x0001, ((void*)0) } , | |||
| 8430 | ||||
| 8431 | #line 557 | |||
| 8432 | SIMD_INTRIN_DEF(cmplt_u16){ "cmplt_u16", simd__intrin_cmplt_u16, 0x0001, ((void*)0) } , | |||
| 8433 | ||||
| 8434 | #line 557 | |||
| 8435 | SIMD_INTRIN_DEF(cmple_u16){ "cmple_u16", simd__intrin_cmple_u16, 0x0001, ((void*)0) } , | |||
| 8436 | ||||
| 8437 | ||||
| 8438 | /*************************** | |||
| 8439 | * Conversion | |||
| 8440 | ***************************/ | |||
| 8441 | SIMD_INTRIN_DEF(cvt_u16_b16){ "cvt_u16_b16", simd__intrin_cvt_u16_b16, 0x0001, ((void*)0) } , | |||
| 8442 | SIMD_INTRIN_DEF(cvt_b16_u16){ "cvt_b16_u16", simd__intrin_cvt_b16_u16, 0x0001, ((void*)0) } , | |||
| 8443 | #if 1 | |||
| 8444 | SIMD_INTRIN_DEF(expand_u32_u16){ "expand_u32_u16", simd__intrin_expand_u32_u16, 0x0001, ((void *)0) } , | |||
| 8445 | #endif // expand_sup | |||
| 8446 | /*************************** | |||
| 8447 | * Arithmetic | |||
| 8448 | ***************************/ | |||
| 8449 | #line 574 | |||
| 8450 | SIMD_INTRIN_DEF(add_u16){ "add_u16", simd__intrin_add_u16, 0x0001, ((void*)0) } , | |||
| 8451 | ||||
| 8452 | #line 574 | |||
| 8453 | SIMD_INTRIN_DEF(sub_u16){ "sub_u16", simd__intrin_sub_u16, 0x0001, ((void*)0) } , | |||
| 8454 | ||||
| 8455 | ||||
| 8456 | #if 1 | |||
| 8457 | #line 581 | |||
| 8458 | SIMD_INTRIN_DEF(adds_u16){ "adds_u16", simd__intrin_adds_u16, 0x0001, ((void*)0) } , | |||
| 8459 | ||||
| 8460 | #line 581 | |||
| 8461 | SIMD_INTRIN_DEF(subs_u16){ "subs_u16", simd__intrin_subs_u16, 0x0001, ((void*)0) } , | |||
| 8462 | ||||
| 8463 | #endif // sat_sup | |||
| 8464 | ||||
| 8465 | #if 1 | |||
| 8466 | SIMD_INTRIN_DEF(mul_u16){ "mul_u16", simd__intrin_mul_u16, 0x0001, ((void*)0) } , | |||
| 8467 | #endif // mul_sup | |||
| 8468 | ||||
| 8469 | #if 0 | |||
| 8470 | SIMD_INTRIN_DEF(div_u16){ "div_u16", simd__intrin_div_u16, 0x0001, ((void*)0) } , | |||
| 8471 | #endif // div_sup | |||
| 8472 | ||||
| 8473 | #if 1 | |||
| 8474 | SIMD_INTRIN_DEF(divisor_u16){ "divisor_u16", simd__intrin_divisor_u16, 0x0001, ((void*)0) } , | |||
| 8475 | SIMD_INTRIN_DEF(divc_u16){ "divc_u16", simd__intrin_divc_u16, 0x0001, ((void*)0) } , | |||
| 8476 | #endif // intdiv_sup | |||
| 8477 | ||||
| 8478 | #if 0 | |||
| 8479 | #line 602 | |||
| 8480 | SIMD_INTRIN_DEF(muladd_u16){ "muladd_u16", simd__intrin_muladd_u16, 0x0001, ((void*)0) } , | |||
| 8481 | ||||
| 8482 | #line 602 | |||
| 8483 | SIMD_INTRIN_DEF(mulsub_u16){ "mulsub_u16", simd__intrin_mulsub_u16, 0x0001, ((void*)0) } , | |||
| 8484 | ||||
| 8485 | #line 602 | |||
| 8486 | SIMD_INTRIN_DEF(nmuladd_u16){ "nmuladd_u16", simd__intrin_nmuladd_u16, 0x0001, ((void*)0) } , | |||
| 8487 | ||||
| 8488 | #line 602 | |||
| 8489 | SIMD_INTRIN_DEF(nmulsub_u16){ "nmulsub_u16", simd__intrin_nmulsub_u16, 0x0001, ((void*)0) } , | |||
| 8490 | ||||
| 8491 | #endif // fused_sup | |||
| 8492 | ||||
| 8493 | #if 0 | |||
| 8494 | SIMD_INTRIN_DEF(sum_u16){ "sum_u16", simd__intrin_sum_u16, 0x0001, ((void*)0) } , | |||
| 8495 | #endif // sum_sup | |||
| 8496 | ||||
| 8497 | #if 1 | |||
| 8498 | SIMD_INTRIN_DEF(sumup_u16){ "sumup_u16", simd__intrin_sumup_u16, 0x0001, ((void*)0) } , | |||
| 8499 | #endif // sumup_sup | |||
| 8500 | /*************************** | |||
| 8501 | * Math | |||
| 8502 | ***************************/ | |||
| 8503 | #if 0 | |||
| 8504 | #line 620 | |||
| 8505 | SIMD_INTRIN_DEF(sqrt_u16){ "sqrt_u16", simd__intrin_sqrt_u16, 0x0001, ((void*)0) } , | |||
| 8506 | ||||
| 8507 | #line 620 | |||
| 8508 | SIMD_INTRIN_DEF(recip_u16){ "recip_u16", simd__intrin_recip_u16, 0x0001, ((void*)0) } , | |||
| 8509 | ||||
| 8510 | #line 620 | |||
| 8511 | SIMD_INTRIN_DEF(abs_u16){ "abs_u16", simd__intrin_abs_u16, 0x0001, ((void*)0) } , | |||
| 8512 | ||||
| 8513 | #line 620 | |||
| 8514 | SIMD_INTRIN_DEF(square_u16){ "square_u16", simd__intrin_square_u16, 0x0001, ((void*)0) } , | |||
| 8515 | ||||
| 8516 | #endif | |||
| 8517 | ||||
| 8518 | #line 627 | |||
| 8519 | SIMD_INTRIN_DEF(max_u16){ "max_u16", simd__intrin_max_u16, 0x0001, ((void*)0) } , | |||
| 8520 | ||||
| 8521 | #line 627 | |||
| 8522 | SIMD_INTRIN_DEF(min_u16){ "min_u16", simd__intrin_min_u16, 0x0001, ((void*)0) } , | |||
| 8523 | ||||
| 8524 | ||||
| 8525 | #if 0 | |||
| 8526 | #line 634 | |||
| 8527 | SIMD_INTRIN_DEF(maxp_u16){ "maxp_u16", simd__intrin_maxp_u16, 0x0001, ((void*)0) } , | |||
| 8528 | ||||
| 8529 | #line 634 | |||
| 8530 | SIMD_INTRIN_DEF(minp_u16){ "minp_u16", simd__intrin_minp_u16, 0x0001, ((void*)0) } , | |||
| 8531 | ||||
| 8532 | #endif | |||
| 8533 | ||||
| 8534 | /*************************** | |||
| 8535 | * Mask operations | |||
| 8536 | ***************************/ | |||
| 8537 | #line 644 | |||
| 8538 | SIMD_INTRIN_DEF(ifadd_u16){ "ifadd_u16", simd__intrin_ifadd_u16, 0x0001, ((void*)0) } , | |||
| 8539 | ||||
| 8540 | #line 644 | |||
| 8541 | SIMD_INTRIN_DEF(ifsub_u16){ "ifsub_u16", simd__intrin_ifsub_u16, 0x0001, ((void*)0) } , | |||
| 8542 | ||||
| 8543 | ||||
| 8544 | #endif // simd_sup | |||
| 8545 | ||||
| 8546 | #line 489 | |||
| 8547 | #if 1 | |||
| 8548 | ||||
| 8549 | /*************************** | |||
| 8550 | * Memory | |||
| 8551 | ***************************/ | |||
| 8552 | #line 497 | |||
| 8553 | SIMD_INTRIN_DEF(load_s16){ "load_s16", simd__intrin_load_s16, 0x0001, ((void*)0) } , | |||
| 8554 | ||||
| 8555 | #line 497 | |||
| 8556 | SIMD_INTRIN_DEF(loada_s16){ "loada_s16", simd__intrin_loada_s16, 0x0001, ((void*)0) } , | |||
| 8557 | ||||
| 8558 | #line 497 | |||
| 8559 | SIMD_INTRIN_DEF(loads_s16){ "loads_s16", simd__intrin_loads_s16, 0x0001, ((void*)0) } , | |||
| 8560 | ||||
| 8561 | #line 497 | |||
| 8562 | SIMD_INTRIN_DEF(loadl_s16){ "loadl_s16", simd__intrin_loadl_s16, 0x0001, ((void*)0) } , | |||
| 8563 | ||||
| 8564 | #line 497 | |||
| 8565 | SIMD_INTRIN_DEF(store_s16){ "store_s16", simd__intrin_store_s16, 0x0001, ((void*)0) } , | |||
| 8566 | ||||
| 8567 | #line 497 | |||
| 8568 | SIMD_INTRIN_DEF(storea_s16){ "storea_s16", simd__intrin_storea_s16, 0x0001, ((void*)0) } , | |||
| 8569 | ||||
| 8570 | #line 497 | |||
| 8571 | SIMD_INTRIN_DEF(stores_s16){ "stores_s16", simd__intrin_stores_s16, 0x0001, ((void*)0) } , | |||
| 8572 | ||||
| 8573 | #line 497 | |||
| 8574 | SIMD_INTRIN_DEF(storel_s16){ "storel_s16", simd__intrin_storel_s16, 0x0001, ((void*)0) } , | |||
| 8575 | ||||
| 8576 | #line 497 | |||
| 8577 | SIMD_INTRIN_DEF(storeh_s16){ "storeh_s16", simd__intrin_storeh_s16, 0x0001, ((void*)0) } , | |||
| 8578 | ||||
| 8579 | ||||
| 8580 | /**************************************** | |||
| 8581 | * Non-contiguous/Partial Memory access | |||
| 8582 | ****************************************/ | |||
| 8583 | #if 0 | |||
| 8584 | #line 508 | |||
| 8585 | SIMD_INTRIN_DEF(load_till_s16){ "load_till_s16", simd__intrin_load_till_s16, 0x0001, ((void *)0) } , | |||
| 8586 | ||||
| 8587 | #line 508 | |||
| 8588 | SIMD_INTRIN_DEF(load_tillz_s16){ "load_tillz_s16", simd__intrin_load_tillz_s16, 0x0001, ((void *)0) } , | |||
| 8589 | ||||
| 8590 | #line 508 | |||
| 8591 | SIMD_INTRIN_DEF(loadn_s16){ "loadn_s16", simd__intrin_loadn_s16, 0x0001, ((void*)0) } , | |||
| 8592 | ||||
| 8593 | #line 508 | |||
| 8594 | SIMD_INTRIN_DEF(loadn_till_s16){ "loadn_till_s16", simd__intrin_loadn_till_s16, 0x0001, ((void *)0) } , | |||
| 8595 | ||||
| 8596 | #line 508 | |||
| 8597 | SIMD_INTRIN_DEF(loadn_tillz_s16){ "loadn_tillz_s16", simd__intrin_loadn_tillz_s16, 0x0001, (( void*)0) } , | |||
| 8598 | ||||
| 8599 | #line 508 | |||
| 8600 | SIMD_INTRIN_DEF(store_till_s16){ "store_till_s16", simd__intrin_store_till_s16, 0x0001, ((void *)0) } , | |||
| 8601 | ||||
| 8602 | #line 508 | |||
| 8603 | SIMD_INTRIN_DEF(storen_s16){ "storen_s16", simd__intrin_storen_s16, 0x0001, ((void*)0) } , | |||
| 8604 | ||||
| 8605 | #line 508 | |||
| 8606 | SIMD_INTRIN_DEF(storen_till_s16){ "storen_till_s16", simd__intrin_storen_till_s16, 0x0001, (( void*)0) } , | |||
| 8607 | ||||
| 8608 | #endif // ncont_sup | |||
| 8609 | ||||
| 8610 | /*************************** | |||
| 8611 | * Misc | |||
| 8612 | ***************************/ | |||
| 8613 | #line 519 | |||
| 8614 | #if 1 | |||
| 8615 | SIMD_INTRIN_DEF(reinterpret_u8_s16){ "reinterpret_u8_s16", simd__intrin_reinterpret_u8_s16, 0x0001 , ((void*)0) } , | |||
| 8616 | #endif // simd_sup2 | |||
| 8617 | ||||
| 8618 | #line 519 | |||
| 8619 | #if 1 | |||
| 8620 | SIMD_INTRIN_DEF(reinterpret_s8_s16){ "reinterpret_s8_s16", simd__intrin_reinterpret_s8_s16, 0x0001 , ((void*)0) } , | |||
| 8621 | #endif // simd_sup2 | |||
| 8622 | ||||
| 8623 | #line 519 | |||
| 8624 | #if 1 | |||
| 8625 | SIMD_INTRIN_DEF(reinterpret_u16_s16){ "reinterpret_u16_s16", simd__intrin_reinterpret_u16_s16, 0x0001 , ((void*)0) } , | |||
| 8626 | #endif // simd_sup2 | |||
| 8627 | ||||
| 8628 | #line 519 | |||
| 8629 | #if 1 | |||
| 8630 | SIMD_INTRIN_DEF(reinterpret_s16_s16){ "reinterpret_s16_s16", simd__intrin_reinterpret_s16_s16, 0x0001 , ((void*)0) } , | |||
| 8631 | #endif // simd_sup2 | |||
| 8632 | ||||
| 8633 | #line 519 | |||
| 8634 | #if 1 | |||
| 8635 | SIMD_INTRIN_DEF(reinterpret_u32_s16){ "reinterpret_u32_s16", simd__intrin_reinterpret_u32_s16, 0x0001 , ((void*)0) } , | |||
| 8636 | #endif // simd_sup2 | |||
| 8637 | ||||
| 8638 | #line 519 | |||
| 8639 | #if 1 | |||
| 8640 | SIMD_INTRIN_DEF(reinterpret_s32_s16){ "reinterpret_s32_s16", simd__intrin_reinterpret_s32_s16, 0x0001 , ((void*)0) } , | |||
| 8641 | #endif // simd_sup2 | |||
| 8642 | ||||
| 8643 | #line 519 | |||
| 8644 | #if 1 | |||
| 8645 | SIMD_INTRIN_DEF(reinterpret_u64_s16){ "reinterpret_u64_s16", simd__intrin_reinterpret_u64_s16, 0x0001 , ((void*)0) } , | |||
| 8646 | #endif // simd_sup2 | |||
| 8647 | ||||
| 8648 | #line 519 | |||
| 8649 | #if 1 | |||
| 8650 | SIMD_INTRIN_DEF(reinterpret_s64_s16){ "reinterpret_s64_s16", simd__intrin_reinterpret_s64_s16, 0x0001 , ((void*)0) } , | |||
| 8651 | #endif // simd_sup2 | |||
| 8652 | ||||
| 8653 | #line 519 | |||
| 8654 | #if 1 | |||
| 8655 | SIMD_INTRIN_DEF(reinterpret_f32_s16){ "reinterpret_f32_s16", simd__intrin_reinterpret_f32_s16, 0x0001 , ((void*)0) } , | |||
| 8656 | #endif // simd_sup2 | |||
| 8657 | ||||
| 8658 | #line 519 | |||
| 8659 | #if NPY_SIMD_F641 | |||
| 8660 | SIMD_INTRIN_DEF(reinterpret_f64_s16){ "reinterpret_f64_s16", simd__intrin_reinterpret_f64_s16, 0x0001 , ((void*)0) } , | |||
| 8661 | #endif // simd_sup2 | |||
| 8662 | ||||
| 8663 | ||||
| 8664 | #line 527 | |||
| 8665 | SIMD_INTRIN_DEF(set_s16){ "set_s16", simd__intrin_set_s16, 0x0001, ((void*)0) } , | |||
| 8666 | ||||
| 8667 | #line 527 | |||
| 8668 | SIMD_INTRIN_DEF(setf_s16){ "setf_s16", simd__intrin_setf_s16, 0x0001, ((void*)0) } , | |||
| 8669 | ||||
| 8670 | #line 527 | |||
| 8671 | SIMD_INTRIN_DEF(setall_s16){ "setall_s16", simd__intrin_setall_s16, 0x0001, ((void*)0) } , | |||
| 8672 | ||||
| 8673 | #line 527 | |||
| 8674 | SIMD_INTRIN_DEF(zero_s16){ "zero_s16", simd__intrin_zero_s16, 0x0001, ((void*)0) } , | |||
| 8675 | ||||
| 8676 | #line 527 | |||
| 8677 | SIMD_INTRIN_DEF(select_s16){ "select_s16", simd__intrin_select_s16, 0x0001, ((void*)0) } , | |||
| 8678 | ||||
| 8679 | ||||
| 8680 | /*************************** | |||
| 8681 | * Reorder | |||
| 8682 | ***************************/ | |||
| 8683 | #line 536 | |||
| 8684 | SIMD_INTRIN_DEF(combinel_s16){ "combinel_s16", simd__intrin_combinel_s16, 0x0001, ((void*) 0) } , | |||
| 8685 | ||||
| 8686 | #line 536 | |||
| 8687 | SIMD_INTRIN_DEF(combineh_s16){ "combineh_s16", simd__intrin_combineh_s16, 0x0001, ((void*) 0) } , | |||
| 8688 | ||||
| 8689 | #line 536 | |||
| 8690 | SIMD_INTRIN_DEF(combine_s16){ "combine_s16", simd__intrin_combine_s16, 0x0001, ((void*)0) } , | |||
| 8691 | ||||
| 8692 | #line 536 | |||
| 8693 | SIMD_INTRIN_DEF(zip_s16){ "zip_s16", simd__intrin_zip_s16, 0x0001, ((void*)0) } , | |||
| 8694 | ||||
| 8695 | ||||
| 8696 | #if 1 | |||
| 8697 | SIMD_INTRIN_DEF(rev64_s16){ "rev64_s16", simd__intrin_rev64_s16, 0x0001, ((void*)0) } , | |||
| 8698 | #endif | |||
| 8699 | ||||
| 8700 | /*************************** | |||
| 8701 | * Operators | |||
| 8702 | ***************************/ | |||
| 8703 | #if 15 > 0 | |||
| 8704 | #line 550 | |||
| 8705 | SIMD_INTRIN_DEF(shl_s16){ "shl_s16", simd__intrin_shl_s16, 0x0001, ((void*)0) } , | |||
| 8706 | ||||
| 8707 | #line 550 | |||
| 8708 | SIMD_INTRIN_DEF(shr_s16){ "shr_s16", simd__intrin_shr_s16, 0x0001, ((void*)0) } , | |||
| 8709 | ||||
| 8710 | #line 550 | |||
| 8711 | SIMD_INTRIN_DEF(shli_s16){ "shli_s16", simd__intrin_shli_s16, 0x0001, ((void*)0) } , | |||
| 8712 | ||||
| 8713 | #line 550 | |||
| 8714 | SIMD_INTRIN_DEF(shri_s16){ "shri_s16", simd__intrin_shri_s16, 0x0001, ((void*)0) } , | |||
| 8715 | ||||
| 8716 | #endif // shl_imm | |||
| 8717 | ||||
| 8718 | #line 557 | |||
| 8719 | SIMD_INTRIN_DEF(and_s16){ "and_s16", simd__intrin_and_s16, 0x0001, ((void*)0) } , | |||
| 8720 | ||||
| 8721 | #line 557 | |||
| 8722 | SIMD_INTRIN_DEF(or_s16){ "or_s16", simd__intrin_or_s16, 0x0001, ((void*)0) } , | |||
| 8723 | ||||
| 8724 | #line 557 | |||
| 8725 | SIMD_INTRIN_DEF(xor_s16){ "xor_s16", simd__intrin_xor_s16, 0x0001, ((void*)0) } , | |||
| 8726 | ||||
| 8727 | #line 557 | |||
| 8728 | SIMD_INTRIN_DEF(not_s16){ "not_s16", simd__intrin_not_s16, 0x0001, ((void*)0) } , | |||
| 8729 | ||||
| 8730 | #line 557 | |||
| 8731 | SIMD_INTRIN_DEF(cmpeq_s16){ "cmpeq_s16", simd__intrin_cmpeq_s16, 0x0001, ((void*)0) } , | |||
| 8732 | ||||
| 8733 | #line 557 | |||
| 8734 | SIMD_INTRIN_DEF(cmpneq_s16){ "cmpneq_s16", simd__intrin_cmpneq_s16, 0x0001, ((void*)0) } , | |||
| 8735 | ||||
| 8736 | #line 557 | |||
| 8737 | SIMD_INTRIN_DEF(cmpgt_s16){ "cmpgt_s16", simd__intrin_cmpgt_s16, 0x0001, ((void*)0) } , | |||
| 8738 | ||||
| 8739 | #line 557 | |||
| 8740 | SIMD_INTRIN_DEF(cmpge_s16){ "cmpge_s16", simd__intrin_cmpge_s16, 0x0001, ((void*)0) } , | |||
| 8741 | ||||
| 8742 | #line 557 | |||
| 8743 | SIMD_INTRIN_DEF(cmplt_s16){ "cmplt_s16", simd__intrin_cmplt_s16, 0x0001, ((void*)0) } , | |||
| 8744 | ||||
| 8745 | #line 557 | |||
| 8746 | SIMD_INTRIN_DEF(cmple_s16){ "cmple_s16", simd__intrin_cmple_s16, 0x0001, ((void*)0) } , | |||
| 8747 | ||||
| 8748 | ||||
| 8749 | /*************************** | |||
| 8750 | * Conversion | |||
| 8751 | ***************************/ | |||
| 8752 | SIMD_INTRIN_DEF(cvt_s16_b16){ "cvt_s16_b16", simd__intrin_cvt_s16_b16, 0x0001, ((void*)0) } , | |||
| 8753 | SIMD_INTRIN_DEF(cvt_b16_s16){ "cvt_b16_s16", simd__intrin_cvt_b16_s16, 0x0001, ((void*)0) } , | |||
| 8754 | #if 0 | |||
| 8755 | SIMD_INTRIN_DEF(expand_s16_s16){ "expand_s16_s16", simd__intrin_expand_s16_s16, 0x0001, ((void *)0) } , | |||
| 8756 | #endif // expand_sup | |||
| 8757 | /*************************** | |||
| 8758 | * Arithmetic | |||
| 8759 | ***************************/ | |||
| 8760 | #line 574 | |||
| 8761 | SIMD_INTRIN_DEF(add_s16){ "add_s16", simd__intrin_add_s16, 0x0001, ((void*)0) } , | |||
| 8762 | ||||
| 8763 | #line 574 | |||
| 8764 | SIMD_INTRIN_DEF(sub_s16){ "sub_s16", simd__intrin_sub_s16, 0x0001, ((void*)0) } , | |||
| 8765 | ||||
| 8766 | ||||
