-: 0:Source:/usr/lib/gcc/x86_64-linux-gnu/14/include/emmintrin.h
-: 0:Graph:tspi4.gcno
-: 0:Data:tspi4.gcda
-: 0:Runs:1
-: 1:/* Copyright (C) 2003-2024 Free Software Foundation, Inc.
-: 2:
-: 3: This file is part of GCC.
-: 4:
-: 5: GCC is free software; you can redistribute it and/or modify
-: 6: it under the terms of the GNU General Public License as published by
-: 7: the Free Software Foundation; either version 3, or (at your option)
-: 8: any later version.
-: 9:
-: 10: GCC is distributed in the hope that it will be useful,
-: 11: but WITHOUT ANY WARRANTY; without even the implied warranty of
-: 12: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-: 13: GNU General Public License for more details.
-: 14:
-: 15: Under Section 7 of GPL version 3, you are granted additional
-: 16: permissions described in the GCC Runtime Library Exception, version
-: 17: 3.1, as published by the Free Software Foundation.
-: 18:
-: 19: You should have received a copy of the GNU General Public License and
-: 20: a copy of the GCC Runtime Library Exception along with this program;
-: 21: see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
-: 22: . */
-: 23:
-: 24:/* Implemented from the specification included in the Intel C++ Compiler
-: 25: User Guide and Reference, version 9.0. */
-: 26:
-: 27:#ifndef _EMMINTRIN_H_INCLUDED
-: 28:#define _EMMINTRIN_H_INCLUDED
-: 29:
-: 30:/* We need definitions from the SSE header files*/
-: 31:#include
-: 32:
-: 33:#ifndef __SSE2__
-: 34:#pragma GCC push_options
-: 35:#pragma GCC target("sse2")
-: 36:#define __DISABLE_SSE2__
-: 37:#endif /* __SSE2__ */
-: 38:
-: 39:/* SSE2 */
-: 40:typedef double __v2df __attribute__ ((__vector_size__ (16)));
-: 41:typedef long long __v2di __attribute__ ((__vector_size__ (16)));
-: 42:typedef unsigned long long __v2du __attribute__ ((__vector_size__ (16)));
-: 43:typedef int __v4si __attribute__ ((__vector_size__ (16)));
-: 44:typedef unsigned int __v4su __attribute__ ((__vector_size__ (16)));
-: 45:typedef short __v8hi __attribute__ ((__vector_size__ (16)));
-: 46:typedef unsigned short __v8hu __attribute__ ((__vector_size__ (16)));
-: 47:typedef char __v16qi __attribute__ ((__vector_size__ (16)));
-: 48:typedef signed char __v16qs __attribute__ ((__vector_size__ (16)));
-: 49:typedef unsigned char __v16qu __attribute__ ((__vector_size__ (16)));
-: 50:
-: 51:/* The Intel API is flexible enough that we must allow aliasing with other
-: 52: vector types, and their scalar components. */
-: 53:typedef long long __m128i __attribute__ ((__vector_size__ (16), __may_alias__));
-: 54:typedef double __m128d __attribute__ ((__vector_size__ (16), __may_alias__));
-: 55:
-: 56:/* Unaligned version of the same types. */
-: 57:typedef long long __m128i_u __attribute__ ((__vector_size__ (16), __may_alias__, __aligned__ (1)));
-: 58:typedef double __m128d_u __attribute__ ((__vector_size__ (16), __may_alias__, __aligned__ (1)));
-: 59:
-: 60:/* Create a selector for use with the SHUFPD instruction. */
-: 61:#define _MM_SHUFFLE2(fp1,fp0) \
-: 62: (((fp1) << 1) | (fp0))
-: 63:
-: 64:/* Create a vector with element 0 as F and the rest zero. */
-: 65:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 66:_mm_set_sd (double __F)
-: 67:{
-: 68: return __extension__ (__m128d){ __F, 0.0 };
-: 69:}
-: 70:
-: 71:/* Create a vector with both elements equal to F. */
-: 72:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 73:_mm_set1_pd (double __F)
-: 74:{
-: 75: return __extension__ (__m128d){ __F, __F };
-: 76:}
-: 77:
-: 78:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 79:_mm_set_pd1 (double __F)
-: 80:{
-: 81: return _mm_set1_pd (__F);
-: 82:}
-: 83:
-: 84:/* Create a vector with the lower value X and upper value W. */
-: 85:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 86:_mm_set_pd (double __W, double __X)