| 8767 | #if 1 | |||
| 8768 | #line 581 | |||
| 8769 | SIMD_INTRIN_DEF(adds_s16){ "adds_s16", simd__intrin_adds_s16, 0x0001, ((void*)0) } , | |||
| 8770 | ||||
| 8771 | #line 581 | |||
| 8772 | SIMD_INTRIN_DEF(subs_s16){ "subs_s16", simd__intrin_subs_s16, 0x0001, ((void*)0) } , | |||
| 8773 | ||||
| 8774 | #endif // sat_sup | |||
| 8775 | ||||
| 8776 | #if 1 | |||
| 8777 | SIMD_INTRIN_DEF(mul_s16){ "mul_s16", simd__intrin_mul_s16, 0x0001, ((void*)0) } , | |||
| 8778 | #endif // mul_sup | |||
| 8779 | ||||
| 8780 | #if 0 | |||
| 8781 | SIMD_INTRIN_DEF(div_s16){ "div_s16", simd__intrin_div_s16, 0x0001, ((void*)0) } , | |||
| 8782 | #endif // div_sup | |||
| 8783 | ||||
| 8784 | #if 1 | |||
| 8785 | SIMD_INTRIN_DEF(divisor_s16){ "divisor_s16", simd__intrin_divisor_s16, 0x0001, ((void*)0) } , | |||
| 8786 | SIMD_INTRIN_DEF(divc_s16){ "divc_s16", simd__intrin_divc_s16, 0x0001, ((void*)0) } , | |||
| 8787 | #endif // intdiv_sup | |||
| 8788 | ||||
| 8789 | #if 0 | |||
| 8790 | #line 602 | |||
| 8791 | SIMD_INTRIN_DEF(muladd_s16){ "muladd_s16", simd__intrin_muladd_s16, 0x0001, ((void*)0) } , | |||
| 8792 | ||||
| 8793 | #line 602 | |||
| 8794 | SIMD_INTRIN_DEF(mulsub_s16){ "mulsub_s16", simd__intrin_mulsub_s16, 0x0001, ((void*)0) } , | |||
| 8795 | ||||
| 8796 | #line 602 | |||
| 8797 | SIMD_INTRIN_DEF(nmuladd_s16){ "nmuladd_s16", simd__intrin_nmuladd_s16, 0x0001, ((void*)0) } , | |||
| 8798 | ||||
| 8799 | #line 602 | |||
| 8800 | SIMD_INTRIN_DEF(nmulsub_s16){ "nmulsub_s16", simd__intrin_nmulsub_s16, 0x0001, ((void*)0) } , | |||
| 8801 | ||||
| 8802 | #endif // fused_sup | |||
| 8803 | ||||
| 8804 | #if 0 | |||
| 8805 | SIMD_INTRIN_DEF(sum_s16){ "sum_s16", simd__intrin_sum_s16, 0x0001, ((void*)0) } , | |||
| 8806 | #endif // sum_sup | |||
| 8807 | ||||
| 8808 | #if 0 | |||
| 8809 | SIMD_INTRIN_DEF(sumup_s16){ "sumup_s16", simd__intrin_sumup_s16, 0x0001, ((void*)0) } , | |||
| 8810 | #endif // sumup_sup | |||
| 8811 | /*************************** | |||
| 8812 | * Math | |||
| 8813 | ***************************/ | |||
| 8814 | #if 0 | |||
| 8815 | #line 620 | |||
| 8816 | SIMD_INTRIN_DEF(sqrt_s16){ "sqrt_s16", simd__intrin_sqrt_s16, 0x0001, ((void*)0) } , | |||
| 8817 | ||||
| 8818 | #line 620 | |||
| 8819 | SIMD_INTRIN_DEF(recip_s16){ "recip_s16", simd__intrin_recip_s16, 0x0001, ((void*)0) } , | |||
| 8820 | ||||
| 8821 | #line 620 | |||
| 8822 | SIMD_INTRIN_DEF(abs_s16){ "abs_s16", simd__intrin_abs_s16, 0x0001, ((void*)0) } , | |||
| 8823 | ||||
| 8824 | #line 620 | |||
| 8825 | SIMD_INTRIN_DEF(square_s16){ "square_s16", simd__intrin_square_s16, 0x0001, ((void*)0) } , | |||
| 8826 | ||||
| 8827 | #endif | |||
| 8828 | ||||
| 8829 | #line 627 | |||
| 8830 | SIMD_INTRIN_DEF(max_s16){ "max_s16", simd__intrin_max_s16, 0x0001, ((void*)0) } , | |||
| 8831 | ||||
| 8832 | #line 627 | |||
| 8833 | SIMD_INTRIN_DEF(min_s16){ "min_s16", simd__intrin_min_s16, 0x0001, ((void*)0) } , | |||
| 8834 | ||||
| 8835 | ||||
| 8836 | #if 0 | |||
| 8837 | #line 634 | |||
| 8838 | SIMD_INTRIN_DEF(maxp_s16){ "maxp_s16", simd__intrin_maxp_s16, 0x0001, ((void*)0) } , | |||
| 8839 | ||||
| 8840 | #line 634 | |||
| 8841 | SIMD_INTRIN_DEF(minp_s16){ "minp_s16", simd__intrin_minp_s16, 0x0001, ((void*)0) } , | |||
| 8842 | ||||
| 8843 | #endif | |||
| 8844 | ||||
| 8845 | /*************************** | |||
| 8846 | * Mask operations | |||
| 8847 | ***************************/ | |||
| 8848 | #line 644 | |||
| 8849 | SIMD_INTRIN_DEF(ifadd_s16){ "ifadd_s16", simd__intrin_ifadd_s16, 0x0001, ((void*)0) } , | |||
| 8850 | ||||
| 8851 | #line 644 | |||
| 8852 | SIMD_INTRIN_DEF(ifsub_s16){ "ifsub_s16", simd__intrin_ifsub_s16, 0x0001, ((void*)0) } , | |||
| 8853 | ||||
| 8854 | ||||
| 8855 | #endif // simd_sup | |||
| 8856 | ||||
| 8857 | #line 489 | |||
| 8858 | #if 1 | |||
| 8859 | ||||
| 8860 | /*************************** | |||
| 8861 | * Memory | |||
| 8862 | ***************************/ | |||
| 8863 | #line 497 | |||
| 8864 | SIMD_INTRIN_DEF(load_u32){ "load_u32", simd__intrin_load_u32, 0x0001, ((void*)0) } , | |||
| 8865 | ||||
| 8866 | #line 497 | |||
| 8867 | SIMD_INTRIN_DEF(loada_u32){ "loada_u32", simd__intrin_loada_u32, 0x0001, ((void*)0) } , | |||
| 8868 | ||||
| 8869 | #line 497 | |||
| 8870 | SIMD_INTRIN_DEF(loads_u32){ "loads_u32", simd__intrin_loads_u32, 0x0001, ((void*)0) } , | |||
| 8871 | ||||
| 8872 | #line 497 | |||
| 8873 | SIMD_INTRIN_DEF(loadl_u32){ "loadl_u32", simd__intrin_loadl_u32, 0x0001, ((void*)0) } , | |||
| 8874 | ||||
| 8875 | #line 497 | |||
| 8876 | SIMD_INTRIN_DEF(store_u32){ "store_u32", simd__intrin_store_u32, 0x0001, ((void*)0) } , | |||
| 8877 | ||||
| 8878 | #line 497 | |||
| 8879 | SIMD_INTRIN_DEF(storea_u32){ "storea_u32", simd__intrin_storea_u32, 0x0001, ((void*)0) } , | |||
| 8880 | ||||
| 8881 | #line 497 | |||
| 8882 | SIMD_INTRIN_DEF(stores_u32){ "stores_u32", simd__intrin_stores_u32, 0x0001, ((void*)0) } , | |||
| 8883 | ||||
| 8884 | #line 497 | |||
| 8885 | SIMD_INTRIN_DEF(storel_u32){ "storel_u32", simd__intrin_storel_u32, 0x0001, ((void*)0) } , | |||
| 8886 | ||||
| 8887 | #line 497 | |||
| 8888 | SIMD_INTRIN_DEF(storeh_u32){ "storeh_u32", simd__intrin_storeh_u32, 0x0001, ((void*)0) } , | |||
| 8889 | ||||
| 8890 | ||||
| 8891 | /**************************************** | |||
| 8892 | * Non-contiguous/Partial Memory access | |||
| 8893 | ****************************************/ | |||
| 8894 | #if 1 | |||
| 8895 | #line 508 | |||
| 8896 | SIMD_INTRIN_DEF(load_till_u32){ "load_till_u32", simd__intrin_load_till_u32, 0x0001, ((void *)0) } , | |||
| 8897 | ||||
| 8898 | #line 508 | |||
| 8899 | SIMD_INTRIN_DEF(load_tillz_u32){ "load_tillz_u32", simd__intrin_load_tillz_u32, 0x0001, ((void *)0) } , | |||
| 8900 | ||||
| 8901 | #line 508 | |||
| 8902 | SIMD_INTRIN_DEF(loadn_u32){ "loadn_u32", simd__intrin_loadn_u32, 0x0001, ((void*)0) } , | |||
| 8903 | ||||
| 8904 | #line 508 | |||
| 8905 | SIMD_INTRIN_DEF(loadn_till_u32){ "loadn_till_u32", simd__intrin_loadn_till_u32, 0x0001, ((void *)0) } , | |||
| 8906 | ||||
| 8907 | #line 508 | |||
| 8908 | SIMD_INTRIN_DEF(loadn_tillz_u32){ "loadn_tillz_u32", simd__intrin_loadn_tillz_u32, 0x0001, (( void*)0) } , | |||
| 8909 | ||||
| 8910 | #line 508 | |||
| 8911 | SIMD_INTRIN_DEF(store_till_u32){ "store_till_u32", simd__intrin_store_till_u32, 0x0001, ((void *)0) } , | |||
| 8912 | ||||
| 8913 | #line 508 | |||
| 8914 | SIMD_INTRIN_DEF(storen_u32){ "storen_u32", simd__intrin_storen_u32, 0x0001, ((void*)0) } , | |||
| 8915 | ||||
| 8916 | #line 508 | |||
| 8917 | SIMD_INTRIN_DEF(storen_till_u32){ "storen_till_u32", simd__intrin_storen_till_u32, 0x0001, (( void*)0) } , | |||
| 8918 | ||||
| 8919 | #endif // ncont_sup | |||
| 8920 | ||||
| 8921 | /*************************** | |||
| 8922 | * Misc | |||
| 8923 | ***************************/ | |||
| 8924 | #line 519 | |||
| 8925 | #if 1 | |||
| 8926 | SIMD_INTRIN_DEF(reinterpret_u8_u32){ "reinterpret_u8_u32", simd__intrin_reinterpret_u8_u32, 0x0001 , ((void*)0) } , | |||
| 8927 | #endif // simd_sup2 | |||
| 8928 | ||||
| 8929 | #line 519 | |||
| 8930 | #if 1 | |||
| 8931 | SIMD_INTRIN_DEF(reinterpret_s8_u32){ "reinterpret_s8_u32", simd__intrin_reinterpret_s8_u32, 0x0001 , ((void*)0) } , | |||
| 8932 | #endif // simd_sup2 | |||
| 8933 | ||||
| 8934 | #line 519 | |||
| 8935 | #if 1 | |||
| 8936 | SIMD_INTRIN_DEF(reinterpret_u16_u32){ "reinterpret_u16_u32", simd__intrin_reinterpret_u16_u32, 0x0001 , ((void*)0) } , | |||
| 8937 | #endif // simd_sup2 | |||
| 8938 | ||||
| 8939 | #line 519 | |||
| 8940 | #if 1 | |||
| 8941 | SIMD_INTRIN_DEF(reinterpret_s16_u32){ "reinterpret_s16_u32", simd__intrin_reinterpret_s16_u32, 0x0001 , ((void*)0) } , | |||
| 8942 | #endif // simd_sup2 | |||
| 8943 | ||||
| 8944 | #line 519 | |||
| 8945 | #if 1 | |||
| 8946 | SIMD_INTRIN_DEF(reinterpret_u32_u32){ "reinterpret_u32_u32", simd__intrin_reinterpret_u32_u32, 0x0001 , ((void*)0) } , | |||
| 8947 | #endif // simd_sup2 | |||
| 8948 | ||||
| 8949 | #line 519 | |||
| 8950 | #if 1 | |||
| 8951 | SIMD_INTRIN_DEF(reinterpret_s32_u32){ "reinterpret_s32_u32", simd__intrin_reinterpret_s32_u32, 0x0001 , ((void*)0) } , | |||
| 8952 | #endif // simd_sup2 | |||
| 8953 | ||||
| 8954 | #line 519 | |||
| 8955 | #if 1 | |||
| 8956 | SIMD_INTRIN_DEF(reinterpret_u64_u32){ "reinterpret_u64_u32", simd__intrin_reinterpret_u64_u32, 0x0001 , ((void*)0) } , | |||
| 8957 | #endif // simd_sup2 | |||
| 8958 | ||||
| 8959 | #line 519 | |||
| 8960 | #if 1 | |||
| 8961 | SIMD_INTRIN_DEF(reinterpret_s64_u32){ "reinterpret_s64_u32", simd__intrin_reinterpret_s64_u32, 0x0001 , ((void*)0) } , | |||
| 8962 | #endif // simd_sup2 | |||
| 8963 | ||||
| 8964 | #line 519 | |||
| 8965 | #if 1 | |||
| 8966 | SIMD_INTRIN_DEF(reinterpret_f32_u32){ "reinterpret_f32_u32", simd__intrin_reinterpret_f32_u32, 0x0001 , ((void*)0) } , | |||
| 8967 | #endif // simd_sup2 | |||
| 8968 | ||||
| 8969 | #line 519 | |||
| 8970 | #if NPY_SIMD_F641 | |||
| 8971 | SIMD_INTRIN_DEF(reinterpret_f64_u32){ "reinterpret_f64_u32", simd__intrin_reinterpret_f64_u32, 0x0001 , ((void*)0) } , | |||
| 8972 | #endif // simd_sup2 | |||
| 8973 | ||||
| 8974 | ||||
| 8975 | #line 527 | |||
| 8976 | SIMD_INTRIN_DEF(set_u32){ "set_u32", simd__intrin_set_u32, 0x0001, ((void*)0) } , | |||
| 8977 | ||||
| 8978 | #line 527 | |||
| 8979 | SIMD_INTRIN_DEF(setf_u32){ "setf_u32", simd__intrin_setf_u32, 0x0001, ((void*)0) } , | |||
| 8980 | ||||
| 8981 | #line 527 | |||
| 8982 | SIMD_INTRIN_DEF(setall_u32){ "setall_u32", simd__intrin_setall_u32, 0x0001, ((void*)0) } , | |||
| 8983 | ||||
| 8984 | #line 527 | |||
| 8985 | SIMD_INTRIN_DEF(zero_u32){ "zero_u32", simd__intrin_zero_u32, 0x0001, ((void*)0) } , | |||
| 8986 | ||||
| 8987 | #line 527 | |||
| 8988 | SIMD_INTRIN_DEF(select_u32){ "select_u32", simd__intrin_select_u32, 0x0001, ((void*)0) } , | |||
| 8989 | ||||
| 8990 | ||||
| 8991 | /*************************** | |||
| 8992 | * Reorder | |||
| 8993 | ***************************/ | |||
| 8994 | #line 536 | |||
| 8995 | SIMD_INTRIN_DEF(combinel_u32){ "combinel_u32", simd__intrin_combinel_u32, 0x0001, ((void*) 0) } , | |||
| 8996 | ||||
| 8997 | #line 536 | |||
| 8998 | SIMD_INTRIN_DEF(combineh_u32){ "combineh_u32", simd__intrin_combineh_u32, 0x0001, ((void*) 0) } , | |||
| 8999 | ||||
| 9000 | #line 536 | |||
| 9001 | SIMD_INTRIN_DEF(combine_u32){ "combine_u32", simd__intrin_combine_u32, 0x0001, ((void*)0) } , | |||
| 9002 | ||||
| 9003 | #line 536 | |||
| 9004 | SIMD_INTRIN_DEF(zip_u32){ "zip_u32", simd__intrin_zip_u32, 0x0001, ((void*)0) } , | |||
| 9005 | ||||
| 9006 | ||||
| 9007 | #if 1 | |||
| 9008 | SIMD_INTRIN_DEF(rev64_u32){ "rev64_u32", simd__intrin_rev64_u32, 0x0001, ((void*)0) } , | |||
| 9009 | #endif | |||
| 9010 | ||||
| 9011 | /*************************** | |||
| 9012 | * Operators | |||
| 9013 | ***************************/ | |||
| 9014 | #if 31 > 0 | |||
| 9015 | #line 550 | |||
| 9016 | SIMD_INTRIN_DEF(shl_u32){ "shl_u32", simd__intrin_shl_u32, 0x0001, ((void*)0) } , | |||
| 9017 | ||||
| 9018 | #line 550 | |||
| 9019 | SIMD_INTRIN_DEF(shr_u32){ "shr_u32", simd__intrin_shr_u32, 0x0001, ((void*)0) } , | |||
| 9020 | ||||
| 9021 | #line 550 | |||
| 9022 | SIMD_INTRIN_DEF(shli_u32){ "shli_u32", simd__intrin_shli_u32, 0x0001, ((void*)0) } , | |||
| 9023 | ||||
| 9024 | #line 550 | |||
| 9025 | SIMD_INTRIN_DEF(shri_u32){ "shri_u32", simd__intrin_shri_u32, 0x0001, ((void*)0) } , | |||
| 9026 | ||||
| 9027 | #endif // shl_imm | |||
| 9028 | ||||
| 9029 | #line 557 | |||
| 9030 | SIMD_INTRIN_DEF(and_u32){ "and_u32", simd__intrin_and_u32, 0x0001, ((void*)0) } , | |||
| 9031 | ||||
| 9032 | #line 557 | |||
| 9033 | SIMD_INTRIN_DEF(or_u32){ "or_u32", simd__intrin_or_u32, 0x0001, ((void*)0) } , | |||
| 9034 | ||||
| 9035 | #line 557 | |||
| 9036 | SIMD_INTRIN_DEF(xor_u32){ "xor_u32", simd__intrin_xor_u32, 0x0001, ((void*)0) } , | |||
| 9037 | ||||
| 9038 | #line 557 | |||
| 9039 | SIMD_INTRIN_DEF(not_u32){ "not_u32", simd__intrin_not_u32, 0x0001, ((void*)0) } , | |||
| 9040 | ||||
| 9041 | #line 557 | |||
| 9042 | SIMD_INTRIN_DEF(cmpeq_u32){ "cmpeq_u32", simd__intrin_cmpeq_u32, 0x0001, ((void*)0) } , | |||
| 9043 | ||||
| 9044 | #line 557 | |||
| 9045 | SIMD_INTRIN_DEF(cmpneq_u32){ "cmpneq_u32", simd__intrin_cmpneq_u32, 0x0001, ((void*)0) } , | |||
| 9046 | ||||
| 9047 | #line 557 | |||
| 9048 | SIMD_INTRIN_DEF(cmpgt_u32){ "cmpgt_u32", simd__intrin_cmpgt_u32, 0x0001, ((void*)0) } , | |||
| 9049 | ||||
| 9050 | #line 557 | |||
| 9051 | SIMD_INTRIN_DEF(cmpge_u32){ "cmpge_u32", simd__intrin_cmpge_u32, 0x0001, ((void*)0) } , | |||
| 9052 | ||||
| 9053 | #line 557 | |||
| 9054 | SIMD_INTRIN_DEF(cmplt_u32){ "cmplt_u32", simd__intrin_cmplt_u32, 0x0001, ((void*)0) } , | |||
| 9055 | ||||
| 9056 | #line 557 | |||
| 9057 | SIMD_INTRIN_DEF(cmple_u32){ "cmple_u32", simd__intrin_cmple_u32, 0x0001, ((void*)0) } , | |||
| 9058 | ||||
| 9059 | ||||
| 9060 | /*************************** | |||
| 9061 | * Conversion | |||
| 9062 | ***************************/ | |||
| 9063 | SIMD_INTRIN_DEF(cvt_u32_b32){ "cvt_u32_b32", simd__intrin_cvt_u32_b32, 0x0001, ((void*)0) } , | |||
| 9064 | SIMD_INTRIN_DEF(cvt_b32_u32){ "cvt_b32_u32", simd__intrin_cvt_b32_u32, 0x0001, ((void*)0) } , | |||
| 9065 | #if 0 | |||
| 9066 | SIMD_INTRIN_DEF(expand_u32_u32){ "expand_u32_u32", simd__intrin_expand_u32_u32, 0x0001, ((void *)0) } , | |||
| 9067 | #endif // expand_sup | |||
| 9068 | /*************************** | |||
| 9069 | * Arithmetic | |||
| 9070 | ***************************/ | |||
| 9071 | #line 574 | |||
| 9072 | SIMD_INTRIN_DEF(add_u32){ "add_u32", simd__intrin_add_u32, 0x0001, ((void*)0) } , | |||
| 9073 | ||||
| 9074 | #line 574 | |||
| 9075 | SIMD_INTRIN_DEF(sub_u32){ "sub_u32", simd__intrin_sub_u32, 0x0001, ((void*)0) } , | |||
| 9076 | ||||
| 9077 | ||||
| 9078 | #if 0 | |||
| 9079 | #line 581 | |||
| 9080 | SIMD_INTRIN_DEF(adds_u32){ "adds_u32", simd__intrin_adds_u32, 0x0001, ((void*)0) } , | |||
| 9081 | ||||
| 9082 | #line 581 | |||
| 9083 | SIMD_INTRIN_DEF(subs_u32){ "subs_u32", simd__intrin_subs_u32, 0x0001, ((void*)0) } , | |||
| 9084 | ||||
| 9085 | #endif // sat_sup | |||
| 9086 | ||||
| 9087 | #if 1 | |||
| 9088 | SIMD_INTRIN_DEF(mul_u32){ "mul_u32", simd__intrin_mul_u32, 0x0001, ((void*)0) } , | |||
| 9089 | #endif // mul_sup | |||
| 9090 | ||||
| 9091 | #if 0 | |||
| 9092 | SIMD_INTRIN_DEF(div_u32){ "div_u32", simd__intrin_div_u32, 0x0001, ((void*)0) } , | |||
| 9093 | #endif // div_sup | |||
| 9094 | ||||
| 9095 | #if 1 | |||
| 9096 | SIMD_INTRIN_DEF(divisor_u32){ "divisor_u32", simd__intrin_divisor_u32, 0x0001, ((void*)0) } , | |||
| 9097 | SIMD_INTRIN_DEF(divc_u32){ "divc_u32", simd__intrin_divc_u32, 0x0001, ((void*)0) } , | |||
| 9098 | #endif // intdiv_sup | |||
| 9099 | ||||
| 9100 | #if 0 | |||
| 9101 | #line 602 | |||
| 9102 | SIMD_INTRIN_DEF(muladd_u32){ "muladd_u32", simd__intrin_muladd_u32, 0x0001, ((void*)0) } , | |||
| 9103 | ||||
| 9104 | #line 602 | |||
| 9105 | SIMD_INTRIN_DEF(mulsub_u32){ "mulsub_u32", simd__intrin_mulsub_u32, 0x0001, ((void*)0) } , | |||
| 9106 | ||||
| 9107 | #line 602 | |||
| 9108 | SIMD_INTRIN_DEF(nmuladd_u32){ "nmuladd_u32", simd__intrin_nmuladd_u32, 0x0001, ((void*)0) } , | |||
| 9109 | ||||
| 9110 | #line 602 | |||
| 9111 | SIMD_INTRIN_DEF(nmulsub_u32){ "nmulsub_u32", simd__intrin_nmulsub_u32, 0x0001, ((void*)0) } , | |||
| 9112 | ||||
| 9113 | #endif // fused_sup | |||
| 9114 | ||||
| 9115 | #if 1 | |||
| 9116 | SIMD_INTRIN_DEF(sum_u32){ "sum_u32", simd__intrin_sum_u32, 0x0001, ((void*)0) } , | |||
| 9117 | #endif // sum_sup | |||
| 9118 | ||||
| 9119 | #if 0 | |||
| 9120 | SIMD_INTRIN_DEF(sumup_u32){ "sumup_u32", simd__intrin_sumup_u32, 0x0001, ((void*)0) } , | |||
| 9121 | #endif // sumup_sup | |||
| 9122 | /*************************** | |||
| 9123 | * Math | |||
| 9124 | ***************************/ | |||
| 9125 | #if 0 | |||
| 9126 | #line 620 | |||
| 9127 | SIMD_INTRIN_DEF(sqrt_u32){ "sqrt_u32", simd__intrin_sqrt_u32, 0x0001, ((void*)0) } , | |||
| 9128 | ||||
| 9129 | #line 620 | |||
| 9130 | SIMD_INTRIN_DEF(recip_u32){ "recip_u32", simd__intrin_recip_u32, 0x0001, ((void*)0) } , | |||
| 9131 | ||||
| 9132 | #line 620 | |||
| 9133 | SIMD_INTRIN_DEF(abs_u32){ "abs_u32", simd__intrin_abs_u32, 0x0001, ((void*)0) } , | |||
| 9134 | ||||
| 9135 | #line 620 | |||
| 9136 | SIMD_INTRIN_DEF(square_u32){ "square_u32", simd__intrin_square_u32, 0x0001, ((void*)0) } , | |||
| 9137 | ||||
| 9138 | #endif | |||
| 9139 | ||||
| 9140 | #line 627 | |||
| 9141 | SIMD_INTRIN_DEF(max_u32){ "max_u32", simd__intrin_max_u32, 0x0001, ((void*)0) } , | |||
| 9142 | ||||
| 9143 | #line 627 | |||
| 9144 | SIMD_INTRIN_DEF(min_u32){ "min_u32", simd__intrin_min_u32, 0x0001, ((void*)0) } , | |||
| 9145 | ||||
| 9146 | ||||
| 9147 | #if 0 | |||
| 9148 | #line 634 | |||
| 9149 | SIMD_INTRIN_DEF(maxp_u32){ "maxp_u32", simd__intrin_maxp_u32, 0x0001, ((void*)0) } , | |||
| 9150 | ||||
| 9151 | #line 634 | |||
| 9152 | SIMD_INTRIN_DEF(minp_u32){ "minp_u32", simd__intrin_minp_u32, 0x0001, ((void*)0) } , | |||
| 9153 | ||||
| 9154 | #endif | |||
| 9155 | ||||
| 9156 | /*************************** | |||
| 9157 | * Mask operations | |||
| 9158 | ***************************/ | |||
| 9159 | #line 644 | |||
| 9160 | SIMD_INTRIN_DEF(ifadd_u32){ "ifadd_u32", simd__intrin_ifadd_u32, 0x0001, ((void*)0) } , | |||
| 9161 | ||||
| 9162 | #line 644 | |||
| 9163 | SIMD_INTRIN_DEF(ifsub_u32){ "ifsub_u32", simd__intrin_ifsub_u32, 0x0001, ((void*)0) } , | |||
| 9164 | ||||
| 9165 | ||||
| 9166 | #endif // simd_sup | |||
| 9167 | ||||
| 9168 | #line 489 | |||
| 9169 | #if 1 | |||
| 9170 | ||||
| 9171 | /*************************** | |||
| 9172 | * Memory | |||
| 9173 | ***************************/ | |||
| 9174 | #line 497 | |||
| 9175 | SIMD_INTRIN_DEF(load_s32){ "load_s32", simd__intrin_load_s32, 0x0001, ((void*)0) } , | |||
| 9176 | ||||
| 9177 | #line 497 | |||
| 9178 | SIMD_INTRIN_DEF(loada_s32){ "loada_s32", simd__intrin_loada_s32, 0x0001, ((void*)0) } , | |||
| 9179 | ||||
| 9180 | #line 497 | |||
| 9181 | SIMD_INTRIN_DEF(loads_s32){ "loads_s32", simd__intrin_loads_s32, 0x0001, ((void*)0) } , | |||
| 9182 | ||||
| 9183 | #line 497 | |||
| 9184 | SIMD_INTRIN_DEF(loadl_s32){ "loadl_s32", simd__intrin_loadl_s32, 0x0001, ((void*)0) } , | |||
| 9185 | ||||
| 9186 | #line 497 | |||
| 9187 | SIMD_INTRIN_DEF(store_s32){ "store_s32", simd__intrin_store_s32, 0x0001, ((void*)0) } , | |||
| 9188 | ||||
| 9189 | #line 497 | |||
| 9190 | SIMD_INTRIN_DEF(storea_s32){ "storea_s32", simd__intrin_storea_s32, 0x0001, ((void*)0) } , | |||
| 9191 | ||||
| 9192 | #line 497 | |||
| 9193 | SIMD_INTRIN_DEF(stores_s32){ "stores_s32", simd__intrin_stores_s32, 0x0001, ((void*)0) } , | |||
| 9194 | ||||
| 9195 | #line 497 | |||
| 9196 | SIMD_INTRIN_DEF(storel_s32){ "storel_s32", simd__intrin_storel_s32, 0x0001, ((void*)0) } , | |||
| 9197 | ||||
| 9198 | #line 497 | |||
| 9199 | SIMD_INTRIN_DEF(storeh_s32){ "storeh_s32", simd__intrin_storeh_s32, 0x0001, ((void*)0) } , | |||
| 9200 | ||||
| 9201 | ||||
| 9202 | /**************************************** | |||
| 9203 | * Non-contiguous/Partial Memory access | |||
| 9204 | ****************************************/ | |||
| 9205 | #if 1 | |||
| 9206 | #line 508 | |||
| 9207 | SIMD_INTRIN_DEF(load_till_s32){ "load_till_s32", simd__intrin_load_till_s32, 0x0001, ((void *)0) } , | |||
| 9208 | ||||
| 9209 | #line 508 | |||
| 9210 | SIMD_INTRIN_DEF(load_tillz_s32){ "load_tillz_s32", simd__intrin_load_tillz_s32, 0x0001, ((void *)0) } , | |||
| 9211 | ||||
| 9212 | #line 508 | |||
| 9213 | SIMD_INTRIN_DEF(loadn_s32){ "loadn_s32", simd__intrin_loadn_s32, 0x0001, ((void*)0) } , | |||
| 9214 | ||||
| 9215 | #line 508 | |||
| 9216 | SIMD_INTRIN_DEF(loadn_till_s32){ "loadn_till_s32", simd__intrin_loadn_till_s32, 0x0001, ((void *)0) } , | |||
| 9217 | ||||
| 9218 | #line 508 | |||
| 9219 | SIMD_INTRIN_DEF(loadn_tillz_s32){ "loadn_tillz_s32", simd__intrin_loadn_tillz_s32, 0x0001, (( void*)0) } , | |||
| 9220 | ||||
| 9221 | #line 508 | |||
| 9222 | SIMD_INTRIN_DEF(store_till_s32){ "store_till_s32", simd__intrin_store_till_s32, 0x0001, ((void *)0) } , | |||
| 9223 | ||||
| 9224 | #line 508 | |||
| 9225 | SIMD_INTRIN_DEF(storen_s32){ "storen_s32", simd__intrin_storen_s32, 0x0001, ((void*)0) } , | |||
| 9226 | ||||
| 9227 | #line 508 | |||
| 9228 | SIMD_INTRIN_DEF(storen_till_s32){ "storen_till_s32", simd__intrin_storen_till_s32, 0x0001, (( void*)0) } , | |||
| 9229 | ||||
| 9230 | #endif // ncont_sup | |||
| 9231 | ||||
| 9232 | /*************************** | |||
| 9233 | * Misc | |||
| 9234 | ***************************/ | |||
| 9235 | #line 519 | |||
| 9236 | #if 1 | |||
| 9237 | SIMD_INTRIN_DEF(reinterpret_u8_s32){ "reinterpret_u8_s32", simd__intrin_reinterpret_u8_s32, 0x0001 , ((void*)0) } , | |||
| 9238 | #endif // simd_sup2 | |||
| 9239 | ||||
| 9240 | #line 519 | |||
| 9241 | #if 1 | |||
| 9242 | SIMD_INTRIN_DEF(reinterpret_s8_s32){ "reinterpret_s8_s32", simd__intrin_reinterpret_s8_s32, 0x0001 , ((void*)0) } , | |||
| 9243 | #endif // simd_sup2 | |||
| 9244 | ||||
| 9245 | #line 519 | |||
| 9246 | #if 1 | |||
| 9247 | SIMD_INTRIN_DEF(reinterpret_u16_s32){ "reinterpret_u16_s32", simd__intrin_reinterpret_u16_s32, 0x0001 , ((void*)0) } , | |||
| 9248 | #endif // simd_sup2 | |||
| 9249 | ||||
| 9250 | #line 519 | |||
| 9251 | #if 1 | |||
| 9252 | SIMD_INTRIN_DEF(reinterpret_s16_s32){ "reinterpret_s16_s32", simd__intrin_reinterpret_s16_s32, 0x0001 , ((void*)0) } , | |||
| 9253 | #endif // simd_sup2 | |||
| 9254 | ||||
| 9255 | #line 519 | |||
| 9256 | #if 1 | |||
| 9257 | SIMD_INTRIN_DEF(reinterpret_u32_s32){ "reinterpret_u32_s32", simd__intrin_reinterpret_u32_s32, 0x0001 , ((void*)0) } , | |||
| 9258 | #endif // simd_sup2 | |||
| 9259 | ||||
| 9260 | #line 519 | |||
| 9261 | #if 1 | |||
| 9262 | SIMD_INTRIN_DEF(reinterpret_s32_s32){ "reinterpret_s32_s32", simd__intrin_reinterpret_s32_s32, 0x0001 , ((void*)0) } , | |||
| 9263 | #endif // simd_sup2 | |||
| 9264 | ||||
| 9265 | #line 519 | |||
| 9266 | #if 1 | |||
| 9267 | SIMD_INTRIN_DEF(reinterpret_u64_s32){ "reinterpret_u64_s32", simd__intrin_reinterpret_u64_s32, 0x0001 , ((void*)0) } , | |||
| 9268 | #endif // simd_sup2 | |||
| 9269 | ||||
| 9270 | #line 519 | |||
| 9271 | #if 1 | |||
| 9272 | SIMD_INTRIN_DEF(reinterpret_s64_s32){ "reinterpret_s64_s32", simd__intrin_reinterpret_s64_s32, 0x0001 , ((void*)0) } , | |||
| 9273 | #endif // simd_sup2 | |||
| 9274 | ||||
| 9275 | #line 519 | |||
| 9276 | #if 1 | |||
| 9277 | SIMD_INTRIN_DEF(reinterpret_f32_s32){ "reinterpret_f32_s32", simd__intrin_reinterpret_f32_s32, 0x0001 , ((void*)0) } , | |||
| 9278 | #endif // simd_sup2 | |||
| 9279 | ||||
| 9280 | #line 519 | |||
| 9281 | #if NPY_SIMD_F641 | |||
| 9282 | SIMD_INTRIN_DEF(reinterpret_f64_s32){ "reinterpret_f64_s32", simd__intrin_reinterpret_f64_s32, 0x0001 , ((void*)0) } , | |||
| 9283 | #endif // simd_sup2 | |||
| 9284 | ||||
| 9285 | ||||
| 9286 | #line 527 | |||
| 9287 | SIMD_INTRIN_DEF(set_s32){ "set_s32", simd__intrin_set_s32, 0x0001, ((void*)0) } , | |||
| 9288 | ||||
| 9289 | #line 527 | |||
| 9290 | SIMD_INTRIN_DEF(setf_s32){ "setf_s32", simd__intrin_setf_s32, 0x0001, ((void*)0) } , | |||
| 9291 | ||||
| 9292 | #line 527 | |||
| 9293 | SIMD_INTRIN_DEF(setall_s32){ "setall_s32", simd__intrin_setall_s32, 0x0001, ((void*)0) } , | |||
| 9294 | ||||
| 9295 | #line 527 | |||
| 9296 | SIMD_INTRIN_DEF(zero_s32){ "zero_s32", simd__intrin_zero_s32, 0x0001, ((void*)0) } , | |||
| 9297 | ||||
| 9298 | #line 527 | |||
| 9299 | SIMD_INTRIN_DEF(select_s32){ "select_s32", simd__intrin_select_s32, 0x0001, ((void*)0) } , | |||
| 9300 | ||||
| 9301 | ||||
| 9302 | /*************************** | |||
| 9303 | * Reorder | |||
| 9304 | ***************************/ | |||
| 9305 | #line 536 | |||
| 9306 | SIMD_INTRIN_DEF(combinel_s32){ "combinel_s32", simd__intrin_combinel_s32, 0x0001, ((void*) 0) } , | |||
| 9307 | ||||
| 9308 | #line 536 | |||
| 9309 | SIMD_INTRIN_DEF(combineh_s32){ "combineh_s32", simd__intrin_combineh_s32, 0x0001, ((void*) 0) } , | |||
| 9310 | ||||
| 9311 | #line 536 | |||
| 9312 | SIMD_INTRIN_DEF(combine_s32){ "combine_s32", simd__intrin_combine_s32, 0x0001, ((void*)0) } , | |||
| 9313 | ||||
| 9314 | #line 536 | |||
| 9315 | SIMD_INTRIN_DEF(zip_s32){ "zip_s32", simd__intrin_zip_s32, 0x0001, ((void*)0) } , | |||
| 9316 | ||||
| 9317 | ||||
| 9318 | #if 1 | |||
| 9319 | SIMD_INTRIN_DEF(rev64_s32){ "rev64_s32", simd__intrin_rev64_s32, 0x0001, ((void*)0) } , | |||
| 9320 | #endif | |||
| 9321 | ||||
| 9322 | /*************************** | |||
| 9323 | * Operators | |||
| 9324 | ***************************/ | |||
| 9325 | #if 31 > 0 | |||
| 9326 | #line 550 | |||
| 9327 | SIMD_INTRIN_DEF(shl_s32){ "shl_s32", simd__intrin_shl_s32, 0x0001, ((void*)0) } , | |||
| 9328 | ||||
| 9329 | #line 550 | |||
| 9330 | SIMD_INTRIN_DEF(shr_s32){ "shr_s32", simd__intrin_shr_s32, 0x0001, ((void*)0) } , | |||
| 9331 | ||||
| 9332 | #line 550 | |||
| 9333 | SIMD_INTRIN_DEF(shli_s32){ "shli_s32", simd__intrin_shli_s32, 0x0001, ((void*)0) } , | |||
| 9334 | ||||
| 9335 | #line 550 | |||
| 9336 | SIMD_INTRIN_DEF(shri_s32){ "shri_s32", simd__intrin_shri_s32, 0x0001, ((void*)0) } , | |||
| 9337 | ||||
| 9338 | #endif // shl_imm | |||
| 9339 | ||||
| 9340 | #line 557 | |||
| 9341 | SIMD_INTRIN_DEF(and_s32){ "and_s32", simd__intrin_and_s32, 0x0001, ((void*)0) } , | |||
| 9342 | ||||
| 9343 | #line 557 | |||
| 9344 | SIMD_INTRIN_DEF(or_s32){ "or_s32", simd__intrin_or_s32, 0x0001, ((void*)0) } , | |||
| 9345 | ||||
| 9346 | #line 557 | |||
| 9347 | SIMD_INTRIN_DEF(xor_s32){ "xor_s32", simd__intrin_xor_s32, 0x0001, ((void*)0) } , | |||
| 9348 | ||||
| 9349 | #line 557 | |||
| 9350 | SIMD_INTRIN_DEF(not_s32){ "not_s32", simd__intrin_not_s32, 0x0001, ((void*)0) } , | |||
| 9351 | ||||
| 9352 | #line 557 | |||
| 9353 | SIMD_INTRIN_DEF(cmpeq_s32){ "cmpeq_s32", simd__intrin_cmpeq_s32, 0x0001, ((void*)0) } , | |||
| 9354 | ||||
| 9355 | #line 557 | |||
| 9356 | SIMD_INTRIN_DEF(cmpneq_s32){ "cmpneq_s32", simd__intrin_cmpneq_s32, 0x0001, ((void*)0) } , | |||
| 9357 | ||||
| 9358 | #line 557 | |||
| 9359 | SIMD_INTRIN_DEF(cmpgt_s32){ "cmpgt_s32", simd__intrin_cmpgt_s32, 0x0001, ((void*)0) } , | |||
| 9360 | ||||
| 9361 | #line 557 | |||
| 9362 | SIMD_INTRIN_DEF(cmpge_s32){ "cmpge_s32", simd__intrin_cmpge_s32, 0x0001, ((void*)0) } , | |||
| 9363 | ||||
| 9364 | #line 557 | |||
| 9365 | SIMD_INTRIN_DEF(cmplt_s32){ "cmplt_s32", simd__intrin_cmplt_s32, 0x0001, ((void*)0) } , | |||
| 9366 | ||||
| 9367 | #line 557 | |||
| 9368 | SIMD_INTRIN_DEF(cmple_s32){ "cmple_s32", simd__intrin_cmple_s32, 0x0001, ((void*)0) } , | |||
| 9369 | ||||
| 9370 | ||||
| 9371 | /*************************** | |||
| 9372 | * Conversion | |||
| 9373 | ***************************/ | |||
| 9374 | SIMD_INTRIN_DEF(cvt_s32_b32){ "cvt_s32_b32", simd__intrin_cvt_s32_b32, 0x0001, ((void*)0) } , | |||
| 9375 | SIMD_INTRIN_DEF(cvt_b32_s32){ "cvt_b32_s32", simd__intrin_cvt_b32_s32, 0x0001, ((void*)0) } , | |||
| 9376 | #if 0 | |||
| 9377 | SIMD_INTRIN_DEF(expand_s32_s32){ "expand_s32_s32", simd__intrin_expand_s32_s32, 0x0001, ((void *)0) } , | |||
| 9378 | #endif // expand_sup | |||
| 9379 | /*************************** | |||
| 9380 | * Arithmetic | |||
| 9381 | ***************************/ | |||
| 9382 | #line 574 | |||
| 9383 | SIMD_INTRIN_DEF(add_s32){ "add_s32", simd__intrin_add_s32, 0x0001, ((void*)0) } , | |||
| 9384 | ||||
| 9385 | #line 574 | |||
| 9386 | SIMD_INTRIN_DEF(sub_s32){ "sub_s32", simd__intrin_sub_s32, 0x0001, ((void*)0) } , | |||
| 9387 | ||||
| 9388 | ||||
| 9389 | #if 0 | |||
| 9390 | #line 581 | |||
| 9391 | SIMD_INTRIN_DEF(adds_s32){ "adds_s32", simd__intrin_adds_s32, 0x0001, ((void*)0) } , | |||
| 9392 | ||||
| 9393 | #line 581 | |||
| 9394 | SIMD_INTRIN_DEF(subs_s32){ "subs_s32", simd__intrin_subs_s32, 0x0001, ((void*)0) } , | |||
| 9395 | ||||
| 9396 | #endif // sat_sup | |||
| 9397 | ||||
| 9398 | #if 1 | |||
| 9399 | SIMD_INTRIN_DEF(mul_s32){ "mul_s32", simd__intrin_mul_s32, 0x0001, ((void*)0) } , | |||
| 9400 | #endif // mul_sup | |||
| 9401 | ||||
| 9402 | #if 0 | |||
| 9403 | SIMD_INTRIN_DEF(div_s32){ "div_s32", simd__intrin_div_s32, 0x0001, ((void*)0) } , | |||
| 9404 | #endif // div_sup | |||
| 9405 | ||||
| 9406 | #if 1 | |||
| 9407 | SIMD_INTRIN_DEF(divisor_s32){ "divisor_s32", simd__intrin_divisor_s32, 0x0001, ((void*)0) } , | |||
| 9408 | SIMD_INTRIN_DEF(divc_s32){ "divc_s32", simd__intrin_divc_s32, 0x0001, ((void*)0) } , | |||
| 9409 | #endif // intdiv_sup | |||
| 9410 | ||||
| 9411 | #if 0 | |||
| 9412 | #line 602 | |||
| 9413 | SIMD_INTRIN_DEF(muladd_s32){ "muladd_s32", simd__intrin_muladd_s32, 0x0001, ((void*)0) } , | |||
| 9414 | ||||
| 9415 | #line 602 | |||
| 9416 | SIMD_INTRIN_DEF(mulsub_s32){ "mulsub_s32", simd__intrin_mulsub_s32, 0x0001, ((void*)0) } , | |||
| 9417 | ||||
| 9418 | #line 602 | |||
| 9419 | SIMD_INTRIN_DEF(nmuladd_s32){ "nmuladd_s32", simd__intrin_nmuladd_s32, 0x0001, ((void*)0) } , | |||
| 9420 | ||||
| 9421 | #line 602 | |||
| 9422 | SIMD_INTRIN_DEF(nmulsub_s32){ "nmulsub_s32", simd__intrin_nmulsub_s32, 0x0001, ((void*)0) } , | |||
| 9423 | ||||
| 9424 | #endif // fused_sup | |||
| 9425 | ||||
| 9426 | #if 0 | |||
| 9427 | SIMD_INTRIN_DEF(sum_s32){ "sum_s32", simd__intrin_sum_s32, 0x0001, ((void*)0) } , | |||
| 9428 | #endif // sum_sup | |||
| 9429 | ||||
| 9430 | #if 0 | |||
| 9431 | SIMD_INTRIN_DEF(sumup_s32){ "sumup_s32", simd__intrin_sumup_s32, 0x0001, ((void*)0) } , | |||
| 9432 | #endif // sumup_sup | |||
| 9433 | /*************************** | |||
| 9434 | * Math | |||
| 9435 | ***************************/ | |||
| 9436 | #if 0 | |||
| 9437 | #line 620 | |||
| 9438 | SIMD_INTRIN_DEF(sqrt_s32){ "sqrt_s32", simd__intrin_sqrt_s32, 0x0001, ((void*)0) } , | |||
| 9439 | ||||
| 9440 | #line 620 | |||
| 9441 | SIMD_INTRIN_DEF(recip_s32){ "recip_s32", simd__intrin_recip_s32, 0x0001, ((void*)0) } , | |||
| 9442 | ||||
| 9443 | #line 620 | |||
| 9444 | SIMD_INTRIN_DEF(abs_s32){ "abs_s32", simd__intrin_abs_s32, 0x0001, ((void*)0) } , | |||
| 9445 | ||||
| 9446 | #line 620 | |||