-: 87:{
-: 88: return __extension__ (__m128d){ __X, __W };
-: 89:}
-: 90:
-: 91:/* Create a vector with the lower value W and upper value X. */
-: 92:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 93:_mm_setr_pd (double __W, double __X)
-: 94:{
-: 95: return __extension__ (__m128d){ __W, __X };
-: 96:}
-: 97:
-: 98:/* Create an undefined vector. */
-: 99:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 100:_mm_undefined_pd (void)
-: 101:{
-: 102:#pragma GCC diagnostic push
-: 103:#pragma GCC diagnostic ignored "-Winit-self"
-: 104: __m128d __Y = __Y;
-: 105:#pragma GCC diagnostic pop
-: 106: return __Y;
-: 107:}
-: 108:
-: 109:/* Create a vector of zeros. */
-: 110:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 111:_mm_setzero_pd (void)
-: 112:{
-: 113: return __extension__ (__m128d){ 0.0, 0.0 };
-: 114:}
-: 115:
-: 116:/* Sets the low DPFP value of A from the low value of B. */
-: 117:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 118:_mm_move_sd (__m128d __A, __m128d __B)
-: 119:{
-: 120: return __extension__ (__m128d) __builtin_shuffle ((__v2df)__A, (__v2df)__B, (__v2di){2, 1});
-: 121:}
-: 122:
-: 123:/* Load two DPFP values from P. The address must be 16-byte aligned. */
-: 124:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 125:_mm_load_pd (double const *__P)
-: 126:{
-: 127: return *(__m128d *)__P;
-: 128:}
-: 129:
-: 130:/* Load two DPFP values from P. The address need not be 16-byte aligned. */
-: 131:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 132:_mm_loadu_pd (double const *__P)
-: 133:{
-: 134: return *(__m128d_u *)__P;
-: 135:}
-: 136:
-: 137:/* Create a vector with all two elements equal to *P. */
-: 138:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 139:_mm_load1_pd (double const *__P)
-: 140:{
-: 141: return _mm_set1_pd (*__P);
-: 142:}
-: 143:
-: 144:/* Create a vector with element 0 as *P and the rest zero. */
-: 145:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 146:_mm_load_sd (double const *__P)
-: 147:{
-: 148: return _mm_set_sd (*__P);
-: 149:}
-: 150:
-: 151:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 152:_mm_load_pd1 (double const *__P)
-: 153:{
-: 154: return _mm_load1_pd (__P);
-: 155:}
-: 156:
-: 157:/* Load two DPFP values in reverse order. The address must be aligned. */
-: 158:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 159:_mm_loadr_pd (double const *__P)
-: 160:{
-: 161: __m128d __tmp = _mm_load_pd (__P);
-: 162: return __builtin_ia32_shufpd (__tmp, __tmp, _MM_SHUFFLE2 (0,1));
-: 163:}
-: 164:
-: 165:/* Store two DPFP values. The address must be 16-byte aligned. */
-: 166:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 167:_mm_store_pd (double *__P, __m128d __A)
-: 168:{
-: 169: *(__m128d *)__P = __A;
-: 170:}
-: 171:
-: 172:/* Store two DPFP values. The address need not be 16-byte aligned. */
-: 173:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 174:_mm_storeu_pd (double *__P, __m128d __A)
-: 175:{
-: 176: *(__m128d_u *)__P = __A;
-: 177:}
-: 178:
-: 179:/* Stores the lower DPFP value. */
-: 180:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 181:_mm_store_sd (double *__P, __m128d __A)
-: 182:{
90182: 183: *__P = ((__v2df)__A)[0];
-: 184:}
-: 185:
-: 186:extern __inline double __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 187:_mm_cvtsd_f64 (__m128d __A)
-: 188:{
-: 189: return ((__v2df)__A)[0];
-: 190:}
-: 191:
-: 192:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 193:_mm_storel_pd (double *__P, __m128d __A)
-: 194:{
-: 195: _mm_store_sd (__P, __A);
-: 196:}
-: 197:
-: 198:/* Stores the upper DPFP value. */
-: 199:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 200:_mm_storeh_pd (double *__P, __m128d __A)
-: 201:{
-: 202: *__P = ((__v2df)__A)[1];
-: 203:}
-: 204:
-: 205:/* Store the lower DPFP value across two words.
-: 206: The address must be 16-byte aligned. */
-: 207:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 208:_mm_store1_pd (double *__P, __m128d __A)
-: 209:{
-: 210: _mm_store_pd (__P, __builtin_ia32_shufpd (__A, __A, _MM_SHUFFLE2 (0,0)));
-: 211:}
-: 212:
-: 213:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 214:_mm_store_pd1 (double *__P, __m128d __A)
-: 215:{
-: 216: _mm_store1_pd (__P, __A);
-: 217:}
-: 218:
-: 219:/* Store two DPFP values in reverse order. The address must be aligned. */
-: 220:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 221:_mm_storer_pd (double *__P, __m128d __A)
-: 222:{
-: 223: _mm_store_pd (__P, __builtin_ia32_shufpd (__A, __A, _MM_SHUFFLE2 (0,1)));
-: 224:}
-: 225:
-: 226:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 227:_mm_cvtsi128_si32 (__m128i __A)
-: 228:{
-: 229: return __builtin_ia32_vec_ext_v4si ((__v4si)__A, 0);
-: 230:}
-: 231:
-: 232:#ifdef __x86_64__
-: 233:/* Intel intrinsic. */
-: 234:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 235:_mm_cvtsi128_si64 (__m128i __A)
-: 236:{
-: 237: return ((__v2di)__A)[0];
-: 238:}
-: 239:
-: 240:/* Microsoft intrinsic. */
-: 241:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 242:_mm_cvtsi128_si64x (__m128i __A)
-: 243:{
-: 244: return ((__v2di)__A)[0];
-: 245:}
-: 246:#endif
-: 247:
-: 248:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 249:_mm_add_pd (__m128d __A, __m128d __B)
-: 250:{
-: 251: return (__m128d) ((__v2df)__A + (__v2df)__B);
-: 252:}
-: 253:
-: 254:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 255:_mm_add_sd (__m128d __A, __m128d __B)
-: 256:{
-: 257: return (__m128d)__builtin_ia32_addsd ((__v2df)__A, (__v2df)__B);
-: 258:}
-: 259:
-: 260:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 261:_mm_sub_pd (__m128d __A, __m128d __B)
-: 262:{
-: 263: return (__m128d) ((__v2df)__A - (__v2df)__B);
-: 264:}
-: 265:
-: 266:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 267:_mm_sub_sd (__m128d __A, __m128d __B)
-: 268:{
-: 269: return (__m128d)__builtin_ia32_subsd ((__v2df)__A, (__v2df)__B);
-: 270:}
-: 271:
-: 272:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 273:_mm_mul_pd (__m128d __A, __m128d __B)
-: 274:{
-: 275: return (__m128d) ((__v2df)__A * (__v2df)__B);
-: 276:}
-: 277:
-: 278:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 279:_mm_mul_sd (__m128d __A, __m128d __B)
-: 280:{
-: 281: return (__m128d)__builtin_ia32_mulsd ((__v2df)__A, (__v2df)__B);
-: 282:}
-: 283:
-: 284:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 285:_mm_div_pd (__m128d __A, __m128d __B)
-: 286:{
-: 287: return (__m128d) ((__v2df)__A / (__v2df)__B);
-: 288:}
-: 289:
-: 290:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 291:_mm_div_sd (__m128d __A, __m128d __B)
-: 292:{
-: 293: return (__m128d)__builtin_ia32_divsd ((__v2df)__A, (__v2df)__B);
-: 294:}
-: 295:
-: 296:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 297:_mm_sqrt_pd (__m128d __A)
-: 298:{
-: 299: return (__m128d)__builtin_ia32_sqrtpd ((__v2df)__A);
-: 300:}
-: 301:
-: 302:/* Return pair {sqrt (B[0]), A[1]}. */
-: 303:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 304:_mm_sqrt_sd (__m128d __A, __m128d __B)
-: 305:{
-: 306: __v2df __tmp = __builtin_ia32_movsd ((__v2df)__A, (__v2df)__B);
-: 307: return (__m128d)__builtin_ia32_sqrtsd ((__v2df)__tmp);
-: 308:}
-: 309:
-: 310:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 311:_mm_min_pd (__m128d __A, __m128d __B)
-: 312:{
-: 313: return (__m128d)__builtin_ia32_minpd ((__v2df)__A, (__v2df)__B);
-: 314:}
-: 315:
-: 316:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 317:_mm_min_sd (__m128d __A, __m128d __B)
-: 318:{
-: 319: return (__m128d)__builtin_ia32_minsd ((__v2df)__A, (__v2df)__B);
-: 320:}
-: 321:
-: 322:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 323:_mm_max_pd (__m128d __A, __m128d __B)