| 9447 | SIMD_INTRIN_DEF(square_s32){ "square_s32", simd__intrin_square_s32, 0x0001, ((void*)0) } , | |||
| 9448 | ||||
| 9449 | #endif | |||
| 9450 | ||||
| 9451 | #line 627 | |||
| 9452 | SIMD_INTRIN_DEF(max_s32){ "max_s32", simd__intrin_max_s32, 0x0001, ((void*)0) } , | |||
| 9453 | ||||
| 9454 | #line 627 | |||
| 9455 | SIMD_INTRIN_DEF(min_s32){ "min_s32", simd__intrin_min_s32, 0x0001, ((void*)0) } , | |||
| 9456 | ||||
| 9457 | ||||
| 9458 | #if 0 | |||
| 9459 | #line 634 | |||
| 9460 | SIMD_INTRIN_DEF(maxp_s32){ "maxp_s32", simd__intrin_maxp_s32, 0x0001, ((void*)0) } , | |||
| 9461 | ||||
| 9462 | #line 634 | |||
| 9463 | SIMD_INTRIN_DEF(minp_s32){ "minp_s32", simd__intrin_minp_s32, 0x0001, ((void*)0) } , | |||
| 9464 | ||||
| 9465 | #endif | |||
| 9466 | ||||
| 9467 | /*************************** | |||
| 9468 | * Mask operations | |||
| 9469 | ***************************/ | |||
| 9470 | #line 644 | |||
| 9471 | SIMD_INTRIN_DEF(ifadd_s32){ "ifadd_s32", simd__intrin_ifadd_s32, 0x0001, ((void*)0) } , | |||
| 9472 | ||||
| 9473 | #line 644 | |||
| 9474 | SIMD_INTRIN_DEF(ifsub_s32){ "ifsub_s32", simd__intrin_ifsub_s32, 0x0001, ((void*)0) } , | |||
| 9475 | ||||
| 9476 | ||||
| 9477 | #endif // simd_sup | |||
| 9478 | ||||
| 9479 | #line 489 | |||
| 9480 | #if 1 | |||
| 9481 | ||||
| 9482 | /*************************** | |||
| 9483 | * Memory | |||
| 9484 | ***************************/ | |||
| 9485 | #line 497 | |||
| 9486 | SIMD_INTRIN_DEF(load_u64){ "load_u64", simd__intrin_load_u64, 0x0001, ((void*)0) } , | |||
| 9487 | ||||
| 9488 | #line 497 | |||
| 9489 | SIMD_INTRIN_DEF(loada_u64){ "loada_u64", simd__intrin_loada_u64, 0x0001, ((void*)0) } , | |||
| 9490 | ||||
| 9491 | #line 497 | |||
| 9492 | SIMD_INTRIN_DEF(loads_u64){ "loads_u64", simd__intrin_loads_u64, 0x0001, ((void*)0) } , | |||
| 9493 | ||||
| 9494 | #line 497 | |||
| 9495 | SIMD_INTRIN_DEF(loadl_u64){ "loadl_u64", simd__intrin_loadl_u64, 0x0001, ((void*)0) } , | |||
| 9496 | ||||
| 9497 | #line 497 | |||
| 9498 | SIMD_INTRIN_DEF(store_u64){ "store_u64", simd__intrin_store_u64, 0x0001, ((void*)0) } , | |||
| 9499 | ||||
| 9500 | #line 497 | |||
| 9501 | SIMD_INTRIN_DEF(storea_u64){ "storea_u64", simd__intrin_storea_u64, 0x0001, ((void*)0) } , | |||
| 9502 | ||||
| 9503 | #line 497 | |||
| 9504 | SIMD_INTRIN_DEF(stores_u64){ "stores_u64", simd__intrin_stores_u64, 0x0001, ((void*)0) } , | |||
| 9505 | ||||
| 9506 | #line 497 | |||
| 9507 | SIMD_INTRIN_DEF(storel_u64){ "storel_u64", simd__intrin_storel_u64, 0x0001, ((void*)0) } , | |||
| 9508 | ||||
| 9509 | #line 497 | |||
| 9510 | SIMD_INTRIN_DEF(storeh_u64){ "storeh_u64", simd__intrin_storeh_u64, 0x0001, ((void*)0) } , | |||
| 9511 | ||||
| 9512 | ||||
| 9513 | /**************************************** | |||
| 9514 | * Non-contiguous/Partial Memory access | |||
| 9515 | ****************************************/ | |||
| 9516 | #if 1 | |||
| 9517 | #line 508 | |||
| 9518 | SIMD_INTRIN_DEF(load_till_u64){ "load_till_u64", simd__intrin_load_till_u64, 0x0001, ((void *)0) } , | |||
| 9519 | ||||
| 9520 | #line 508 | |||
| 9521 | SIMD_INTRIN_DEF(load_tillz_u64){ "load_tillz_u64", simd__intrin_load_tillz_u64, 0x0001, ((void *)0) } , | |||
| 9522 | ||||
| 9523 | #line 508 | |||
| 9524 | SIMD_INTRIN_DEF(loadn_u64){ "loadn_u64", simd__intrin_loadn_u64, 0x0001, ((void*)0) } , | |||
| 9525 | ||||
| 9526 | #line 508 | |||
| 9527 | SIMD_INTRIN_DEF(loadn_till_u64){ "loadn_till_u64", simd__intrin_loadn_till_u64, 0x0001, ((void *)0) } , | |||
| 9528 | ||||
| 9529 | #line 508 | |||
| 9530 | SIMD_INTRIN_DEF(loadn_tillz_u64){ "loadn_tillz_u64", simd__intrin_loadn_tillz_u64, 0x0001, (( void*)0) } , | |||
| 9531 | ||||
| 9532 | #line 508 | |||
| 9533 | SIMD_INTRIN_DEF(store_till_u64){ "store_till_u64", simd__intrin_store_till_u64, 0x0001, ((void *)0) } , | |||
| 9534 | ||||
| 9535 | #line 508 | |||
| 9536 | SIMD_INTRIN_DEF(storen_u64){ "storen_u64", simd__intrin_storen_u64, 0x0001, ((void*)0) } , | |||
| 9537 | ||||
| 9538 | #line 508 | |||
| 9539 | SIMD_INTRIN_DEF(storen_till_u64){ "storen_till_u64", simd__intrin_storen_till_u64, 0x0001, (( void*)0) } , | |||
| 9540 | ||||
| 9541 | #endif // ncont_sup | |||
| 9542 | ||||
| 9543 | /*************************** | |||
| 9544 | * Misc | |||
| 9545 | ***************************/ | |||
| 9546 | #line 519 | |||
| 9547 | #if 1 | |||
| 9548 | SIMD_INTRIN_DEF(reinterpret_u8_u64){ "reinterpret_u8_u64", simd__intrin_reinterpret_u8_u64, 0x0001 , ((void*)0) } , | |||
| 9549 | #endif // simd_sup2 | |||
| 9550 | ||||
| 9551 | #line 519 | |||
| 9552 | #if 1 | |||
| 9553 | SIMD_INTRIN_DEF(reinterpret_s8_u64){ "reinterpret_s8_u64", simd__intrin_reinterpret_s8_u64, 0x0001 , ((void*)0) } , | |||
| 9554 | #endif // simd_sup2 | |||
| 9555 | ||||
| 9556 | #line 519 | |||
| 9557 | #if 1 | |||
| 9558 | SIMD_INTRIN_DEF(reinterpret_u16_u64){ "reinterpret_u16_u64", simd__intrin_reinterpret_u16_u64, 0x0001 , ((void*)0) } , | |||
| 9559 | #endif // simd_sup2 | |||
| 9560 | ||||
| 9561 | #line 519 | |||
| 9562 | #if 1 | |||
| 9563 | SIMD_INTRIN_DEF(reinterpret_s16_u64){ "reinterpret_s16_u64", simd__intrin_reinterpret_s16_u64, 0x0001 , ((void*)0) } , | |||
| 9564 | #endif // simd_sup2 | |||
| 9565 | ||||
| 9566 | #line 519 | |||
| 9567 | #if 1 | |||
| 9568 | SIMD_INTRIN_DEF(reinterpret_u32_u64){ "reinterpret_u32_u64", simd__intrin_reinterpret_u32_u64, 0x0001 , ((void*)0) } , | |||
| 9569 | #endif // simd_sup2 | |||
| 9570 | ||||
| 9571 | #line 519 | |||
| 9572 | #if 1 | |||
| 9573 | SIMD_INTRIN_DEF(reinterpret_s32_u64){ "reinterpret_s32_u64", simd__intrin_reinterpret_s32_u64, 0x0001 , ((void*)0) } , | |||
| 9574 | #endif // simd_sup2 | |||
| 9575 | ||||
| 9576 | #line 519 | |||
| 9577 | #if 1 | |||
| 9578 | SIMD_INTRIN_DEF(reinterpret_u64_u64){ "reinterpret_u64_u64", simd__intrin_reinterpret_u64_u64, 0x0001 , ((void*)0) } , | |||
| 9579 | #endif // simd_sup2 | |||
| 9580 | ||||
| 9581 | #line 519 | |||
| 9582 | #if 1 | |||
| 9583 | SIMD_INTRIN_DEF(reinterpret_s64_u64){ "reinterpret_s64_u64", simd__intrin_reinterpret_s64_u64, 0x0001 , ((void*)0) } , | |||
| 9584 | #endif // simd_sup2 | |||
| 9585 | ||||
| 9586 | #line 519 | |||
| 9587 | #if 1 | |||
| 9588 | SIMD_INTRIN_DEF(reinterpret_f32_u64){ "reinterpret_f32_u64", simd__intrin_reinterpret_f32_u64, 0x0001 , ((void*)0) } , | |||
| 9589 | #endif // simd_sup2 | |||
| 9590 | ||||
| 9591 | #line 519 | |||
| 9592 | #if NPY_SIMD_F641 | |||
| 9593 | SIMD_INTRIN_DEF(reinterpret_f64_u64){ "reinterpret_f64_u64", simd__intrin_reinterpret_f64_u64, 0x0001 , ((void*)0) } , | |||
| 9594 | #endif // simd_sup2 | |||
| 9595 | ||||
| 9596 | ||||
| 9597 | #line 527 | |||
| 9598 | SIMD_INTRIN_DEF(set_u64){ "set_u64", simd__intrin_set_u64, 0x0001, ((void*)0) } , | |||
| 9599 | ||||
| 9600 | #line 527 | |||
| 9601 | SIMD_INTRIN_DEF(setf_u64){ "setf_u64", simd__intrin_setf_u64, 0x0001, ((void*)0) } , | |||
| 9602 | ||||
| 9603 | #line 527 | |||
| 9604 | SIMD_INTRIN_DEF(setall_u64){ "setall_u64", simd__intrin_setall_u64, 0x0001, ((void*)0) } , | |||
| 9605 | ||||
| 9606 | #line 527 | |||
| 9607 | SIMD_INTRIN_DEF(zero_u64){ "zero_u64", simd__intrin_zero_u64, 0x0001, ((void*)0) } , | |||
| 9608 | ||||
| 9609 | #line 527 | |||
| 9610 | SIMD_INTRIN_DEF(select_u64){ "select_u64", simd__intrin_select_u64, 0x0001, ((void*)0) } , | |||
| 9611 | ||||
| 9612 | ||||
| 9613 | /*************************** | |||
| 9614 | * Reorder | |||
| 9615 | ***************************/ | |||
| 9616 | #line 536 | |||
| 9617 | SIMD_INTRIN_DEF(combinel_u64){ "combinel_u64", simd__intrin_combinel_u64, 0x0001, ((void*) 0) } , | |||
| 9618 | ||||
| 9619 | #line 536 | |||
| 9620 | SIMD_INTRIN_DEF(combineh_u64){ "combineh_u64", simd__intrin_combineh_u64, 0x0001, ((void*) 0) } , | |||
| 9621 | ||||
| 9622 | #line 536 | |||
| 9623 | SIMD_INTRIN_DEF(combine_u64){ "combine_u64", simd__intrin_combine_u64, 0x0001, ((void*)0) } , | |||
| 9624 | ||||
| 9625 | #line 536 | |||
| 9626 | SIMD_INTRIN_DEF(zip_u64){ "zip_u64", simd__intrin_zip_u64, 0x0001, ((void*)0) } , | |||
| 9627 | ||||
| 9628 | ||||
| 9629 | #if 0 | |||
| 9630 | SIMD_INTRIN_DEF(rev64_u64){ "rev64_u64", simd__intrin_rev64_u64, 0x0001, ((void*)0) } , | |||
| 9631 | #endif | |||
| 9632 | ||||
| 9633 | /*************************** | |||
| 9634 | * Operators | |||
| 9635 | ***************************/ | |||
| 9636 | #if 63 > 0 | |||
| 9637 | #line 550 | |||
| 9638 | SIMD_INTRIN_DEF(shl_u64){ "shl_u64", simd__intrin_shl_u64, 0x0001, ((void*)0) } , | |||
| 9639 | ||||
| 9640 | #line 550 | |||
| 9641 | SIMD_INTRIN_DEF(shr_u64){ "shr_u64", simd__intrin_shr_u64, 0x0001, ((void*)0) } , | |||
| 9642 | ||||
| 9643 | #line 550 | |||
| 9644 | SIMD_INTRIN_DEF(shli_u64){ "shli_u64", simd__intrin_shli_u64, 0x0001, ((void*)0) } , | |||
| 9645 | ||||
| 9646 | #line 550 | |||
| 9647 | SIMD_INTRIN_DEF(shri_u64){ "shri_u64", simd__intrin_shri_u64, 0x0001, ((void*)0) } , | |||
| 9648 | ||||
| 9649 | #endif // shl_imm | |||
| 9650 | ||||
| 9651 | #line 557 | |||
| 9652 | SIMD_INTRIN_DEF(and_u64){ "and_u64", simd__intrin_and_u64, 0x0001, ((void*)0) } , | |||
| 9653 | ||||
| 9654 | #line 557 | |||
| 9655 | SIMD_INTRIN_DEF(or_u64){ "or_u64", simd__intrin_or_u64, 0x0001, ((void*)0) } , | |||
| 9656 | ||||
| 9657 | #line 557 | |||
| 9658 | SIMD_INTRIN_DEF(xor_u64){ "xor_u64", simd__intrin_xor_u64, 0x0001, ((void*)0) } , | |||
| 9659 | ||||
| 9660 | #line 557 | |||
| 9661 | SIMD_INTRIN_DEF(not_u64){ "not_u64", simd__intrin_not_u64, 0x0001, ((void*)0) } , | |||
| 9662 | ||||
| 9663 | #line 557 | |||
| 9664 | SIMD_INTRIN_DEF(cmpeq_u64){ "cmpeq_u64", simd__intrin_cmpeq_u64, 0x0001, ((void*)0) } , | |||
| 9665 | ||||
| 9666 | #line 557 | |||
| 9667 | SIMD_INTRIN_DEF(cmpneq_u64){ "cmpneq_u64", simd__intrin_cmpneq_u64, 0x0001, ((void*)0) } , | |||
| 9668 | ||||
| 9669 | #line 557 | |||
| 9670 | SIMD_INTRIN_DEF(cmpgt_u64){ "cmpgt_u64", simd__intrin_cmpgt_u64, 0x0001, ((void*)0) } , | |||
| 9671 | ||||
| 9672 | #line 557 | |||
| 9673 | SIMD_INTRIN_DEF(cmpge_u64){ "cmpge_u64", simd__intrin_cmpge_u64, 0x0001, ((void*)0) } , | |||
| 9674 | ||||
| 9675 | #line 557 | |||
| 9676 | SIMD_INTRIN_DEF(cmplt_u64){ "cmplt_u64", simd__intrin_cmplt_u64, 0x0001, ((void*)0) } , | |||
| 9677 | ||||
| 9678 | #line 557 | |||
| 9679 | SIMD_INTRIN_DEF(cmple_u64){ "cmple_u64", simd__intrin_cmple_u64, 0x0001, ((void*)0) } , | |||
| 9680 | ||||
| 9681 | ||||
| 9682 | /*************************** | |||
| 9683 | * Conversion | |||
| 9684 | ***************************/ | |||
| 9685 | SIMD_INTRIN_DEF(cvt_u64_b64){ "cvt_u64_b64", simd__intrin_cvt_u64_b64, 0x0001, ((void*)0) } , | |||
| 9686 | SIMD_INTRIN_DEF(cvt_b64_u64){ "cvt_b64_u64", simd__intrin_cvt_b64_u64, 0x0001, ((void*)0) } , | |||
| 9687 | #if 0 | |||
| 9688 | SIMD_INTRIN_DEF(expand_u64_u64){ "expand_u64_u64", simd__intrin_expand_u64_u64, 0x0001, ((void *)0) } , | |||
| 9689 | #endif // expand_sup | |||
| 9690 | /*************************** | |||
| 9691 | * Arithmetic | |||
| 9692 | ***************************/ | |||
| 9693 | #line 574 | |||
| 9694 | SIMD_INTRIN_DEF(add_u64){ "add_u64", simd__intrin_add_u64, 0x0001, ((void*)0) } , | |||
| 9695 | ||||
| 9696 | #line 574 | |||
| 9697 | SIMD_INTRIN_DEF(sub_u64){ "sub_u64", simd__intrin_sub_u64, 0x0001, ((void*)0) } , | |||
| 9698 | ||||
| 9699 | ||||
| 9700 | #if 0 | |||
| 9701 | #line 581 | |||
| 9702 | SIMD_INTRIN_DEF(adds_u64){ "adds_u64", simd__intrin_adds_u64, 0x0001, ((void*)0) } , | |||
| 9703 | ||||
| 9704 | #line 581 | |||
| 9705 | SIMD_INTRIN_DEF(subs_u64){ "subs_u64", simd__intrin_subs_u64, 0x0001, ((void*)0) } , | |||
| 9706 | ||||
| 9707 | #endif // sat_sup | |||
| 9708 | ||||
| 9709 | #if 0 | |||
| 9710 | SIMD_INTRIN_DEF(mul_u64){ "mul_u64", simd__intrin_mul_u64, 0x0001, ((void*)0) } , | |||
| 9711 | #endif // mul_sup | |||
| 9712 | ||||
| 9713 | #if 0 | |||
| 9714 | SIMD_INTRIN_DEF(div_u64){ "div_u64", simd__intrin_div_u64, 0x0001, ((void*)0) } , | |||
| 9715 | #endif // div_sup | |||
| 9716 | ||||
| 9717 | #if 1 | |||
| 9718 | SIMD_INTRIN_DEF(divisor_u64){ "divisor_u64", simd__intrin_divisor_u64, 0x0001, ((void*)0) } , | |||
| 9719 | SIMD_INTRIN_DEF(divc_u64){ "divc_u64", simd__intrin_divc_u64, 0x0001, ((void*)0) } , | |||
| 9720 | #endif // intdiv_sup | |||
| 9721 | ||||
| 9722 | #if 0 | |||
| 9723 | #line 602 | |||
| 9724 | SIMD_INTRIN_DEF(muladd_u64){ "muladd_u64", simd__intrin_muladd_u64, 0x0001, ((void*)0) } , | |||
| 9725 | ||||
| 9726 | #line 602 | |||
| 9727 | SIMD_INTRIN_DEF(mulsub_u64){ "mulsub_u64", simd__intrin_mulsub_u64, 0x0001, ((void*)0) } , | |||
| 9728 | ||||
| 9729 | #line 602 | |||
| 9730 | SIMD_INTRIN_DEF(nmuladd_u64){ "nmuladd_u64", simd__intrin_nmuladd_u64, 0x0001, ((void*)0) } , | |||
| 9731 | ||||
| 9732 | #line 602 | |||
| 9733 | SIMD_INTRIN_DEF(nmulsub_u64){ "nmulsub_u64", simd__intrin_nmulsub_u64, 0x0001, ((void*)0) } , | |||
| 9734 | ||||
| 9735 | #endif // fused_sup | |||
| 9736 | ||||
| 9737 | #if 1 | |||
| 9738 | SIMD_INTRIN_DEF(sum_u64){ "sum_u64", simd__intrin_sum_u64, 0x0001, ((void*)0) } , | |||
| 9739 | #endif // sum_sup | |||
| 9740 | ||||
| 9741 | #if 0 | |||
| 9742 | SIMD_INTRIN_DEF(sumup_u64){ "sumup_u64", simd__intrin_sumup_u64, 0x0001, ((void*)0) } , | |||
| 9743 | #endif // sumup_sup | |||
| 9744 | /*************************** | |||
| 9745 | * Math | |||
| 9746 | ***************************/ | |||
| 9747 | #if 0 | |||
| 9748 | #line 620 | |||
| 9749 | SIMD_INTRIN_DEF(sqrt_u64){ "sqrt_u64", simd__intrin_sqrt_u64, 0x0001, ((void*)0) } , | |||
| 9750 | ||||
| 9751 | #line 620 | |||
| 9752 | SIMD_INTRIN_DEF(recip_u64){ "recip_u64", simd__intrin_recip_u64, 0x0001, ((void*)0) } , | |||
| 9753 | ||||
| 9754 | #line 620 | |||
| 9755 | SIMD_INTRIN_DEF(abs_u64){ "abs_u64", simd__intrin_abs_u64, 0x0001, ((void*)0) } , | |||
| 9756 | ||||
| 9757 | #line 620 | |||
| 9758 | SIMD_INTRIN_DEF(square_u64){ "square_u64", simd__intrin_square_u64, 0x0001, ((void*)0) } , | |||
| 9759 | ||||
| 9760 | #endif | |||
| 9761 | ||||
| 9762 | #line 627 | |||
| 9763 | SIMD_INTRIN_DEF(max_u64){ "max_u64", simd__intrin_max_u64, 0x0001, ((void*)0) } , | |||
| 9764 | ||||
| 9765 | #line 627 | |||
| 9766 | SIMD_INTRIN_DEF(min_u64){ "min_u64", simd__intrin_min_u64, 0x0001, ((void*)0) } , | |||
| 9767 | ||||
| 9768 | ||||
| 9769 | #if 0 | |||
| 9770 | #line 634 | |||
| 9771 | SIMD_INTRIN_DEF(maxp_u64){ "maxp_u64", simd__intrin_maxp_u64, 0x0001, ((void*)0) } , | |||
| 9772 | ||||
| 9773 | #line 634 | |||
| 9774 | SIMD_INTRIN_DEF(minp_u64){ "minp_u64", simd__intrin_minp_u64, 0x0001, ((void*)0) } , | |||
| 9775 | ||||
| 9776 | #endif | |||
| 9777 | ||||
| 9778 | /*************************** | |||
| 9779 | * Mask operations | |||
| 9780 | ***************************/ | |||
| 9781 | #line 644 | |||
| 9782 | SIMD_INTRIN_DEF(ifadd_u64){ "ifadd_u64", simd__intrin_ifadd_u64, 0x0001, ((void*)0) } , | |||
| 9783 | ||||
| 9784 | #line 644 | |||
| 9785 | SIMD_INTRIN_DEF(ifsub_u64){ "ifsub_u64", simd__intrin_ifsub_u64, 0x0001, ((void*)0) } , | |||
| 9786 | ||||
| 9787 | ||||
| 9788 | #endif // simd_sup | |||
| 9789 | ||||
| 9790 | #line 489 | |||