-: 324:{
-: 325: return (__m128d)__builtin_ia32_maxpd ((__v2df)__A, (__v2df)__B);
-: 326:}
-: 327:
-: 328:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 329:_mm_max_sd (__m128d __A, __m128d __B)
-: 330:{
-: 331: return (__m128d)__builtin_ia32_maxsd ((__v2df)__A, (__v2df)__B);
-: 332:}
-: 333:
-: 334:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 335:_mm_and_pd (__m128d __A, __m128d __B)
-: 336:{
-: 337: return (__m128d)__builtin_ia32_andpd ((__v2df)__A, (__v2df)__B);
-: 338:}
-: 339:
-: 340:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 341:_mm_andnot_pd (__m128d __A, __m128d __B)
-: 342:{
-: 343: return (__m128d)__builtin_ia32_andnpd ((__v2df)__A, (__v2df)__B);
-: 344:}
-: 345:
-: 346:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 347:_mm_or_pd (__m128d __A, __m128d __B)
-: 348:{
-: 349: return (__m128d)__builtin_ia32_orpd ((__v2df)__A, (__v2df)__B);
-: 350:}
-: 351:
-: 352:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 353:_mm_xor_pd (__m128d __A, __m128d __B)
-: 354:{
-: 355: return (__m128d)__builtin_ia32_xorpd ((__v2df)__A, (__v2df)__B);
-: 356:}
-: 357:
-: 358:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 359:_mm_cmpeq_pd (__m128d __A, __m128d __B)
-: 360:{
-: 361: return (__m128d)__builtin_ia32_cmpeqpd ((__v2df)__A, (__v2df)__B);
-: 362:}
-: 363:
-: 364:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 365:_mm_cmplt_pd (__m128d __A, __m128d __B)
-: 366:{
-: 367: return (__m128d)__builtin_ia32_cmpltpd ((__v2df)__A, (__v2df)__B);
-: 368:}
-: 369:
-: 370:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 371:_mm_cmple_pd (__m128d __A, __m128d __B)
-: 372:{
-: 373: return (__m128d)__builtin_ia32_cmplepd ((__v2df)__A, (__v2df)__B);
-: 374:}
-: 375:
-: 376:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 377:_mm_cmpgt_pd (__m128d __A, __m128d __B)
-: 378:{
-: 379: return (__m128d)__builtin_ia32_cmpgtpd ((__v2df)__A, (__v2df)__B);
-: 380:}
-: 381:
-: 382:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 383:_mm_cmpge_pd (__m128d __A, __m128d __B)
-: 384:{
-: 385: return (__m128d)__builtin_ia32_cmpgepd ((__v2df)__A, (__v2df)__B);
-: 386:}
-: 387:
-: 388:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 389:_mm_cmpneq_pd (__m128d __A, __m128d __B)
-: 390:{
-: 391: return (__m128d)__builtin_ia32_cmpneqpd ((__v2df)__A, (__v2df)__B);
-: 392:}
-: 393:
-: 394:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 395:_mm_cmpnlt_pd (__m128d __A, __m128d __B)
-: 396:{
-: 397: return (__m128d)__builtin_ia32_cmpnltpd ((__v2df)__A, (__v2df)__B);
-: 398:}
-: 399:
-: 400:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 401:_mm_cmpnle_pd (__m128d __A, __m128d __B)
-: 402:{
-: 403: return (__m128d)__builtin_ia32_cmpnlepd ((__v2df)__A, (__v2df)__B);
-: 404:}
-: 405:
-: 406:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 407:_mm_cmpngt_pd (__m128d __A, __m128d __B)
-: 408:{
-: 409: return (__m128d)__builtin_ia32_cmpngtpd ((__v2df)__A, (__v2df)__B);
-: 410:}
-: 411:
-: 412:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 413:_mm_cmpnge_pd (__m128d __A, __m128d __B)
-: 414:{
-: 415: return (__m128d)__builtin_ia32_cmpngepd ((__v2df)__A, (__v2df)__B);
-: 416:}
-: 417:
-: 418:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 419:_mm_cmpord_pd (__m128d __A, __m128d __B)
-: 420:{
-: 421: return (__m128d)__builtin_ia32_cmpordpd ((__v2df)__A, (__v2df)__B);
-: 422:}
-: 423:
-: 424:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 425:_mm_cmpunord_pd (__m128d __A, __m128d __B)
-: 426:{
-: 427: return (__m128d)__builtin_ia32_cmpunordpd ((__v2df)__A, (__v2df)__B);
-: 428:}
-: 429:
-: 430:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 431:_mm_cmpeq_sd (__m128d __A, __m128d __B)
-: 432:{
-: 433: return (__m128d)__builtin_ia32_cmpeqsd ((__v2df)__A, (__v2df)__B);
-: 434:}
-: 435:
-: 436:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 437:_mm_cmplt_sd (__m128d __A, __m128d __B)
-: 438:{
-: 439: return (__m128d)__builtin_ia32_cmpltsd ((__v2df)__A, (__v2df)__B);
-: 440:}
-: 441:
-: 442:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 443:_mm_cmple_sd (__m128d __A, __m128d __B)
-: 444:{
-: 445: return (__m128d)__builtin_ia32_cmplesd ((__v2df)__A, (__v2df)__B);
-: 446:}
-: 447:
-: 448:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 449:_mm_cmpgt_sd (__m128d __A, __m128d __B)
-: 450:{
-: 451: return (__m128d) __builtin_ia32_movsd ((__v2df) __A,
-: 452: (__v2df)
-: 453: __builtin_ia32_cmpltsd ((__v2df) __B,
-: 454: (__v2df)
-: 455: __A));
-: 456:}
-: 457:
-: 458:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 459:_mm_cmpge_sd (__m128d __A, __m128d __B)
-: 460:{
-: 461: return (__m128d) __builtin_ia32_movsd ((__v2df) __A,
-: 462: (__v2df)
-: 463: __builtin_ia32_cmplesd ((__v2df) __B,
-: 464: (__v2df)
-: 465: __A));
-: 466:}
-: 467:
-: 468:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 469:_mm_cmpneq_sd (__m128d __A, __m128d __B)
-: 470:{
-: 471: return (__m128d)__builtin_ia32_cmpneqsd ((__v2df)__A, (__v2df)__B);
-: 472:}
-: 473:
-: 474:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 475:_mm_cmpnlt_sd (__m128d __A, __m128d __B)
-: 476:{
-: 477: return (__m128d)__builtin_ia32_cmpnltsd ((__v2df)__A, (__v2df)__B);
-: 478:}
-: 479:
-: 480:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 481:_mm_cmpnle_sd (__m128d __A, __m128d __B)
-: 482:{
-: 483: return (__m128d)__builtin_ia32_cmpnlesd ((__v2df)__A, (__v2df)__B);
-: 484:}
-: 485:
-: 486:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 487:_mm_cmpngt_sd (__m128d __A, __m128d __B)
-: 488:{
-: 489: return (__m128d) __builtin_ia32_movsd ((__v2df) __A,
-: 490: (__v2df)
-: 491: __builtin_ia32_cmpnltsd ((__v2df) __B,
-: 492: (__v2df)
-: 493: __A));
-: 494:}
-: 495:
-: 496:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 497:_mm_cmpnge_sd (__m128d __A, __m128d __B)
-: 498:{
-: 499: return (__m128d) __builtin_ia32_movsd ((__v2df) __A,
-: 500: (__v2df)
-: 501: __builtin_ia32_cmpnlesd ((__v2df) __B,
-: 502: (__v2df)
-: 503: __A));
-: 504:}
-: 505:
-: 506:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 507:_mm_cmpord_sd (__m128d __A, __m128d __B)
-: 508:{
-: 509: return (__m128d)__builtin_ia32_cmpordsd ((__v2df)__A, (__v2df)__B);
-: 510:}
-: 511:
-: 512:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 513:_mm_cmpunord_sd (__m128d __A, __m128d __B)
-: 514:{
-: 515: return (__m128d)__builtin_ia32_cmpunordsd ((__v2df)__A, (__v2df)__B);
-: 516:}
-: 517:
-: 518:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 519:_mm_comieq_sd (__m128d __A, __m128d __B)
-: 520:{
-: 521: return __builtin_ia32_comisdeq ((__v2df)__A, (__v2df)__B);
-: 522:}
-: 523:
-: 524:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 525:_mm_comilt_sd (__m128d __A, __m128d __B)
-: 526:{
-: 527: return __builtin_ia32_comisdlt ((__v2df)__A, (__v2df)__B);
-: 528:}
-: 529:
-: 530:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 531:_mm_comile_sd (__m128d __A, __m128d __B)
-: 532:{
-: 533: return __builtin_ia32_comisdle ((__v2df)__A, (__v2df)__B);
-: 534:}
-: 535:
-: 536:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 537:_mm_comigt_sd (__m128d __A, __m128d __B)
-: 538:{
-: 539: return __builtin_ia32_comisdgt ((__v2df)__A, (__v2df)__B);
-: 540:}
-: 541:
-: 542:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 543:_mm_comige_sd (__m128d __A, __m128d __B)
-: 544:{
-: 545: return __builtin_ia32_comisdge ((__v2df)__A, (__v2df)__B);
-: 546:}
-: 547:
-: 548:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 549:_mm_comineq_sd (__m128d __A, __m128d __B)
-: 550:{
-: 551: return __builtin_ia32_comisdneq ((__v2df)__A, (__v2df)__B);