| 9791 | #if 1 | |||
| 9792 | ||||
| 9793 | /*************************** | |||
| 9794 | * Memory | |||
| 9795 | ***************************/ | |||
| 9796 | #line 497 | |||
| 9797 | SIMD_INTRIN_DEF(load_s64){ "load_s64", simd__intrin_load_s64, 0x0001, ((void*)0) } , | |||
| 9798 | ||||
| 9799 | #line 497 | |||
| 9800 | SIMD_INTRIN_DEF(loada_s64){ "loada_s64", simd__intrin_loada_s64, 0x0001, ((void*)0) } , | |||
| 9801 | ||||
| 9802 | #line 497 | |||
| 9803 | SIMD_INTRIN_DEF(loads_s64){ "loads_s64", simd__intrin_loads_s64, 0x0001, ((void*)0) } , | |||
| 9804 | ||||
| 9805 | #line 497 | |||
| 9806 | SIMD_INTRIN_DEF(loadl_s64){ "loadl_s64", simd__intrin_loadl_s64, 0x0001, ((void*)0) } , | |||
| 9807 | ||||
| 9808 | #line 497 | |||
| 9809 | SIMD_INTRIN_DEF(store_s64){ "store_s64", simd__intrin_store_s64, 0x0001, ((void*)0) } , | |||
| 9810 | ||||
| 9811 | #line 497 | |||
| 9812 | SIMD_INTRIN_DEF(storea_s64){ "storea_s64", simd__intrin_storea_s64, 0x0001, ((void*)0) } , | |||
| 9813 | ||||
| 9814 | #line 497 | |||
| 9815 | SIMD_INTRIN_DEF(stores_s64){ "stores_s64", simd__intrin_stores_s64, 0x0001, ((void*)0) } , | |||
| 9816 | ||||
| 9817 | #line 497 | |||
| 9818 | SIMD_INTRIN_DEF(storel_s64){ "storel_s64", simd__intrin_storel_s64, 0x0001, ((void*)0) } , | |||
| 9819 | ||||
| 9820 | #line 497 | |||
| 9821 | SIMD_INTRIN_DEF(storeh_s64){ "storeh_s64", simd__intrin_storeh_s64, 0x0001, ((void*)0) } , | |||
| 9822 | ||||
| 9823 | ||||
| 9824 | /**************************************** | |||
| 9825 | * Non-contiguous/Partial Memory access | |||
| 9826 | ****************************************/ | |||
| 9827 | #if 1 | |||
| 9828 | #line 508 | |||
| 9829 | SIMD_INTRIN_DEF(load_till_s64){ "load_till_s64", simd__intrin_load_till_s64, 0x0001, ((void *)0) } , | |||
| 9830 | ||||
| 9831 | #line 508 | |||
| 9832 | SIMD_INTRIN_DEF(load_tillz_s64){ "load_tillz_s64", simd__intrin_load_tillz_s64, 0x0001, ((void *)0) } , | |||
| 9833 | ||||
| 9834 | #line 508 | |||
| 9835 | SIMD_INTRIN_DEF(loadn_s64){ "loadn_s64", simd__intrin_loadn_s64, 0x0001, ((void*)0) } , | |||
| 9836 | ||||
| 9837 | #line 508 | |||
| 9838 | SIMD_INTRIN_DEF(loadn_till_s64){ "loadn_till_s64", simd__intrin_loadn_till_s64, 0x0001, ((void *)0) } , | |||
| 9839 | ||||
| 9840 | #line 508 | |||
| 9841 | SIMD_INTRIN_DEF(loadn_tillz_s64){ "loadn_tillz_s64", simd__intrin_loadn_tillz_s64, 0x0001, (( void*)0) } , | |||
| 9842 | ||||
| 9843 | #line 508 | |||
| 9844 | SIMD_INTRIN_DEF(store_till_s64){ "store_till_s64", simd__intrin_store_till_s64, 0x0001, ((void *)0) } , | |||
| 9845 | ||||
| 9846 | #line 508 | |||
| 9847 | SIMD_INTRIN_DEF(storen_s64){ "storen_s64", simd__intrin_storen_s64, 0x0001, ((void*)0) } , | |||
| 9848 | ||||
| 9849 | #line 508 | |||
| 9850 | SIMD_INTRIN_DEF(storen_till_s64){ "storen_till_s64", simd__intrin_storen_till_s64, 0x0001, (( void*)0) } , | |||
| 9851 | ||||
| 9852 | #endif // ncont_sup | |||
| 9853 | ||||
| 9854 | /*************************** | |||
| 9855 | * Misc | |||
| 9856 | ***************************/ | |||
| 9857 | #line 519 | |||
| 9858 | #if 1 | |||
| 9859 | SIMD_INTRIN_DEF(reinterpret_u8_s64){ "reinterpret_u8_s64", simd__intrin_reinterpret_u8_s64, 0x0001 , ((void*)0) } , | |||
| 9860 | #endif // simd_sup2 | |||
| 9861 | ||||
| 9862 | #line 519 | |||
| 9863 | #if 1 | |||
| 9864 | SIMD_INTRIN_DEF(reinterpret_s8_s64){ "reinterpret_s8_s64", simd__intrin_reinterpret_s8_s64, 0x0001 , ((void*)0) } , | |||
| 9865 | #endif // simd_sup2 | |||
| 9866 | ||||
| 9867 | #line 519 | |||
| 9868 | #if 1 | |||
| 9869 | SIMD_INTRIN_DEF(reinterpret_u16_s64){ "reinterpret_u16_s64", simd__intrin_reinterpret_u16_s64, 0x0001 , ((void*)0) } , | |||
| 9870 | #endif // simd_sup2 | |||
| 9871 | ||||
| 9872 | #line 519 | |||
| 9873 | #if 1 | |||
| 9874 | SIMD_INTRIN_DEF(reinterpret_s16_s64){ "reinterpret_s16_s64", simd__intrin_reinterpret_s16_s64, 0x0001 , ((void*)0) } , | |||
| 9875 | #endif // simd_sup2 | |||
| 9876 | ||||
| 9877 | #line 519 | |||
| 9878 | #if 1 | |||
| 9879 | SIMD_INTRIN_DEF(reinterpret_u32_s64){ "reinterpret_u32_s64", simd__intrin_reinterpret_u32_s64, 0x0001 , ((void*)0) } , | |||
| 9880 | #endif // simd_sup2 | |||
| 9881 | ||||
| 9882 | #line 519 | |||
| 9883 | #if 1 | |||
| 9884 | SIMD_INTRIN_DEF(reinterpret_s32_s64){ "reinterpret_s32_s64", simd__intrin_reinterpret_s32_s64, 0x0001 , ((void*)0) } , | |||
| 9885 | #endif // simd_sup2 | |||
| 9886 | ||||
| 9887 | #line 519 | |||
| 9888 | #if 1 | |||
| 9889 | SIMD_INTRIN_DEF(reinterpret_u64_s64){ "reinterpret_u64_s64", simd__intrin_reinterpret_u64_s64, 0x0001 , ((void*)0) } , | |||
| 9890 | #endif // simd_sup2 | |||
| 9891 | ||||
| 9892 | #line 519 | |||
| 9893 | #if 1 | |||
| 9894 | SIMD_INTRIN_DEF(reinterpret_s64_s64){ "reinterpret_s64_s64", simd__intrin_reinterpret_s64_s64, 0x0001 , ((void*)0) } , | |||
| 9895 | #endif // simd_sup2 | |||
| 9896 | ||||
| 9897 | #line 519 | |||
| 9898 | #if 1 | |||
| 9899 | SIMD_INTRIN_DEF(reinterpret_f32_s64){ "reinterpret_f32_s64", simd__intrin_reinterpret_f32_s64, 0x0001 , ((void*)0) } , | |||
| 9900 | #endif // simd_sup2 | |||
| 9901 | ||||
| 9902 | #line 519 | |||
| 9903 | #if NPY_SIMD_F641 | |||
| 9904 | SIMD_INTRIN_DEF(reinterpret_f64_s64){ "reinterpret_f64_s64", simd__intrin_reinterpret_f64_s64, 0x0001 , ((void*)0) } , | |||
| 9905 | #endif // simd_sup2 | |||
| 9906 | ||||
| 9907 | ||||
| 9908 | #line 527 | |||
| 9909 | SIMD_INTRIN_DEF(set_s64){ "set_s64", simd__intrin_set_s64, 0x0001, ((void*)0) } , | |||
| 9910 | ||||
| 9911 | #line 527 | |||
| 9912 | SIMD_INTRIN_DEF(setf_s64){ "setf_s64", simd__intrin_setf_s64, 0x0001, ((void*)0) } , | |||
| 9913 | ||||
| 9914 | #line 527 | |||
| 9915 | SIMD_INTRIN_DEF(setall_s64){ "setall_s64", simd__intrin_setall_s64, 0x0001, ((void*)0) } , | |||
| 9916 | ||||
| 9917 | #line 527 | |||
| 9918 | SIMD_INTRIN_DEF(zero_s64){ "zero_s64", simd__intrin_zero_s64, 0x0001, ((void*)0) } , | |||
| 9919 | ||||
| 9920 | #line 527 | |||
| 9921 | SIMD_INTRIN_DEF(select_s64){ "select_s64", simd__intrin_select_s64, 0x0001, ((void*)0) } , | |||
| 9922 | ||||
| 9923 | ||||
| 9924 | /*************************** | |||
| 9925 | * Reorder | |||
| 9926 | ***************************/ | |||
| 9927 | #line 536 | |||
| 9928 | SIMD_INTRIN_DEF(combinel_s64){ "combinel_s64", simd__intrin_combinel_s64, 0x0001, ((void*) 0) } , | |||
| 9929 | ||||
| 9930 | #line 536 | |||
| 9931 | SIMD_INTRIN_DEF(combineh_s64){ "combineh_s64", simd__intrin_combineh_s64, 0x0001, ((void*) 0) } , | |||
| 9932 | ||||
| 9933 | #line 536 | |||
| 9934 | SIMD_INTRIN_DEF(combine_s64){ "combine_s64", simd__intrin_combine_s64, 0x0001, ((void*)0) } , | |||
| 9935 | ||||
| 9936 | #line 536 | |||
| 9937 | SIMD_INTRIN_DEF(zip_s64){ "zip_s64", simd__intrin_zip_s64, 0x0001, ((void*)0) } , | |||
| 9938 | ||||
| 9939 | ||||
| 9940 | #if 0 | |||
| 9941 | SIMD_INTRIN_DEF(rev64_s64){ "rev64_s64", simd__intrin_rev64_s64, 0x0001, ((void*)0) } , | |||
| 9942 | #endif | |||
| 9943 | ||||
| 9944 | /*************************** | |||
| 9945 | * Operators | |||
| 9946 | ***************************/ | |||
| 9947 | #if 63 > 0 | |||
| 9948 | #line 550 | |||
| 9949 | SIMD_INTRIN_DEF(shl_s64){ "shl_s64", simd__intrin_shl_s64, 0x0001, ((void*)0) } , | |||
| 9950 | ||||
| 9951 | #line 550 | |||
| 9952 | SIMD_INTRIN_DEF(shr_s64){ "shr_s64", simd__intrin_shr_s64, 0x0001, ((void*)0) } , | |||
| 9953 | ||||
| 9954 | #line 550 | |||
| 9955 | SIMD_INTRIN_DEF(shli_s64){ "shli_s64", simd__intrin_shli_s64, 0x0001, ((void*)0) } , | |||
| 9956 | ||||
| 9957 | #line 550 | |||
| 9958 | SIMD_INTRIN_DEF(shri_s64){ "shri_s64", simd__intrin_shri_s64, 0x0001, ((void*)0) } , | |||
| 9959 | ||||
| 9960 | #endif // shl_imm | |||
| 9961 | ||||
| 9962 | #line 557 | |||
| 9963 | SIMD_INTRIN_DEF(and_s64){ "and_s64", simd__intrin_and_s64, 0x0001, ((void*)0) } , | |||
| 9964 | ||||
| 9965 | #line 557 | |||
| 9966 | SIMD_INTRIN_DEF(or_s64){ "or_s64", simd__intrin_or_s64, 0x0001, ((void*)0) } , | |||
| 9967 | ||||
| 9968 | #line 557 | |||
| 9969 | SIMD_INTRIN_DEF(xor_s64){ "xor_s64", simd__intrin_xor_s64, 0x0001, ((void*)0) } , | |||
| 9970 | ||||
| 9971 | #line 557 | |||
| 9972 | SIMD_INTRIN_DEF(not_s64){ "not_s64", simd__intrin_not_s64, 0x0001, ((void*)0) } , | |||
| 9973 | ||||
| 9974 | #line 557 | |||
| 9975 | SIMD_INTRIN_DEF(cmpeq_s64){ "cmpeq_s64", simd__intrin_cmpeq_s64, 0x0001, ((void*)0) } , | |||
| 9976 | ||||
| 9977 | #line 557 | |||
| 9978 | SIMD_INTRIN_DEF(cmpneq_s64){ "cmpneq_s64", simd__intrin_cmpneq_s64, 0x0001, ((void*)0) } , | |||
| 9979 | ||||
| 9980 | #line 557 | |||
| 9981 | SIMD_INTRIN_DEF(cmpgt_s64){ "cmpgt_s64", simd__intrin_cmpgt_s64, 0x0001, ((void*)0) } , | |||
| 9982 | ||||
| 9983 | #line 557 | |||
| 9984 | SIMD_INTRIN_DEF(cmpge_s64){ "cmpge_s64", simd__intrin_cmpge_s64, 0x0001, ((void*)0) } , | |||
| 9985 | ||||
| 9986 | #line 557 | |||
| 9987 | SIMD_INTRIN_DEF(cmplt_s64){ "cmplt_s64", simd__intrin_cmplt_s64, 0x0001, ((void*)0) } , | |||
| 9988 | ||||
| 9989 | #line 557 | |||
| 9990 | SIMD_INTRIN_DEF(cmple_s64){ "cmple_s64", simd__intrin_cmple_s64, 0x0001, ((void*)0) } , | |||
| 9991 | ||||
| 9992 | ||||
| 9993 | /*************************** | |||
| 9994 | * Conversion | |||
| 9995 | ***************************/ | |||
| 9996 | SIMD_INTRIN_DEF(cvt_s64_b64){ "cvt_s64_b64", simd__intrin_cvt_s64_b64, 0x0001, ((void*)0) } , | |||
| 9997 | SIMD_INTRIN_DEF(cvt_b64_s64){ "cvt_b64_s64", simd__intrin_cvt_b64_s64, 0x0001, ((void*)0) } , | |||
| 9998 | #if 0 | |||
| 9999 | SIMD_INTRIN_DEF(expand_s64_s64){ "expand_s64_s64", simd__intrin_expand_s64_s64, 0x0001, ((void *)0) } , | |||
| 10000 | #endif // expand_sup | |||
| 10001 | /*************************** | |||
| 10002 | * Arithmetic | |||
| 10003 | ***************************/ | |||
| 10004 | #line 574 | |||
| 10005 | SIMD_INTRIN_DEF(add_s64){ "add_s64", simd__intrin_add_s64, 0x0001, ((void*)0) } , | |||
| 10006 | ||||
| 10007 | #line 574 | |||
| 10008 | SIMD_INTRIN_DEF(sub_s64){ "sub_s64", simd__intrin_sub_s64, 0x0001, ((void*)0) } , | |||
| 10009 | ||||
| 10010 | ||||
| 10011 | #if 0 | |||
| 10012 | #line 581 | |||
| 10013 | SIMD_INTRIN_DEF(adds_s64){ "adds_s64", simd__intrin_adds_s64, 0x0001, ((void*)0) } , | |||
| 10014 | ||||
| 10015 | #line 581 | |||
| 10016 | SIMD_INTRIN_DEF(subs_s64){ "subs_s64", simd__intrin_subs_s64, 0x0001, ((void*)0) } , | |||
| 10017 | ||||
| 10018 | #endif // sat_sup | |||
| 10019 | ||||
| 10020 | #if 0 | |||
| 10021 | SIMD_INTRIN_DEF(mul_s64){ "mul_s64", simd__intrin_mul_s64, 0x0001, ((void*)0) } , | |||
| 10022 | #endif // mul_sup | |||
| 10023 | ||||
| 10024 | #if 0 | |||
| 10025 | SIMD_INTRIN_DEF(div_s64){ "div_s64", simd__intrin_div_s64, 0x0001, ((void*)0) } , | |||
| 10026 | #endif // div_sup | |||
| 10027 | ||||
| 10028 | #if 1 | |||
| 10029 | SIMD_INTRIN_DEF(divisor_s64){ "divisor_s64", simd__intrin_divisor_s64, 0x0001, ((void*)0) } , | |||
| 10030 | SIMD_INTRIN_DEF(divc_s64){ "divc_s64", simd__intrin_divc_s64, 0x0001, ((void*)0) } , | |||
| 10031 | #endif // intdiv_sup | |||
| 10032 | ||||
| 10033 | #if 0 | |||
| 10034 | #line 602 | |||
| 10035 | SIMD_INTRIN_DEF(muladd_s64){ "muladd_s64", simd__intrin_muladd_s64, 0x0001, ((void*)0) } , | |||
| 10036 | ||||
| 10037 | #line 602 | |||
| 10038 | SIMD_INTRIN_DEF(mulsub_s64){ "mulsub_s64", simd__intrin_mulsub_s64, 0x0001, ((void*)0) } , | |||
| 10039 | ||||
| 10040 | #line 602 | |||
| 10041 | SIMD_INTRIN_DEF(nmuladd_s64){ "nmuladd_s64", simd__intrin_nmuladd_s64, 0x0001, ((void*)0) } , | |||
| 10042 | ||||
| 10043 | #line 602 | |||
| 10044 | SIMD_INTRIN_DEF(nmulsub_s64){ "nmulsub_s64", simd__intrin_nmulsub_s64, 0x0001, ((void*)0) } , | |||
| 10045 | ||||
| 10046 | #endif // fused_sup | |||
| 10047 | ||||
| 10048 | #if 0 | |||
| 10049 | SIMD_INTRIN_DEF(sum_s64){ "sum_s64", simd__intrin_sum_s64, 0x0001, ((void*)0) } , | |||
| 10050 | #endif // sum_sup | |||
| 10051 | ||||
| 10052 | #if 0 | |||
| 10053 | SIMD_INTRIN_DEF(sumup_s64){ "sumup_s64", simd__intrin_sumup_s64, 0x0001, ((void*)0) } , | |||
| 10054 | #endif // sumup_sup | |||
| 10055 | /*************************** | |||
| 10056 | * Math | |||
| 10057 | ***************************/ | |||
| 10058 | #if 0 | |||
| 10059 | #line 620 | |||
| 10060 | SIMD_INTRIN_DEF(sqrt_s64){ "sqrt_s64", simd__intrin_sqrt_s64, 0x0001, ((void*)0) } , | |||
| 10061 | ||||
| 10062 | #line 620 | |||
| 10063 | SIMD_INTRIN_DEF(recip_s64){ "recip_s64", simd__intrin_recip_s64, 0x0001, ((void*)0) } , | |||
| 10064 | ||||
| 10065 | #line 620 | |||
| 10066 | SIMD_INTRIN_DEF(abs_s64){ "abs_s64", simd__intrin_abs_s64, 0x0001, ((void*)0) } , | |||
| 10067 | ||||
| 10068 | #line 620 | |||
| 10069 | SIMD_INTRIN_DEF(square_s64){ "square_s64", simd__intrin_square_s64, 0x0001, ((void*)0) } , | |||
| 10070 | ||||
| 10071 | #endif | |||
| 10072 | ||||
| 10073 | #line 627 | |||
| 10074 | SIMD_INTRIN_DEF(max_s64){ "max_s64", simd__intrin_max_s64, 0x0001, ((void*)0) } , | |||
| 10075 | ||||
| 10076 | #line 627 | |||
| 10077 | SIMD_INTRIN_DEF(min_s64){ "min_s64", simd__intrin_min_s64, 0x0001, ((void*)0) } , | |||
| 10078 | ||||
| 10079 | ||||
| 10080 | #if 0 | |||
| 10081 | #line 634 | |||
| 10082 | SIMD_INTRIN_DEF(maxp_s64){ "maxp_s64", simd__intrin_maxp_s64, 0x0001, ((void*)0) } , | |||
| 10083 | ||||
| 10084 | #line 634 | |||
| 10085 | SIMD_INTRIN_DEF(minp_s64){ "minp_s64", simd__intrin_minp_s64, 0x0001, ((void*)0) } , | |||
| 10086 | ||||
| 10087 | #endif | |||
| 10088 | ||||
| 10089 | /*************************** | |||
| 10090 | * Mask operations | |||
| 10091 | ***************************/ | |||
| 10092 | #line 644 | |||
| 10093 | SIMD_INTRIN_DEF(ifadd_s64){ "ifadd_s64", simd__intrin_ifadd_s64, 0x0001, ((void*)0) } , | |||
| 10094 | ||||
| 10095 | #line 644 | |||
| 10096 | SIMD_INTRIN_DEF(ifsub_s64){ "ifsub_s64", simd__intrin_ifsub_s64, 0x0001, ((void*)0) } , | |||
| 10097 | ||||
| 10098 | ||||
| 10099 | #endif // simd_sup | |||
| 10100 | ||||
| 10101 | #line 489 | |||
| 10102 | #if 1 | |||
| 10103 | ||||
| 10104 | /*************************** | |||
| 10105 | * Memory | |||
| 10106 | ***************************/ | |||
| 10107 | #line 497 | |||
| 10108 | SIMD_INTRIN_DEF(load_f32){ "load_f32", simd__intrin_load_f32, 0x0001, ((void*)0) } , | |||
| 10109 | ||||
| 10110 | #line 497 | |||
| 10111 | SIMD_INTRIN_DEF(loada_f32){ "loada_f32", simd__intrin_loada_f32, 0x0001, ((void*)0) } , | |||
| 10112 | ||||
| 10113 | #line 497 | |||
| 10114 | SIMD_INTRIN_DEF(loads_f32){ "loads_f32", simd__intrin_loads_f32, 0x0001, ((void*)0) } , | |||
| 10115 | ||||
| 10116 | #line 497 | |||
| 10117 | SIMD_INTRIN_DEF(loadl_f32){ "loadl_f32", simd__intrin_loadl_f32, 0x0001, ((void*)0) } , | |||
| 10118 | ||||
| 10119 | #line 497 | |||
| 10120 | SIMD_INTRIN_DEF(store_f32){ "store_f32", simd__intrin_store_f32, 0x0001, ((void*)0) } , | |||
| 10121 | ||||
| 10122 | #line 497 | |||
| 10123 | SIMD_INTRIN_DEF(storea_f32){ "storea_f32", simd__intrin_storea_f32, 0x0001, ((void*)0) } , | |||
| 10124 | ||||
| 10125 | #line 497 | |||
| 10126 | SIMD_INTRIN_DEF(stores_f32){ "stores_f32", simd__intrin_stores_f32, 0x0001, ((void*)0) } , | |||