-: 552:}
-: 553:
-: 554:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 555:_mm_ucomieq_sd (__m128d __A, __m128d __B)
-: 556:{
-: 557: return __builtin_ia32_ucomisdeq ((__v2df)__A, (__v2df)__B);
-: 558:}
-: 559:
-: 560:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 561:_mm_ucomilt_sd (__m128d __A, __m128d __B)
-: 562:{
-: 563: return __builtin_ia32_ucomisdlt ((__v2df)__A, (__v2df)__B);
-: 564:}
-: 565:
-: 566:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 567:_mm_ucomile_sd (__m128d __A, __m128d __B)
-: 568:{
-: 569: return __builtin_ia32_ucomisdle ((__v2df)__A, (__v2df)__B);
-: 570:}
-: 571:
-: 572:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 573:_mm_ucomigt_sd (__m128d __A, __m128d __B)
-: 574:{
-: 575: return __builtin_ia32_ucomisdgt ((__v2df)__A, (__v2df)__B);
-: 576:}
-: 577:
-: 578:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 579:_mm_ucomige_sd (__m128d __A, __m128d __B)
-: 580:{
-: 581: return __builtin_ia32_ucomisdge ((__v2df)__A, (__v2df)__B);
-: 582:}
-: 583:
-: 584:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 585:_mm_ucomineq_sd (__m128d __A, __m128d __B)
-: 586:{
-: 587: return __builtin_ia32_ucomisdneq ((__v2df)__A, (__v2df)__B);
-: 588:}
-: 589:
-: 590:/* Create a vector of Qi, where i is the element number. */
-: 591:
-: 592:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 593:_mm_set_epi64x (long long __q1, long long __q0)
-: 594:{
-: 595: return __extension__ (__m128i)(__v2di){ __q0, __q1 };
-: 596:}
-: 597:
-: 598:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 599:_mm_set_epi64 (__m64 __q1, __m64 __q0)
-: 600:{
-: 601: return _mm_set_epi64x ((long long)__q1, (long long)__q0);
-: 602:}
-: 603:
-: 604:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 605:_mm_set_epi32 (int __q3, int __q2, int __q1, int __q0)
-: 606:{
-: 607: return __extension__ (__m128i)(__v4si){ __q0, __q1, __q2, __q3 };
-: 608:}
-: 609:
-: 610:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 611:_mm_set_epi16 (short __q7, short __q6, short __q5, short __q4,
-: 612: short __q3, short __q2, short __q1, short __q0)
-: 613:{
-: 614: return __extension__ (__m128i)(__v8hi){
-: 615: __q0, __q1, __q2, __q3, __q4, __q5, __q6, __q7 };
-: 616:}
-: 617:
-: 618:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 619:_mm_set_epi8 (char __q15, char __q14, char __q13, char __q12,
-: 620: char __q11, char __q10, char __q09, char __q08,
-: 621: char __q07, char __q06, char __q05, char __q04,
-: 622: char __q03, char __q02, char __q01, char __q00)
-: 623:{
-: 624: return __extension__ (__m128i)(__v16qi){
-: 625: __q00, __q01, __q02, __q03, __q04, __q05, __q06, __q07,
-: 626: __q08, __q09, __q10, __q11, __q12, __q13, __q14, __q15
-: 627: };
-: 628:}
-: 629:
-: 630:/* Set all of the elements of the vector to A. */
-: 631:
-: 632:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 633:_mm_set1_epi64x (long long __A)
-: 634:{
-: 635: return _mm_set_epi64x (__A, __A);
-: 636:}
-: 637:
-: 638:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 639:_mm_set1_epi64 (__m64 __A)
-: 640:{
-: 641: return _mm_set_epi64 (__A, __A);
-: 642:}
-: 643:
-: 644:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 645:_mm_set1_epi32 (int __A)
-: 646:{
-: 647: return _mm_set_epi32 (__A, __A, __A, __A);
-: 648:}
-: 649:
-: 650:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 651:_mm_set1_epi16 (short __A)
-: 652:{
-: 653: return _mm_set_epi16 (__A, __A, __A, __A, __A, __A, __A, __A);
-: 654:}
-: 655:
-: 656:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 657:_mm_set1_epi8 (char __A)
-: 658:{
-: 659: return _mm_set_epi8 (__A, __A, __A, __A, __A, __A, __A, __A,
-: 660: __A, __A, __A, __A, __A, __A, __A, __A);
-: 661:}
-: 662:
-: 663:/* Create a vector of Qi, where i is the element number.
-: 664: The parameter order is reversed from the _mm_set_epi* functions. */
-: 665:
-: 666:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 667:_mm_setr_epi64 (__m64 __q0, __m64 __q1)
-: 668:{
-: 669: return _mm_set_epi64 (__q1, __q0);
-: 670:}
-: 671:
-: 672:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 673:_mm_setr_epi32 (int __q0, int __q1, int __q2, int __q3)
-: 674:{
-: 675: return _mm_set_epi32 (__q3, __q2, __q1, __q0);
-: 676:}
-: 677:
-: 678:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 679:_mm_setr_epi16 (short __q0, short __q1, short __q2, short __q3,
-: 680: short __q4, short __q5, short __q6, short __q7)
-: 681:{
-: 682: return _mm_set_epi16 (__q7, __q6, __q5, __q4, __q3, __q2, __q1, __q0);
-: 683:}
-: 684:
-: 685:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 686:_mm_setr_epi8 (char __q00, char __q01, char __q02, char __q03,
-: 687: char __q04, char __q05, char __q06, char __q07,
-: 688: char __q08, char __q09, char __q10, char __q11,
-: 689: char __q12, char __q13, char __q14, char __q15)
-: 690:{
-: 691: return _mm_set_epi8 (__q15, __q14, __q13, __q12, __q11, __q10, __q09, __q08,
-: 692: __q07, __q06, __q05, __q04, __q03, __q02, __q01, __q00);
-: 693:}
-: 694:
-: 695:/* Create a vector with element 0 as *P and the rest zero. */
-: 696:
-: 697:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 698:_mm_load_si128 (__m128i const *__P)
-: 699:{
-: 700: return *__P;
-: 701:}
-: 702:
-: 703:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 704:_mm_loadu_si128 (__m128i_u const *__P)
-: 705:{
-: 706: return *__P;
-: 707:}
-: 708:
-: 709:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 710:_mm_loadl_epi64 (__m128i_u const *__P)
-: 711:{
-: 712: return _mm_set_epi64 ((__m64)0LL, *(__m64_u *)__P);
-: 713:}
-: 714:
-: 715:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 716:_mm_loadu_si64 (void const *__P)
-: 717:{
-: 718: return _mm_loadl_epi64 ((__m128i_u *)__P);
-: 719:}
-: 720:
-: 721:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 722:_mm_loadu_si32 (void const *__P)
-: 723:{
-: 724: return _mm_set_epi32 (0, 0, 0, (*(__m32_u *)__P)[0]);
-: 725:}
-: 726:
-: 727:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 728:_mm_loadu_si16 (void const *__P)
-: 729:{
-: 730: return _mm_set_epi16 (0, 0, 0, 0, 0, 0, 0, (*(__m16_u *)__P)[0]);
-: 731:}
-: 732:
-: 733:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 734:_mm_store_si128 (__m128i *__P, __m128i __B)
-: 735:{
-: 736: *__P = __B;
-: 737:}
-: 738:
-: 739:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 740:_mm_storeu_si128 (__m128i_u *__P, __m128i __B)
-: 741:{
-: 742: *__P = __B;
-: 743:}
-: 744:
-: 745:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 746:_mm_storel_epi64 (__m128i_u *__P, __m128i __B)
-: 747:{
-: 748: *(__m64_u *)__P = (__m64) ((__v2di)__B)[0];
-: 749:}
-: 750:
-: 751:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 752:_mm_storeu_si64 (void *__P, __m128i __B)
-: 753:{
-: 754: _mm_storel_epi64 ((__m128i_u *)__P, __B);
-: 755:}
-: 756:
-: 757:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 758:_mm_storeu_si32 (void *__P, __m128i __B)
-: 759:{
-: 760: *(__m32_u *)__P = (__m32) ((__v4si)__B)[0];
-: 761:}
-: 762:
-: 763:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 764:_mm_storeu_si16 (void *__P, __m128i __B)
-: 765:{
-: 766: *(__m16_u *)__P = (__m16) ((__v8hi)__B)[0];
-: 767:}
-: 768:
-: 769:extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 770:_mm_movepi64_pi64 (__m128i __B)
-: 771:{
-: 772: return (__m64) ((__v2di)__B)[0];
-: 773:}
-: 774:
-: 775:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 776:_mm_movpi64_epi64 (__m64 __A)
-: 777:{
-: 778: return _mm_set_epi64 ((__m64)0LL, __A);
-: 779:}
-: 780:
-: 781:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 782:_mm_move_epi64 (__m128i __A)
-: 783:{
-: 784: return (__m128i)__builtin_ia32_movq128 ((__v2di) __A);
-: 785:}
-: 786:
-: 787:/* Create an undefined vector. */
-: 788:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 789:_mm_undefined_si128 (void)
-: 790:{
-: 791:#pragma GCC diagnostic push
-: 792:#pragma GCC diagnostic ignored "-Winit-self"
-: 793: __m128i __Y = __Y;
-: 794:#pragma GCC diagnostic pop
-: 795: return __Y;
-: 796:}
-: 797:
-: 798:/* Create a vector of zeros. */