| 10127 | ||||
| 10128 | #line 497 | |||
| 10129 | SIMD_INTRIN_DEF(storel_f32){ "storel_f32", simd__intrin_storel_f32, 0x0001, ((void*)0) } , | |||
| 10130 | ||||
| 10131 | #line 497 | |||
| 10132 | SIMD_INTRIN_DEF(storeh_f32){ "storeh_f32", simd__intrin_storeh_f32, 0x0001, ((void*)0) } , | |||
| 10133 | ||||
| 10134 | ||||
| 10135 | /**************************************** | |||
| 10136 | * Non-contiguous/Partial Memory access | |||
| 10137 | ****************************************/ | |||
| 10138 | #if 1 | |||
| 10139 | #line 508 | |||
| 10140 | SIMD_INTRIN_DEF(load_till_f32){ "load_till_f32", simd__intrin_load_till_f32, 0x0001, ((void *)0) } , | |||
| 10141 | ||||
| 10142 | #line 508 | |||
| 10143 | SIMD_INTRIN_DEF(load_tillz_f32){ "load_tillz_f32", simd__intrin_load_tillz_f32, 0x0001, ((void *)0) } , | |||
| 10144 | ||||
| 10145 | #line 508 | |||
| 10146 | SIMD_INTRIN_DEF(loadn_f32){ "loadn_f32", simd__intrin_loadn_f32, 0x0001, ((void*)0) } , | |||
| 10147 | ||||
| 10148 | #line 508 | |||
| 10149 | SIMD_INTRIN_DEF(loadn_till_f32){ "loadn_till_f32", simd__intrin_loadn_till_f32, 0x0001, ((void *)0) } , | |||
| 10150 | ||||
| 10151 | #line 508 | |||
| 10152 | SIMD_INTRIN_DEF(loadn_tillz_f32){ "loadn_tillz_f32", simd__intrin_loadn_tillz_f32, 0x0001, (( void*)0) } , | |||
| 10153 | ||||
| 10154 | #line 508 | |||
| 10155 | SIMD_INTRIN_DEF(store_till_f32){ "store_till_f32", simd__intrin_store_till_f32, 0x0001, ((void *)0) } , | |||
| 10156 | ||||
| 10157 | #line 508 | |||
| 10158 | SIMD_INTRIN_DEF(storen_f32){ "storen_f32", simd__intrin_storen_f32, 0x0001, ((void*)0) } , | |||
| 10159 | ||||
| 10160 | #line 508 | |||
| 10161 | SIMD_INTRIN_DEF(storen_till_f32){ "storen_till_f32", simd__intrin_storen_till_f32, 0x0001, (( void*)0) } , | |||
| 10162 | ||||
| 10163 | #endif // ncont_sup | |||
| 10164 | ||||
| 10165 | /*************************** | |||
| 10166 | * Misc | |||
| 10167 | ***************************/ | |||
| 10168 | #line 519 | |||
| 10169 | #if 1 | |||
| 10170 | SIMD_INTRIN_DEF(reinterpret_u8_f32){ "reinterpret_u8_f32", simd__intrin_reinterpret_u8_f32, 0x0001 , ((void*)0) } , | |||
| 10171 | #endif // simd_sup2 | |||
| 10172 | ||||
| 10173 | #line 519 | |||
| 10174 | #if 1 | |||
| 10175 | SIMD_INTRIN_DEF(reinterpret_s8_f32){ "reinterpret_s8_f32", simd__intrin_reinterpret_s8_f32, 0x0001 , ((void*)0) } , | |||
| 10176 | #endif // simd_sup2 | |||
| 10177 | ||||
| 10178 | #line 519 | |||
| 10179 | #if 1 | |||
| 10180 | SIMD_INTRIN_DEF(reinterpret_u16_f32){ "reinterpret_u16_f32", simd__intrin_reinterpret_u16_f32, 0x0001 , ((void*)0) } , | |||
| 10181 | #endif // simd_sup2 | |||
| 10182 | ||||
| 10183 | #line 519 | |||
| 10184 | #if 1 | |||
| 10185 | SIMD_INTRIN_DEF(reinterpret_s16_f32){ "reinterpret_s16_f32", simd__intrin_reinterpret_s16_f32, 0x0001 , ((void*)0) } , | |||
| 10186 | #endif // simd_sup2 | |||
| 10187 | ||||
| 10188 | #line 519 | |||
| 10189 | #if 1 | |||
| 10190 | SIMD_INTRIN_DEF(reinterpret_u32_f32){ "reinterpret_u32_f32", simd__intrin_reinterpret_u32_f32, 0x0001 , ((void*)0) } , | |||
| 10191 | #endif // simd_sup2 | |||
| 10192 | ||||
| 10193 | #line 519 | |||
| 10194 | #if 1 | |||
| 10195 | SIMD_INTRIN_DEF(reinterpret_s32_f32){ "reinterpret_s32_f32", simd__intrin_reinterpret_s32_f32, 0x0001 , ((void*)0) } , | |||
| 10196 | #endif // simd_sup2 | |||
| 10197 | ||||
| 10198 | #line 519 | |||
| 10199 | #if 1 | |||
| 10200 | SIMD_INTRIN_DEF(reinterpret_u64_f32){ "reinterpret_u64_f32", simd__intrin_reinterpret_u64_f32, 0x0001 , ((void*)0) } , | |||
| 10201 | #endif // simd_sup2 | |||
| 10202 | ||||
| 10203 | #line 519 | |||
| 10204 | #if 1 | |||
| 10205 | SIMD_INTRIN_DEF(reinterpret_s64_f32){ "reinterpret_s64_f32", simd__intrin_reinterpret_s64_f32, 0x0001 , ((void*)0) } , | |||
| 10206 | #endif // simd_sup2 | |||
| 10207 | ||||
| 10208 | #line 519 | |||
| 10209 | #if 1 | |||
| 10210 | SIMD_INTRIN_DEF(reinterpret_f32_f32){ "reinterpret_f32_f32", simd__intrin_reinterpret_f32_f32, 0x0001 , ((void*)0) } , | |||
| 10211 | #endif // simd_sup2 | |||
| 10212 | ||||
| 10213 | #line 519 | |||
| 10214 | #if NPY_SIMD_F641 | |||
| 10215 | SIMD_INTRIN_DEF(reinterpret_f64_f32){ "reinterpret_f64_f32", simd__intrin_reinterpret_f64_f32, 0x0001 , ((void*)0) } , | |||
| 10216 | #endif // simd_sup2 | |||
| 10217 | ||||
| 10218 | ||||
| 10219 | #line 527 | |||
| 10220 | SIMD_INTRIN_DEF(set_f32){ "set_f32", simd__intrin_set_f32, 0x0001, ((void*)0) } , | |||
| 10221 | ||||
| 10222 | #line 527 | |||
| 10223 | SIMD_INTRIN_DEF(setf_f32){ "setf_f32", simd__intrin_setf_f32, 0x0001, ((void*)0) } , | |||
| 10224 | ||||
| 10225 | #line 527 | |||
| 10226 | SIMD_INTRIN_DEF(setall_f32){ "setall_f32", simd__intrin_setall_f32, 0x0001, ((void*)0) } , | |||
| 10227 | ||||
| 10228 | #line 527 | |||
| 10229 | SIMD_INTRIN_DEF(zero_f32){ "zero_f32", simd__intrin_zero_f32, 0x0001, ((void*)0) } , | |||
| 10230 | ||||
| 10231 | #line 527 | |||
| 10232 | SIMD_INTRIN_DEF(select_f32){ "select_f32", simd__intrin_select_f32, 0x0001, ((void*)0) } , | |||
| 10233 | ||||
| 10234 | ||||
| 10235 | /*************************** | |||
| 10236 | * Reorder | |||
| 10237 | ***************************/ | |||
| 10238 | #line 536 | |||
| 10239 | SIMD_INTRIN_DEF(combinel_f32){ "combinel_f32", simd__intrin_combinel_f32, 0x0001, ((void*) 0) } , | |||
| 10240 | ||||
| 10241 | #line 536 | |||
| 10242 | SIMD_INTRIN_DEF(combineh_f32){ "combineh_f32", simd__intrin_combineh_f32, 0x0001, ((void*) 0) } , | |||
| 10243 | ||||
| 10244 | #line 536 | |||
| 10245 | SIMD_INTRIN_DEF(combine_f32){ "combine_f32", simd__intrin_combine_f32, 0x0001, ((void*)0) } , | |||
| 10246 | ||||
| 10247 | #line 536 | |||
| 10248 | SIMD_INTRIN_DEF(zip_f32){ "zip_f32", simd__intrin_zip_f32, 0x0001, ((void*)0) } , | |||
| 10249 | ||||
| 10250 | ||||
| 10251 | #if 1 | |||
| 10252 | SIMD_INTRIN_DEF(rev64_f32){ "rev64_f32", simd__intrin_rev64_f32, 0x0001, ((void*)0) } , | |||
| 10253 | #endif | |||
| 10254 | ||||
| 10255 | /*************************** | |||
| 10256 | * Operators | |||
| 10257 | ***************************/ | |||
| 10258 | #if 0 > 0 | |||
| 10259 | #line 550 | |||
| 10260 | SIMD_INTRIN_DEF(shl_f32){ "shl_f32", simd__intrin_shl_f32, 0x0001, ((void*)0) } , | |||
| 10261 | ||||
| 10262 | #line 550 | |||
| 10263 | SIMD_INTRIN_DEF(shr_f32){ "shr_f32", simd__intrin_shr_f32, 0x0001, ((void*)0) } , | |||
| 10264 | ||||
| 10265 | #line 550 | |||
| 10266 | SIMD_INTRIN_DEF(shli_f32){ "shli_f32", simd__intrin_shli_f32, 0x0001, ((void*)0) } , | |||
| 10267 | ||||
| 10268 | #line 550 | |||
| 10269 | SIMD_INTRIN_DEF(shri_f32){ "shri_f32", simd__intrin_shri_f32, 0x0001, ((void*)0) } , | |||
| 10270 | ||||
| 10271 | #endif // shl_imm | |||
| 10272 | ||||
| 10273 | #line 557 | |||
| 10274 | SIMD_INTRIN_DEF(and_f32){ "and_f32", simd__intrin_and_f32, 0x0001, ((void*)0) } , | |||
| 10275 | ||||
| 10276 | #line 557 | |||
| 10277 | SIMD_INTRIN_DEF(or_f32){ "or_f32", simd__intrin_or_f32, 0x0001, ((void*)0) } , | |||
| 10278 | ||||
| 10279 | #line 557 | |||
| 10280 | SIMD_INTRIN_DEF(xor_f32){ "xor_f32", simd__intrin_xor_f32, 0x0001, ((void*)0) } , | |||
| 10281 | ||||
| 10282 | #line 557 | |||
| 10283 | SIMD_INTRIN_DEF(not_f32){ "not_f32", simd__intrin_not_f32, 0x0001, ((void*)0) } , | |||
| 10284 | ||||
| 10285 | #line 557 | |||
| 10286 | SIMD_INTRIN_DEF(cmpeq_f32){ "cmpeq_f32", simd__intrin_cmpeq_f32, 0x0001, ((void*)0) } , | |||
| 10287 | ||||
| 10288 | #line 557 | |||
| 10289 | SIMD_INTRIN_DEF(cmpneq_f32){ "cmpneq_f32", simd__intrin_cmpneq_f32, 0x0001, ((void*)0) } , | |||
| 10290 | ||||
| 10291 | #line 557 | |||
| 10292 | SIMD_INTRIN_DEF(cmpgt_f32){ "cmpgt_f32", simd__intrin_cmpgt_f32, 0x0001, ((void*)0) } , | |||
| 10293 | ||||
| 10294 | #line 557 | |||
| 10295 | SIMD_INTRIN_DEF(cmpge_f32){ "cmpge_f32", simd__intrin_cmpge_f32, 0x0001, ((void*)0) } , | |||
| 10296 | ||||
| 10297 | #line 557 | |||
| 10298 | SIMD_INTRIN_DEF(cmplt_f32){ "cmplt_f32", simd__intrin_cmplt_f32, 0x0001, ((void*)0) } , | |||
| 10299 | ||||
| 10300 | #line 557 | |||
| 10301 | SIMD_INTRIN_DEF(cmple_f32){ "cmple_f32", simd__intrin_cmple_f32, 0x0001, ((void*)0) } , | |||
| 10302 | ||||
| 10303 | ||||
| 10304 | /*************************** | |||
| 10305 | * Conversion | |||
| 10306 | ***************************/ | |||
| 10307 | SIMD_INTRIN_DEF(cvt_f32_b32){ "cvt_f32_b32", simd__intrin_cvt_f32_b32, 0x0001, ((void*)0) } , | |||
| 10308 | SIMD_INTRIN_DEF(cvt_b32_f32){ "cvt_b32_f32", simd__intrin_cvt_b32_f32, 0x0001, ((void*)0) } , | |||
| 10309 | #if 0 | |||
| 10310 | SIMD_INTRIN_DEF(expand_f32_f32){ "expand_f32_f32", simd__intrin_expand_f32_f32, 0x0001, ((void *)0) } , | |||
| 10311 | #endif // expand_sup | |||
| 10312 | /*************************** | |||
| 10313 | * Arithmetic | |||
| 10314 | ***************************/ | |||
| 10315 | #line 574 | |||
| 10316 | SIMD_INTRIN_DEF(add_f32){ "add_f32", simd__intrin_add_f32, 0x0001, ((void*)0) } , | |||
| 10317 | ||||
| 10318 | #line 574 | |||
| 10319 | SIMD_INTRIN_DEF(sub_f32){ "sub_f32", simd__intrin_sub_f32, 0x0001, ((void*)0) } , | |||
| 10320 | ||||
| 10321 | ||||
| 10322 | #if 0 | |||
| 10323 | #line 581 | |||
| 10324 | SIMD_INTRIN_DEF(adds_f32){ "adds_f32", simd__intrin_adds_f32, 0x0001, ((void*)0) } , | |||
| 10325 | ||||
| 10326 | #line 581 | |||
| 10327 | SIMD_INTRIN_DEF(subs_f32){ "subs_f32", simd__intrin_subs_f32, 0x0001, ((void*)0) } , | |||
| 10328 | ||||
| 10329 | #endif // sat_sup | |||
| 10330 | ||||
| 10331 | #if 1 | |||
| 10332 | SIMD_INTRIN_DEF(mul_f32){ "mul_f32", simd__intrin_mul_f32, 0x0001, ((void*)0) } , | |||
| 10333 | #endif // mul_sup | |||
| 10334 | ||||
| 10335 | #if 1 | |||
| 10336 | SIMD_INTRIN_DEF(div_f32){ "div_f32", simd__intrin_div_f32, 0x0001, ((void*)0) } , | |||
| 10337 | #endif // div_sup | |||
| 10338 | ||||
| 10339 | #if 0 | |||
| 10340 | SIMD_INTRIN_DEF(divisor_f32){ "divisor_f32", simd__intrin_divisor_f32, 0x0001, ((void*)0) } , | |||
| 10341 | SIMD_INTRIN_DEF(divc_f32){ "divc_f32", simd__intrin_divc_f32, 0x0001, ((void*)0) } , | |||
| 10342 | #endif // intdiv_sup | |||
| 10343 | ||||
| 10344 | #if 1 | |||
| 10345 | #line 602 | |||
| 10346 | SIMD_INTRIN_DEF(muladd_f32){ "muladd_f32", simd__intrin_muladd_f32, 0x0001, ((void*)0) } , | |||
| 10347 | ||||
| 10348 | #line 602 | |||
| 10349 | SIMD_INTRIN_DEF(mulsub_f32){ "mulsub_f32", simd__intrin_mulsub_f32, 0x0001, ((void*)0) } , | |||
| 10350 | ||||
| 10351 | #line 602 | |||
| 10352 | SIMD_INTRIN_DEF(nmuladd_f32){ "nmuladd_f32", simd__intrin_nmuladd_f32, 0x0001, ((void*)0) } , | |||
| 10353 | ||||
| 10354 | #line 602 | |||
| 10355 | SIMD_INTRIN_DEF(nmulsub_f32){ "nmulsub_f32", simd__intrin_nmulsub_f32, 0x0001, ((void*)0) } , | |||
| 10356 | ||||
| 10357 | #endif // fused_sup | |||
| 10358 | ||||
| 10359 | #if 1 | |||
| 10360 | SIMD_INTRIN_DEF(sum_f32){ "sum_f32", simd__intrin_sum_f32, 0x0001, ((void*)0) } , | |||
| 10361 | #endif // sum_sup | |||
| 10362 | ||||
| 10363 | #if 0 | |||
| 10364 | SIMD_INTRIN_DEF(sumup_f32){ "sumup_f32", simd__intrin_sumup_f32, 0x0001, ((void*)0) } , | |||
| 10365 | #endif // sumup_sup | |||
| 10366 | /*************************** | |||
| 10367 | * Math | |||
| 10368 | ***************************/ | |||
| 10369 | #if 1 | |||
| 10370 | #line 620 | |||
| 10371 | SIMD_INTRIN_DEF(sqrt_f32){ "sqrt_f32", simd__intrin_sqrt_f32, 0x0001, ((void*)0) } , | |||
| 10372 | ||||
| 10373 | #line 620 | |||
| 10374 | SIMD_INTRIN_DEF(recip_f32){ "recip_f32", simd__intrin_recip_f32, 0x0001, ((void*)0) } , | |||
| 10375 | ||||
| 10376 | #line 620 | |||
| 10377 | SIMD_INTRIN_DEF(abs_f32){ "abs_f32", simd__intrin_abs_f32, 0x0001, ((void*)0) } , | |||
| 10378 | ||||
| 10379 | #line 620 | |||
| 10380 | SIMD_INTRIN_DEF(square_f32){ "square_f32", simd__intrin_square_f32, 0x0001, ((void*)0) } , | |||
| 10381 | ||||
| 10382 | #endif | |||
| 10383 | ||||
| 10384 | #line 627 | |||
| 10385 | SIMD_INTRIN_DEF(max_f32){ "max_f32", simd__intrin_max_f32, 0x0001, ((void*)0) } , | |||
| 10386 | ||||
| 10387 | #line 627 | |||
| 10388 | SIMD_INTRIN_DEF(min_f32){ "min_f32", simd__intrin_min_f32, 0x0001, ((void*)0) } , | |||
| 10389 | ||||
| 10390 | ||||
| 10391 | #if 1 | |||
| 10392 | #line 634 | |||
| 10393 | SIMD_INTRIN_DEF(maxp_f32){ "maxp_f32", simd__intrin_maxp_f32, 0x0001, ((void*)0) } , | |||
| 10394 | ||||
| 10395 | #line 634 | |||
| 10396 | SIMD_INTRIN_DEF(minp_f32){ "minp_f32", simd__intrin_minp_f32, 0x0001, ((void*)0) } , | |||
| 10397 | ||||
| 10398 | #endif | |||
| 10399 | ||||
| 10400 | /*************************** | |||
| 10401 | * Mask operations | |||
| 10402 | ***************************/ | |||
| 10403 | #line 644 | |||
| 10404 | SIMD_INTRIN_DEF(ifadd_f32){ "ifadd_f32", simd__intrin_ifadd_f32, 0x0001, ((void*)0) } , | |||
| 10405 | ||||
| 10406 | #line 644 | |||
| 10407 | SIMD_INTRIN_DEF(ifsub_f32){ "ifsub_f32", simd__intrin_ifsub_f32, 0x0001, ((void*)0) } , | |||
| 10408 | ||||
| 10409 | ||||
| 10410 | #endif // simd_sup | |||
| 10411 | ||||
| 10412 | #line 489 | |||
| 10413 | #if NPY_SIMD_F641 | |||
| 10414 | ||||
| 10415 | /*************************** | |||
| 10416 | * Memory | |||
| 10417 | ***************************/ | |||
| 10418 | #line 497 | |||
| 10419 | SIMD_INTRIN_DEF(load_f64){ "load_f64", simd__intrin_load_f64, 0x0001, ((void*)0) } , | |||
| 10420 | ||||
| 10421 | #line 497 | |||
| 10422 | SIMD_INTRIN_DEF(loada_f64){ "loada_f64", simd__intrin_loada_f64, 0x0001, ((void*)0) } , | |||
| 10423 | ||||
| 10424 | #line 497 | |||
| 10425 | SIMD_INTRIN_DEF(loads_f64){ "loads_f64", simd__intrin_loads_f64, 0x0001, ((void*)0) } , | |||
| 10426 | ||||
| 10427 | #line 497 | |||
| 10428 | SIMD_INTRIN_DEF(loadl_f64){ "loadl_f64", simd__intrin_loadl_f64, 0x0001, ((void*)0) } , | |||
| 10429 | ||||
| 10430 | #line 497 | |||
| 10431 | SIMD_INTRIN_DEF(store_f64){ "store_f64", simd__intrin_store_f64, 0x0001, ((void*)0) } , | |||
| 10432 | ||||
| 10433 | #line 497 | |||
| 10434 | SIMD_INTRIN_DEF(storea_f64){ "storea_f64", simd__intrin_storea_f64, 0x0001, ((void*)0) } , | |||
| 10435 | ||||
| 10436 | #line 497 | |||
| 10437 | SIMD_INTRIN_DEF(stores_f64){ "stores_f64", simd__intrin_stores_f64, 0x0001, ((void*)0) } , | |||
| 10438 | ||||
| 10439 | #line 497 | |||
| 10440 | SIMD_INTRIN_DEF(storel_f64){ "storel_f64", simd__intrin_storel_f64, 0x0001, ((void*)0) } , | |||
| 10441 | ||||
| 10442 | #line 497 | |||
| 10443 | SIMD_INTRIN_DEF(storeh_f64){ "storeh_f64", simd__intrin_storeh_f64, 0x0001, ((void*)0) } , | |||
| 10444 | ||||
| 10445 | ||||
| 10446 | /**************************************** | |||
| 10447 | * Non-contiguous/Partial Memory access | |||
| 10448 | ****************************************/ | |||
| 10449 | #if 1 | |||
| 10450 | #line 508 | |||
| 10451 | SIMD_INTRIN_DEF(load_till_f64){ "load_till_f64", simd__intrin_load_till_f64, 0x0001, ((void *)0) } , | |||
| 10452 | ||||
| 10453 | #line 508 | |||
| 10454 | SIMD_INTRIN_DEF(load_tillz_f64){ "load_tillz_f64", simd__intrin_load_tillz_f64, 0x0001, ((void *)0) } , | |||
| 10455 | ||||
| 10456 | #line 508 | |||
| 10457 | SIMD_INTRIN_DEF(loadn_f64){ "loadn_f64", simd__intrin_loadn_f64, 0x0001, ((void*)0) } , | |||
| 10458 | ||||
| 10459 | #line 508 | |||
| 10460 | SIMD_INTRIN_DEF(loadn_till_f64){ "loadn_till_f64", simd__intrin_loadn_till_f64, 0x0001, ((void *)0) } , | |||
| 10461 | ||||
| 10462 | #line 508 | |||
| 10463 | SIMD_INTRIN_DEF(loadn_tillz_f64){ "loadn_tillz_f64", simd__intrin_loadn_tillz_f64, 0x0001, (( void*)0) } , | |||
| 10464 | ||||
| 10465 | #line 508 | |||
| 10466 | SIMD_INTRIN_DEF(store_till_f64){ "store_till_f64", simd__intrin_store_till_f64, 0x0001, ((void *)0) } , | |||
| 10467 | ||||
| 10468 | #line 508 | |||
| 10469 | SIMD_INTRIN_DEF(storen_f64){ "storen_f64", simd__intrin_storen_f64, 0x0001, ((void*)0) } , | |||
| 10470 | ||||
| 10471 | #line 508 | |||
| 10472 | SIMD_INTRIN_DEF(storen_till_f64){ "storen_till_f64", simd__intrin_storen_till_f64, 0x0001, (( void*)0) } , | |||
| 10473 | ||||
| 10474 | #endif // ncont_sup | |||
| 10475 | ||||
| 10476 | /*************************** | |||
| 10477 | * Misc | |||
| 10478 | ***************************/ | |||
| 10479 | #line 519 | |||
| 10480 | #if 1 | |||
| 10481 | SIMD_INTRIN_DEF(reinterpret_u8_f64){ "reinterpret_u8_f64", simd__intrin_reinterpret_u8_f64, 0x0001 , ((void*)0) } , | |||
| 10482 | #endif // simd_sup2 | |||
| 10483 | ||||
| 10484 | #line 519 | |||
| 10485 | #if 1 | |||
| 10486 | SIMD_INTRIN_DEF(reinterpret_s8_f64){ "reinterpret_s8_f64", simd__intrin_reinterpret_s8_f64, 0x0001 , ((void*)0) } , | |||
| 10487 | #endif // simd_sup2 | |||
| 10488 | ||||
| 10489 | #line 519 | |||
| 10490 | #if 1 | |||
| 10491 | SIMD_INTRIN_DEF(reinterpret_u16_f64){ "reinterpret_u16_f64", simd__intrin_reinterpret_u16_f64, 0x0001 , ((void*)0) } , | |||
| 10492 | #endif // simd_sup2 | |||
| 10493 | ||||
| 10494 | #line 519 | |||
| 10495 | #if 1 | |||
| 10496 | SIMD_INTRIN_DEF(reinterpret_s16_f64){ "reinterpret_s16_f64", simd__intrin_reinterpret_s16_f64, 0x0001 , ((void*)0) } , | |||
| 10497 | #endif // simd_sup2 | |||
| 10498 | ||||
| 10499 | #line 519 | |||
| 10500 | #if 1 | |||
| 10501 | SIMD_INTRIN_DEF(reinterpret_u32_f64){ "reinterpret_u32_f64", simd__intrin_reinterpret_u32_f64, 0x0001 , ((void*)0) } , | |||
| 10502 | #endif // simd_sup2 | |||
| 10503 | ||||
| 10504 | #line 519 | |||
| 10505 | #if 1 | |||
| 10506 | SIMD_INTRIN_DEF(reinterpret_s32_f64){ "reinterpret_s32_f64", simd__intrin_reinterpret_s32_f64, 0x0001 , ((void*)0) } , | |||
| 10507 | #endif // simd_sup2 | |||
| 10508 | ||||
| 10509 | #line 519 | |||
| 10510 | #if 1 | |||
| 10511 | SIMD_INTRIN_DEF(reinterpret_u64_f64){ "reinterpret_u64_f64", simd__intrin_reinterpret_u64_f64, 0x0001 , ((void*)0) } , | |||
| 10512 | #endif // simd_sup2 | |||
| 10513 | ||||
| 10514 | #line 519 | |||
| 10515 | #if 1 | |||
| 10516 | SIMD_INTRIN_DEF(reinterpret_s64_f64){ "reinterpret_s64_f64", simd__intrin_reinterpret_s64_f64, 0x0001 , ((void*)0) } , | |||
| 10517 | #endif // simd_sup2 | |||
| 10518 | ||||
| 10519 | #line 519 | |||
| 10520 | #if 1 | |||
| 10521 | SIMD_INTRIN_DEF(reinterpret_f32_f64){ "reinterpret_f32_f64", simd__intrin_reinterpret_f32_f64, 0x0001 , ((void*)0) } , | |||
| 10522 | #endif // simd_sup2 | |||
| 10523 | ||||
| 10524 | #line 519 | |||
| 10525 | #if NPY_SIMD_F641 | |||
| 10526 | SIMD_INTRIN_DEF(reinterpret_f64_f64){ "reinterpret_f64_f64", simd__intrin_reinterpret_f64_f64, 0x0001 , ((void*)0) } , | |||
| 10527 | #endif // simd_sup2 | |||
| 10528 | ||||
| 10529 | ||||
| 10530 | #line 527 | |||
| 10531 | SIMD_INTRIN_DEF(set_f64){ "set_f64", simd__intrin_set_f64, 0x0001, ((void*)0) } , | |||
| 10532 | ||||
| 10533 | #line 527 | |||
| 10534 | SIMD_INTRIN_DEF(setf_f64){ "setf_f64", simd__intrin_setf_f64, 0x0001, ((void*)0) } , | |||
| 10535 | ||||
| 10536 | #line 527 | |||
| 10537 | SIMD_INTRIN_DEF(setall_f64){ "setall_f64", simd__intrin_setall_f64, 0x0001, ((void*)0) } , | |||
| 10538 | ||||
| 10539 | #line 527 | |||
| 10540 | SIMD_INTRIN_DEF(zero_f64){ "zero_f64", simd__intrin_zero_f64, 0x0001, ((void*)0) } , | |||
| 10541 | ||||
| 10542 | #line 527 | |||
| 10543 | SIMD_INTRIN_DEF(select_f64){ "select_f64", simd__intrin_select_f64, 0x0001, ((void*)0) } , | |||
| 10544 | ||||
| 10545 | ||||
| 10546 | /*************************** | |||
| 10547 | * Reorder | |||
| 10548 | ***************************/ | |||
| 10549 | #line 536 | |||
| 10550 | SIMD_INTRIN_DEF(combinel_f64){ "combinel_f64", simd__intrin_combinel_f64, 0x0001, ((void*) 0) } , | |||
| 10551 | ||||
| 10552 | #line 536 | |||
| 10553 | SIMD_INTRIN_DEF(combineh_f64){ "combineh_f64", simd__intrin_combineh_f64, 0x0001, ((void*) 0) } , | |||
| 10554 | ||||
| 10555 | #line 536 | |||
| 10556 | SIMD_INTRIN_DEF(combine_f64){ "combine_f64", simd__intrin_combine_f64, 0x0001, ((void*)0) } , | |||
| 10557 | ||||
| 10558 | #line 536 | |||
| 10559 | SIMD_INTRIN_DEF(zip_f64){ "zip_f64", simd__intrin_zip_f64, 0x0001, ((void*)0) } , | |||
| 10560 | ||||
| 10561 | ||||
| 10562 | #if 0 | |||
| 10563 | SIMD_INTRIN_DEF(rev64_f64){ "rev64_f64", simd__intrin_rev64_f64, 0x0001, ((void*)0) } , | |||
| 10564 | #endif | |||
| 10565 | ||||
| 10566 | /*************************** | |||
| 10567 | * Operators | |||
| 10568 | ***************************/ | |||
| 10569 | #if 0 > 0 | |||
| 10570 | #line 550 | |||
| 10571 | SIMD_INTRIN_DEF(shl_f64){ "shl_f64", simd__intrin_shl_f64, 0x0001, ((void*)0) } , | |||
| 10572 | ||||
| 10573 | #line 550 | |||
| 10574 | SIMD_INTRIN_DEF(shr_f64){ "shr_f64", simd__intrin_shr_f64, 0x0001, ((void*)0) } , | |||
| 10575 | ||||
| 10576 | #line 550 | |||
| 10577 | SIMD_INTRIN_DEF(shli_f64){ "shli_f64", simd__intrin_shli_f64, 0x0001, ((void*)0) } , | |||
| 10578 | ||||
| 10579 | #line 550 | |||
| 10580 | SIMD_INTRIN_DEF(shri_f64){ "shri_f64", simd__intrin_shri_f64, 0x0001, ((void*)0) } , | |||
| 10581 | ||||
| 10582 | #endif // shl_imm | |||
| 10583 | ||||
| 10584 | #line 557 | |||
| 10585 | SIMD_INTRIN_DEF(and_f64){ "and_f64", simd__intrin_and_f64, 0x0001, ((void*)0) } , | |||
| 10586 | ||||
| 10587 | #line 557 | |||
| 10588 | SIMD_INTRIN_DEF(or_f64){ "or_f64", simd__intrin_or_f64, 0x0001, ((void*)0) } , | |||
| 10589 | ||||
| 10590 | #line 557 | |||
| 10591 | SIMD_INTRIN_DEF(xor_f64){ "xor_f64", simd__intrin_xor_f64, 0x0001, ((void*)0) } , | |||
| 10592 | ||||
| 10593 | #line 557 | |||
| 10594 | SIMD_INTRIN_DEF(not_f64){ "not_f64", simd__intrin_not_f64, 0x0001, ((void*)0) } , | |||
| 10595 | ||||
| 10596 | #line 557 | |||
| 10597 | SIMD_INTRIN_DEF(cmpeq_f64){ "cmpeq_f64", simd__intrin_cmpeq_f64, 0x0001, ((void*)0) } , | |||
| 10598 | ||||
| 10599 | #line 557 | |||
| 10600 | SIMD_INTRIN_DEF(cmpneq_f64){ "cmpneq_f64", simd__intrin_cmpneq_f64, 0x0001, ((void*)0) } , | |||
| 10601 | ||||
| 10602 | #line 557 | |||
| 10603 | SIMD_INTRIN_DEF(cmpgt_f64){ "cmpgt_f64", simd__intrin_cmpgt_f64, 0x0001, ((void*)0) } , | |||
| 10604 | ||||
| 10605 | #line 557 | |||
| 10606 | SIMD_INTRIN_DEF(cmpge_f64){ "cmpge_f64", simd__intrin_cmpge_f64, 0x0001, ((void*)0) } , | |||
| 10607 | ||||
| 10608 | #line 557 | |||
| 10609 | SIMD_INTRIN_DEF(cmplt_f64){ "cmplt_f64", simd__intrin_cmplt_f64, 0x0001, ((void*)0) } , | |||
| 10610 | ||||
| 10611 | #line 557 | |||
| 10612 | SIMD_INTRIN_DEF(cmple_f64){ "cmple_f64", simd__intrin_cmple_f64, 0x0001, ((void*)0) } , | |||
| 10613 | ||||
| 10614 | ||||
| 10615 | /*************************** | |||
| 10616 | * Conversion | |||
| 10617 | ***************************/ | |||
| 10618 | SIMD_INTRIN_DEF(cvt_f64_b64){ "cvt_f64_b64", simd__intrin_cvt_f64_b64, 0x0001, ((void*)0) } , | |||
| 10619 | SIMD_INTRIN_DEF(cvt_b64_f64){ "cvt_b64_f64", simd__intrin_cvt_b64_f64, 0x0001, ((void*)0) } , | |||
| 10620 | #if 0 | |||
| 10621 | SIMD_INTRIN_DEF(expand_f64_f64){ "expand_f64_f64", simd__intrin_expand_f64_f64, 0x0001, ((void *)0) } , | |||
| 10622 | #endif // expand_sup | |||
| 10623 | /*************************** | |||
| 10624 | * Arithmetic | |||
| 10625 | ***************************/ | |||
| 10626 | #line 574 | |||
| 10627 | SIMD_INTRIN_DEF(add_f64){ "add_f64", simd__intrin_add_f64, 0x0001, ((void*)0) } , | |||
| 10628 | ||||
| 10629 | #line 574 | |||
| 10630 | SIMD_INTRIN_DEF(sub_f64){ "sub_f64", simd__intrin_sub_f64, 0x0001, ((void*)0) } , | |||
| 10631 | ||||
| 10632 | ||||
| 10633 | #if 0 | |||
| 10634 | #line 581 | |||
| 10635 | SIMD_INTRIN_DEF(adds_f64){ "adds_f64", simd__intrin_adds_f64, 0x0001, ((void*)0) } , | |||
| 10636 | ||||
| 10637 | #line 581 | |||
| 10638 | SIMD_INTRIN_DEF(subs_f64){ "subs_f64", simd__intrin_subs_f64, 0x0001, ((void*)0) } , | |||
| 10639 | ||||
| 10640 | #endif // sat_sup | |||
| 10641 | ||||
| 10642 | #if 1 | |||
| 10643 | SIMD_INTRIN_DEF(mul_f64){ "mul_f64", simd__intrin_mul_f64, 0x0001, ((void*)0) } , | |||
| 10644 | #endif // mul_sup | |||
| 10645 | ||||
| 10646 | #if 1 | |||
| 10647 | SIMD_INTRIN_DEF(div_f64){ "div_f64", simd__intrin_div_f64, 0x0001, ((void*)0) } , | |||
| 10648 | #endif // div_sup | |||
| 10649 | ||||
| 10650 | #if 0 | |||
| 10651 | SIMD_INTRIN_DEF(divisor_f64){ "divisor_f64", simd__intrin_divisor_f64, 0x0001, ((void*)0) } , | |||
| 10652 | SIMD_INTRIN_DEF(divc_f64){ "divc_f64", simd__intrin_divc_f64, 0x0001, ((void*)0) } , | |||
| 10653 | #endif // intdiv_sup | |||
| 10654 | ||||
| 10655 | #if 1 | |||
| 10656 | #line 602 | |||
| 10657 | SIMD_INTRIN_DEF(muladd_f64){ "muladd_f64", simd__intrin_muladd_f64, 0x0001, ((void*)0) } , | |||
| 10658 | ||||
| 10659 | #line 602 | |||
| 10660 | SIMD_INTRIN_DEF(mulsub_f64){ "mulsub_f64", simd__intrin_mulsub_f64, 0x0001, ((void*)0) } , | |||
| 10661 | ||||
| 10662 | #line 602 | |||
| 10663 | SIMD_INTRIN_DEF(nmuladd_f64){ "nmuladd_f64", simd__intrin_nmuladd_f64, 0x0001, ((void*)0) } , | |||
| 10664 | ||||
| 10665 | #line 602 | |||
| 10666 | SIMD_INTRIN_DEF(nmulsub_f64){ "nmulsub_f64", simd__intrin_nmulsub_f64, 0x0001, ((void*)0) } , | |||
| 10667 | ||||
| 10668 | #endif // fused_sup | |||
| 10669 | ||||
| 10670 | #if 1 | |||
| 10671 | SIMD_INTRIN_DEF(sum_f64){ "sum_f64", simd__intrin_sum_f64, 0x0001, ((void*)0) } , | |||
| 10672 | #endif // sum_sup | |||
| 10673 | ||||
| 10674 | #if 0 | |||
| 10675 | SIMD_INTRIN_DEF(sumup_f64){ "sumup_f64", simd__intrin_sumup_f64, 0x0001, ((void*)0) } , | |||
| 10676 | #endif // sumup_sup | |||
| 10677 | /*************************** | |||
| 10678 | * Math | |||
| 10679 | ***************************/ | |||
| 10680 | #if 1 | |||
| 10681 | #line 620 | |||
| 10682 | SIMD_INTRIN_DEF(sqrt_f64){ "sqrt_f64", simd__intrin_sqrt_f64, 0x0001, ((void*)0) } , | |||
| 10683 | ||||
| 10684 | #line 620 | |||
| 10685 | SIMD_INTRIN_DEF(recip_f64){ "recip_f64", simd__intrin_recip_f64, 0x0001, ((void*)0) } , | |||
| 10686 | ||||
| 10687 | #line 620 | |||
| 10688 | SIMD_INTRIN_DEF(abs_f64){ "abs_f64", simd__intrin_abs_f64, 0x0001, ((void*)0) } , | |||
| 10689 | ||||
| 10690 | #line 620 | |||
| 10691 | SIMD_INTRIN_DEF(square_f64){ "square_f64", simd__intrin_square_f64, 0x0001, ((void*)0) } , | |||
| 10692 | ||||
| 10693 | #endif | |||
| 10694 | ||||
| 10695 | #line 627 | |||
| 10696 | SIMD_INTRIN_DEF(max_f64){ "max_f64", simd__intrin_max_f64, 0x0001, ((void*)0) } , | |||
| 10697 | ||||
| 10698 | #line 627 | |||
| 10699 | SIMD_INTRIN_DEF(min_f64){ "min_f64", simd__intrin_min_f64, 0x0001, ((void*)0) } , | |||
| 10700 | ||||
| 10701 | ||||
| 10702 | #if 1 | |||
| 10703 | #line 634 | |||
| 10704 | SIMD_INTRIN_DEF(maxp_f64){ "maxp_f64", simd__intrin_maxp_f64, 0x0001, ((void*)0) } , | |||
| 10705 | ||||
| 10706 | #line 634 | |||
| 10707 | SIMD_INTRIN_DEF(minp_f64){ "minp_f64", simd__intrin_minp_f64, 0x0001, ((void*)0) } , | |||
| 10708 | ||||
| 10709 | #endif | |||
| 10710 | ||||
| 10711 | /*************************** | |||
| 10712 | * Mask operations | |||
| 10713 | ***************************/ | |||
| 10714 | #line 644 | |||
| 10715 | SIMD_INTRIN_DEF(ifadd_f64){ "ifadd_f64", simd__intrin_ifadd_f64, 0x0001, ((void*)0) } , | |||
| 10716 | ||||
| 10717 | #line 644 | |||
| 10718 | SIMD_INTRIN_DEF(ifsub_f64){ "ifsub_f64", simd__intrin_ifsub_f64, 0x0001, ((void*)0) } , | |||
| 10719 | ||||
| 10720 | ||||
| 10721 | #endif // simd_sup | |||
| 10722 | ||||
| 10723 | /************************************************************************* | |||
| 10724 | * Variant | |||
| 10725 | ************************************************************************/ | |||
| 10726 | SIMD_INTRIN_DEF(cleanup){ "cleanup", simd__intrin_cleanup, 0x0001, ((void*)0) } , | |||
| 10727 | ||||
| 10728 | /************************************************************************* | |||
| 10729 | * A special section for f32/f64 intrinsics outside the main repeater | |||
| 10730 | ************************************************************************/ | |||
| 10731 | /*************************** | |||
| 10732 | * Operators | |||
| 10733 | ***************************/ | |||
| 10734 | // check special cases | |||
| 10735 | SIMD_INTRIN_DEF(notnan_f32){ "notnan_f32", simd__intrin_notnan_f32, 0x0001, ((void*)0) } , | |||
| 10736 | #if NPY_SIMD_F641 | |||
| 10737 | SIMD_INTRIN_DEF(notnan_f64){ "notnan_f64", simd__intrin_notnan_f64, 0x0001, ((void*)0) } , | |||
| 10738 | #endif | |||
| 10739 | /*************************** | |||
| 10740 | * Conversions | |||
| 10741 | ***************************/ | |||
| 10742 | // round to nearest integer (assume even) | |||
| 10743 | SIMD_INTRIN_DEF(round_s32_f32){ "round_s32_f32", simd__intrin_round_s32_f32, 0x0001, ((void *)0) } , | |||
| 10744 | #if NPY_SIMD_F641 | |||
| 10745 | SIMD_INTRIN_DEF(round_s32_f64){ "round_s32_f64", simd__intrin_round_s32_f64, 0x0001, ((void *)0) } , | |||
| 10746 | #endif | |||
| 10747 | ||||
| 10748 | /************************************************************************* | |||
| 10749 | * A special section for boolean intrinsics outside the main repeater | |||
| 10750 | ************************************************************************/ | |||
| 10751 | /*************************** | |||
| 10752 | * Operators | |||
| 10753 | ***************************/ | |||
| 10754 | // Logical | |||
| 10755 | #line 684 | |||
| 10756 | SIMD_INTRIN_DEF(and_b8){ "and_b8", simd__intrin_and_b8, 0x0001, ((void*)0) } , | |||
| 10757 | SIMD_INTRIN_DEF(or_b8){ "or_b8", simd__intrin_or_b8, 0x0001, ((void*)0) } , | |||
| 10758 | SIMD_INTRIN_DEF(xor_b8){ "xor_b8", simd__intrin_xor_b8, 0x0001, ((void*)0) } , | |||
| 10759 | SIMD_INTRIN_DEF(not_b8){ "not_b8", simd__intrin_not_b8, 0x0001, ((void*)0) } , | |||
| 10760 | ||||
| 10761 | #line 684 | |||
| 10762 | SIMD_INTRIN_DEF(and_b16){ "and_b16", simd__intrin_and_b16, 0x0001, ((void*)0) } , | |||
| 10763 | SIMD_INTRIN_DEF(or_b16){ "or_b16", simd__intrin_or_b16, 0x0001, ((void*)0) } , | |||
| 10764 | SIMD_INTRIN_DEF(xor_b16){ "xor_b16", simd__intrin_xor_b16, 0x0001, ((void*)0) } , | |||
| 10765 | SIMD_INTRIN_DEF(not_b16){ "not_b16", simd__intrin_not_b16, 0x0001, ((void*)0) } , | |||
| 10766 | ||||
| 10767 | #line 684 | |||
| 10768 | SIMD_INTRIN_DEF(and_b32){ "and_b32", simd__intrin_and_b32, 0x0001, ((void*)0) } , | |||
| 10769 | SIMD_INTRIN_DEF(or_b32){ "or_b32", simd__intrin_or_b32, 0x0001, ((void*)0) } , | |||
| 10770 | SIMD_INTRIN_DEF(xor_b32){ "xor_b32", simd__intrin_xor_b32, 0x0001, ((void*)0) } , | |||
| 10771 | SIMD_INTRIN_DEF(not_b32){ "not_b32", simd__intrin_not_b32, 0x0001, ((void*)0) } , | |||
| 10772 | ||||
| 10773 | #line 684 | |||
| 10774 | SIMD_INTRIN_DEF(and_b64){ "and_b64", simd__intrin_and_b64, 0x0001, ((void*)0) } , | |||
| 10775 | SIMD_INTRIN_DEF(or_b64){ "or_b64", simd__intrin_or_b64, 0x0001, ((void*)0) } , | |||
| 10776 | SIMD_INTRIN_DEF(xor_b64){ "xor_b64", simd__intrin_xor_b64, 0x0001, ((void*)0) } , | |||
| 10777 | SIMD_INTRIN_DEF(not_b64){ "not_b64", simd__intrin_not_b64, 0x0001, ((void*)0) } , | |||
| 10778 | ||||
| 10779 | /*************************** | |||
| 10780 | * Conversions | |||
| 10781 | ***************************/ | |||
| 10782 | // Convert mask vector to integer bitfield | |||
| 10783 | #line 696 | |||
| 10784 | SIMD_INTRIN_DEF(tobits_b8){ "tobits_b8", simd__intrin_tobits_b8, 0x0001, ((void*)0) } , | |||
| 10785 | ||||
| 10786 | #line 696 | |||
| 10787 | SIMD_INTRIN_DEF(tobits_b16){ "tobits_b16", simd__intrin_tobits_b16, 0x0001, ((void*)0) } , | |||
| 10788 | ||||
| 10789 | #line 696 | |||
| 10790 | SIMD_INTRIN_DEF(tobits_b32){ "tobits_b32", simd__intrin_tobits_b32, 0x0001, ((void*)0) } , | |||
| 10791 | ||||
| 10792 | #line 696 | |||
| 10793 | SIMD_INTRIN_DEF(tobits_b64){ "tobits_b64", simd__intrin_tobits_b64, 0x0001, ((void*)0) } , | |||
| 10794 | ||||
| 10795 | ||||
| 10796 | /************************************************************************/ | |||
| 10797 | {NULL((void*)0), NULL((void*)0), 0, NULL((void*)0)} | |||
| 10798 | }; // PyMethodDef | |||
| 10799 | ||||
| 10800 | #endif // NPY_SIMD | |||
| 10801 | ||||
| 10802 | //######################################################################### | |||
| 10803 | //## Defining a separate module for each target | |||
| 10804 | //######################################################################### | |||
| 10805 | NPY_VISIBILITY_HIDDEN__attribute__((visibility("hidden"))) PyObject * | |||
| 10806 | NPY_CPU_DISPATCH_CURFX(simd_create_module)simd_create_module(void) | |||
| 10807 | { | |||
| 10808 | static struct PyModuleDef defs = { | |||
| 10809 | .m_base = PyModuleDef_HEAD_INIT{ { 1, ((void*)0) }, ((void*)0), 0, ((void*)0), }, | |||
| 10810 | .m_size = -1, | |||
| 10811 | #ifdef NPY__CPU_TARGET_CURRENT | |||
| 10812 | .m_name = "numpy.core._simd." NPY_TOSTRING(NPY__CPU_TARGET_CURRENT)"NPY__CPU_TARGET_CURRENT", | |||
| 10813 | #else | |||
| 10814 | .m_name = "numpy.core._simd.baseline", | |||
| 10815 | #endif | |||
| 10816 | #if NPY_SIMD128 | |||
| 10817 | .m_methods = simd__intrinsics_methods | |||
| 10818 | #else | |||
| 10819 | .m_methods = NULL((void*)0) | |||
| 10820 | #endif | |||
| 10821 | }; | |||
| 10822 | PyObject *m = PyModule_Create(&defs)PyModule_Create2(&defs, 1013); | |||
| 10823 | if (m == NULL((void*)0)) { | |||
| 10824 | return NULL((void*)0); | |||
| 10825 | } | |||
| 10826 | if (PyModule_AddIntConstant(m, "simd", NPY_SIMD128)) { | |||
| 10827 | goto err; | |||
| 10828 | } | |||
| 10829 | if (PyModule_AddIntConstant(m, "simd_f64", NPY_SIMD_F641)) { | |||
| 10830 | goto err; | |||
| 10831 | } | |||
| 10832 | if (PyModule_AddIntConstant(m, "simd_fma3", NPY_SIMD_FMA30)) { | |||
| 10833 | goto err; | |||
| 10834 | } | |||
| 10835 | if (PyModule_AddIntConstant(m, "simd_width", NPY_SIMD_WIDTH16)) { | |||
| 10836 | goto err; | |||
| 10837 | } | |||
| 10838 | #if NPY_SIMD128 | |||
| 10839 | if (PySIMDVectorType_Init(m)) { | |||
| 10840 | goto err; | |||
| 10841 | } | |||
| 10842 | #line 748 | |||
| 10843 | if (PyModule_AddIntConstant(m, "nlanes_u8", npyv_nlanes_u816)) { | |||
| 10844 | goto err; | |||
| 10845 | } | |||
| 10846 | ||||
| 10847 | #line 748 | |||
| 10848 | if (PyModule_AddIntConstant(m, "nlanes_s8", npyv_nlanes_s816)) { | |||
| 10849 | goto err; | |||
| 10850 | } | |||
| 10851 | ||||
| 10852 | #line 748 | |||
| 10853 | if (PyModule_AddIntConstant(m, "nlanes_u16", npyv_nlanes_u168)) { | |||
| 10854 | goto err; | |||
| 10855 | } | |||
| 10856 | ||||
| 10857 | #line 748 | |||
| 10858 | if (PyModule_AddIntConstant(m, "nlanes_s16", npyv_nlanes_s168)) { | |||
| 10859 | goto err; | |||
| 10860 | } | |||
| 10861 | ||||
| 10862 | #line 748 | |||
| 10863 | if (PyModule_AddIntConstant(m, "nlanes_u32", npyv_nlanes_u324)) { | |||
| 10864 | goto err; | |||
| 10865 | } | |||
| 10866 | ||||
| 10867 | #line 748 | |||
| 10868 | if (PyModule_AddIntConstant(m, "nlanes_s32", npyv_nlanes_s324)) { | |||
| 10869 | goto err; | |||
| 10870 | } | |||
| 10871 | ||||
| 10872 | #line 748 | |||
| 10873 | if (PyModule_AddIntConstant(m, "nlanes_u64", npyv_nlanes_u642)) { | |||
| 10874 | goto err; | |||
| 10875 | } | |||
| 10876 | ||||
| 10877 | #line 748 | |||
| 10878 | if (PyModule_AddIntConstant(m, "nlanes_s64", npyv_nlanes_s642)) { | |||
| 10879 | goto err; | |||
| 10880 | } | |||
| 10881 | ||||
| 10882 | #line 748 | |||
| 10883 | if (PyModule_AddIntConstant(m, "nlanes_f32", npyv_nlanes_f324)) { | |||
| 10884 | goto err; | |||
| 10885 | } | |||
| 10886 | ||||
| 10887 | #line 748 | |||
| 10888 | if (PyModule_AddIntConstant(m, "nlanes_f64", npyv_nlanes_f642)) { | |||
| 10889 | goto err; | |||
| 10890 | } | |||
| 10891 | ||||
| 10892 | #endif // NPY_SIMD | |||
| 10893 | return m; | |||
| 10894 | err: | |||
| 10895 | Py_DECREF(m)_Py_DECREF(((PyObject*)(m))); | |||
| 10896 | return NULL((void*)0); | |||
| 10897 | } | |||
| 10898 |
| 1 | /** | |||
| 2 | * This file is included by `_simd.dispatch.c.src`. Its contents are affected by the simd configuration, and | |||
| 3 | * therefore must be built multiple times. Making it a standalone `.c` file with `NPY_VISIBILITY_HIDDEN` | |||
| 4 | * symbols would require judicious use of `NPY_CPU_DISPATCH_DECLARE` and `NPY_CPU_DISPATCH_CURFX`, which was | |||
| 5 | * deemed too harmful to readability. | |||
| 6 | */ | |||
| 7 | /************************************ | |||
| 8 | ** Protected Definitions | |||
| 9 | ************************************/ | |||
| 10 | static simd_data | |||
| 11 | simd_scalar_from_number(PyObject *obj, simd_data_type dtype) | |||
| 12 | { | |||
| 13 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 14 | assert(info->is_scalar && info->lane_size > 0)((void) (0)); | |||
| 15 | simd_data data; | |||
| 16 | if (info->is_float) { | |||
| 17 | data.f64 = PyFloat_AsDouble(obj); | |||
| 18 | if (dtype == simd_data_f32){ | |||
| 19 | data.f32 = (float)data.f64; | |||
| 20 | } | |||
| 21 | } else { | |||
| 22 | data.u64 = PyLong_AsUnsignedLongLongMask(obj); | |||
| 23 | } | |||
| 24 | return data; | |||
| 25 | } | |||
| 26 | ||||
| 27 | static PyObject * | |||
| 28 | simd_scalar_to_number(simd_data data, simd_data_type dtype) | |||
| 29 | { | |||
| 30 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 31 | assert(info->is_scalar && info->lane_size > 0)((void) (0)); | |||
| 32 | if (info->is_float) { | |||
| 33 | if (dtype == simd_data_f32) { | |||
| 34 | return PyFloat_FromDouble(data.f32); | |||
| 35 | } | |||
| 36 | return PyFloat_FromDouble(data.f64); | |||
| 37 | } | |||
| 38 | int leftb = (sizeof(npyv_lanetype_u64) - info->lane_size) * 8; | |||
| 39 | data.u64 <<= leftb; | |||
| 40 | if (info->is_signed) { | |||
| 41 | return PyLong_FromLongLong(data.s64 >> leftb); | |||
| 42 | } | |||
| 43 | return PyLong_FromUnsignedLongLong(data.u64 >> leftb); | |||
| 44 | } | |||
| 45 | ||||
| 46 | typedef struct { | |||
| 47 | Py_ssize_t len; | |||
| 48 | void *ptr; | |||
| 49 | } simd__alloc_data; | |||
| 50 | ||||
| 51 | static void * | |||
| 52 | simd_sequence_new(Py_ssize_t len, simd_data_type dtype) | |||
| 53 | { | |||
| 54 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 55 | assert(len > 0 && info->is_sequence && info->lane_size > 0)((void) (0)); | |||
| 56 | size_t size = sizeof(simd__alloc_data) + len * info->lane_size + NPY_SIMD_WIDTH16; | |||
| 57 | void *ptr = malloc(size); | |||
| 58 | if (ptr == NULL((void*)0)) { | |||
| 59 | return PyErr_NoMemory(); | |||
| 60 | } | |||
| 61 | // align the pointer | |||
| 62 | simd__alloc_data *a_ptr = (simd__alloc_data *)( | |||
| 63 | ((uintptr_t)ptr + sizeof(simd__alloc_data) + NPY_SIMD_WIDTH16) & ~(uintptr_t)(NPY_SIMD_WIDTH16-1) | |||
| 64 | ); | |||
| 65 | a_ptr[-1].len = len; | |||
| 66 | a_ptr[-1].ptr = ptr; | |||
| 67 | return a_ptr; | |||
| 68 | } | |||
| 69 | ||||
| 70 | static Py_ssize_t | |||
| 71 | simd_sequence_len(void const *ptr) | |||
| 72 | { | |||
| 73 | return ((simd__alloc_data const*)ptr)[-1].len; | |||
| 74 | } | |||
| 75 | ||||
| 76 | static void | |||
| 77 | simd_sequence_free(void *ptr) | |||
| 78 | { | |||
| 79 | free(((simd__alloc_data *)ptr)[-1].ptr); | |||
| 80 | } | |||
| 81 | ||||
| 82 | static void * | |||
| 83 | simd_sequence_from_iterable(PyObject *obj, simd_data_type dtype, Py_ssize_t min_size) | |||
| 84 | { | |||
| 85 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 86 | assert(info->is_sequence && info->lane_size > 0)((void) (0)); | |||
| 87 | PyObject *seq_obj = PySequence_Fast(obj, "expected a sequence"); | |||
| ||||
| 88 | if (seq_obj == NULL((void*)0)) { | |||
| 89 | return NULL((void*)0); | |||
| 90 | } | |||
| 91 | Py_ssize_t seq_size = PySequence_Fast_GET_SIZE(seq_obj)(((((((PyObject*)(seq_obj))->ob_type))->tp_flags & ( (1UL << 25))) != 0) ? (((void) (0)), (((PyVarObject*)(seq_obj ))->ob_size)) : (((PyVarObject*)((((void) (0)), (PyTupleObject *)(seq_obj))))->ob_size)); | |||
| 92 | if (seq_size < min_size) { | |||
| 93 | PyErr_Format(PyExc_ValueError, | |||
| 94 | "minimum acceptable size of the required sequence is %d, given(%d)", | |||
| 95 | min_size, seq_size | |||
| 96 | ); | |||
| 97 | return NULL((void*)0); | |||
| 98 | } | |||
| 99 | npyv_lanetype_u8 *dst = simd_sequence_new(seq_size, dtype); | |||
| 100 | if (dst == NULL((void*)0)) { | |||
| 101 | return NULL((void*)0); | |||
| 102 | } | |||
| 103 | PyObject **seq_items = PySequence_Fast_ITEMS(seq_obj)(((((((PyObject*)(seq_obj))->ob_type))->tp_flags & ( (1UL << 25))) != 0) ? ((PyListObject *)(seq_obj))->ob_item : ((PyTupleObject *)(seq_obj))->ob_item); | |||
| 104 | for (Py_ssize_t i = 0; i < seq_size; ++i) { | |||
| 105 | simd_data data = simd_scalar_from_number(seq_items[i], info->to_scalar); | |||
| 106 | npyv_lanetype_u8 *sdst = dst + i * info->lane_size; | |||
| 107 | memcpy(sdst, &data.u64, info->lane_size); | |||
| 108 | } | |||
| 109 | Py_DECREF(seq_obj)_Py_DECREF(((PyObject*)(seq_obj))); | |||
| 110 | ||||
| 111 | if (PyErr_Occurred()) { | |||
| 112 | simd_sequence_free(dst); | |||
| 113 | return NULL((void*)0); | |||
| 114 | } | |||
| 115 | return dst; | |||
| 116 | } | |||
| 117 | ||||
| 118 | static int | |||
| 119 | simd_sequence_fill_iterable(PyObject *obj, const void *ptr, simd_data_type dtype) | |||
| 120 | { | |||
| 121 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 122 | if (!PySequence_Check(obj)) { | |||
| 123 | PyErr_Format(PyExc_TypeError, | |||
| 124 | "a sequence object is required to fill %s", info->pyname | |||
| 125 | ); | |||
| 126 | return -1; | |||
| 127 | } | |||
| 128 | const npyv_lanetype_u8 *src = ptr; | |||
| 129 | Py_ssize_t seq_len = simd_sequence_len(ptr); | |||
| 130 | for (Py_ssize_t i = 0; i < seq_len; ++i) { | |||
| 131 | const npyv_lanetype_u8 *ssrc = src + i * info->lane_size; | |||
| 132 | simd_data data; | |||
| 133 | memcpy(&data.u64, ssrc, info->lane_size); | |||
| 134 | PyObject *item = simd_scalar_to_number(data, info->to_scalar); | |||
| 135 | if (item == NULL((void*)0)) { | |||
| 136 | return -1; | |||
| 137 | } | |||
| 138 | int res = PySequence_SetItem(obj, i, item); | |||
| 139 | Py_DECREF(item)_Py_DECREF(((PyObject*)(item))); | |||
| 140 | if (res < 0) { | |||
| 141 | return -1; | |||
| 142 | } | |||
| 143 | } | |||
| 144 | return 0; | |||
| 145 | } | |||
| 146 | ||||
| 147 | static PyObject * | |||
| 148 | simd_sequence_to_list(const void *ptr, simd_data_type dtype) | |||
| 149 | { | |||
| 150 | PyObject *list = PyList_New(simd_sequence_len(ptr)); | |||
| 151 | if (list == NULL((void*)0)) { | |||
| 152 | return NULL((void*)0); | |||
| 153 | } | |||
| 154 | if (simd_sequence_fill_iterable(list, ptr, dtype) < 0) { | |||
| 155 | Py_DECREF(list)_Py_DECREF(((PyObject*)(list))); | |||
| 156 | return NULL((void*)0); | |||
| 157 | } | |||
| 158 | return list; | |||
| 159 | } | |||
| 160 | ||||
| 161 | static simd_data | |||
| 162 | simd_vectorx_from_tuple(PyObject *obj, simd_data_type dtype) | |||
| 163 | { | |||
| 164 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 165 | // NPYV currently only supports x2 and x3 | |||
| 166 | assert(info->is_vectorx > 1 && info->is_vectorx < 4)((void) (0)); | |||
| 167 | ||||
| 168 | simd_data data = {.u64 = 0}; | |||
| 169 | if (!PyTuple_Check(obj)((((((PyObject*)(obj))->ob_type))->tp_flags & ((1UL << 26))) != 0) || PyTuple_GET_SIZE(obj)(((PyVarObject*)((((void) (0)), (PyTupleObject *)(obj))))-> ob_size) != info->is_vectorx) { | |||
| 170 | PyErr_Format(PyExc_TypeError, | |||
| 171 | "a tuple of %d vector type %s is required", | |||
| 172 | info->is_vectorx, simd_data_getinfo(info->to_vector)->pyname | |||
| 173 | ); | |||
| 174 | return data; | |||
| 175 | } | |||
| 176 | for (int i = 0; i < info->is_vectorx; ++i) { | |||
| 177 | PyObject *item = PyTuple_GET_ITEM(obj, i)((((void) (0)), (PyTupleObject *)(obj))->ob_item[i]); | |||
| 178 | // get the max multi-vec and let the compiler do the rest | |||
| 179 | data.vu64x3.val[i] = PySIMDVector_AsData((PySIMDVectorObject*)item, info->to_vector).vu64; | |||
| 180 | if (PyErr_Occurred()) { | |||
| 181 | return data; | |||
| 182 | } | |||
| 183 | } | |||
| 184 | return data; | |||
| 185 | } | |||
| 186 | ||||
| 187 | static PyObject * | |||
| 188 | simd_vectorx_to_tuple(simd_data data, simd_data_type dtype) | |||
| 189 | { | |||
| 190 | const simd_data_info *info = simd_data_getinfo(dtype); | |||
| 191 | // NPYV currently only supports x2 and x3 | |||
| 192 | assert(info->is_vectorx > 1 && info->is_vectorx < 4)((void) (0)); | |||
| 193 | ||||
| 194 | PyObject *tuple = PyTuple_New(info->is_vectorx); | |||
| 195 | if (tuple == NULL((void*)0)) { | |||
| 196 | return NULL((void*)0); | |||
| 197 | } | |||
| 198 | for (int i = 0; i < info->is_vectorx; ++i) { | |||
| 199 | // get the max multi-vector and let the compiler handle the rest | |||
| 200 | simd_data vdata = {.vu64 = data.vu64x3.val[i]}; | |||
| 201 | PyObject *item = (PyObject*)PySIMDVector_FromData(vdata, info->to_vector); | |||
| 202 | if (item == NULL((void*)0)) { | |||
| 203 | // TODO: improve log add item number | |||
| 204 | Py_DECREF(tuple)_Py_DECREF(((PyObject*)(tuple))); | |||
| 205 | return NULL((void*)0); | |||
| 206 | } | |||
| 207 | PyTuple_SET_ITEM(tuple, i, item)PyTuple_SetItem(tuple, i, item); | |||
| 208 | } | |||
| 209 | return tuple; | |||
| 210 | } |
| 1 | #ifndef PySequence_Fast |
| 2 | struct _object; |
| 3 | typedef struct _object PyObject; |
| 4 | PyObject* clang_analyzer_PyObject_New_Reference(); |
| 5 | PyObject* PySequence_Fast(PyObject *o, const char *m) { |
| 6 | return clang_analyzer_PyObject_New_Reference(); |
| 7 | } |
| 8 | #else |
| 9 | #warning "API PySequence_Fast is defined as a macro." |
| 10 | #endif |