-: 799:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 800:_mm_setzero_si128 (void)
-: 801:{
-: 802: return __extension__ (__m128i)(__v4si){ 0, 0, 0, 0 };
-: 803:}
-: 804:
-: 805:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 806:_mm_cvtepi32_pd (__m128i __A)
-: 807:{
-: 808: return (__m128d)__builtin_ia32_cvtdq2pd ((__v4si) __A);
-: 809:}
-: 810:
-: 811:extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 812:_mm_cvtepi32_ps (__m128i __A)
-: 813:{
-: 814: return (__m128)__builtin_ia32_cvtdq2ps ((__v4si) __A);
-: 815:}
-: 816:
-: 817:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 818:_mm_cvtpd_epi32 (__m128d __A)
-: 819:{
-: 820: return (__m128i)__builtin_ia32_cvtpd2dq ((__v2df) __A);
-: 821:}
-: 822:
-: 823:extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 824:_mm_cvtpd_pi32 (__m128d __A)
-: 825:{
-: 826: return (__m64)__builtin_ia32_cvtpd2pi ((__v2df) __A);
-: 827:}
-: 828:
-: 829:extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 830:_mm_cvtpd_ps (__m128d __A)
-: 831:{
-: 832: return (__m128)__builtin_ia32_cvtpd2ps ((__v2df) __A);
-: 833:}
-: 834:
-: 835:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 836:_mm_cvttpd_epi32 (__m128d __A)
-: 837:{
-: 838: return (__m128i)__builtin_ia32_cvttpd2dq ((__v2df) __A);
-: 839:}
-: 840:
-: 841:extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 842:_mm_cvttpd_pi32 (__m128d __A)
-: 843:{
-: 844: return (__m64)__builtin_ia32_cvttpd2pi ((__v2df) __A);
-: 845:}
-: 846:
-: 847:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 848:_mm_cvtpi32_pd (__m64 __A)
-: 849:{
-: 850: return (__m128d)__builtin_ia32_cvtpi2pd ((__v2si) __A);
-: 851:}
-: 852:
-: 853:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 854:_mm_cvtps_epi32 (__m128 __A)
-: 855:{
-: 856: return (__m128i)__builtin_ia32_cvtps2dq ((__v4sf) __A);
-: 857:}
-: 858:
-: 859:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 860:_mm_cvttps_epi32 (__m128 __A)
-: 861:{
-: 862: return (__m128i)__builtin_ia32_cvttps2dq ((__v4sf) __A);
-: 863:}
-: 864:
-: 865:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 866:_mm_cvtps_pd (__m128 __A)
-: 867:{
-: 868: return (__m128d)__builtin_ia32_cvtps2pd ((__v4sf) __A);
-: 869:}
-: 870:
-: 871:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 872:_mm_cvtsd_si32 (__m128d __A)
-: 873:{
-: 874: return __builtin_ia32_cvtsd2si ((__v2df) __A);
-: 875:}
-: 876:
-: 877:#ifdef __x86_64__
-: 878:/* Intel intrinsic. */
-: 879:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 880:_mm_cvtsd_si64 (__m128d __A)
-: 881:{
-: 882: return __builtin_ia32_cvtsd2si64 ((__v2df) __A);
-: 883:}
-: 884:
-: 885:/* Microsoft intrinsic. */
-: 886:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 887:_mm_cvtsd_si64x (__m128d __A)
-: 888:{
-: 889: return __builtin_ia32_cvtsd2si64 ((__v2df) __A);
-: 890:}
-: 891:#endif
-: 892:
-: 893:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 894:_mm_cvttsd_si32 (__m128d __A)
-: 895:{
-: 896: return __builtin_ia32_cvttsd2si ((__v2df) __A);
-: 897:}
-: 898:
-: 899:#ifdef __x86_64__
-: 900:/* Intel intrinsic. */
-: 901:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 902:_mm_cvttsd_si64 (__m128d __A)
-: 903:{
-: 904: return __builtin_ia32_cvttsd2si64 ((__v2df) __A);
-: 905:}
-: 906:
-: 907:/* Microsoft intrinsic. */
-: 908:extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 909:_mm_cvttsd_si64x (__m128d __A)
-: 910:{
-: 911: return __builtin_ia32_cvttsd2si64 ((__v2df) __A);
-: 912:}
-: 913:#endif
-: 914:
-: 915:extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 916:_mm_cvtsd_ss (__m128 __A, __m128d __B)
-: 917:{
-: 918: return (__m128)__builtin_ia32_cvtsd2ss ((__v4sf) __A, (__v2df) __B);
-: 919:}
-: 920:
-: 921:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 922:_mm_cvtsi32_sd (__m128d __A, int __B)
-: 923:{
-: 924: return (__m128d)__builtin_ia32_cvtsi2sd ((__v2df) __A, __B);
-: 925:}
-: 926:
-: 927:#ifdef __x86_64__
-: 928:/* Intel intrinsic. */
-: 929:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 930:_mm_cvtsi64_sd (__m128d __A, long long __B)
-: 931:{
-: 932: return (__m128d)__builtin_ia32_cvtsi642sd ((__v2df) __A, __B);
-: 933:}
-: 934:
-: 935:/* Microsoft intrinsic. */
-: 936:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 937:_mm_cvtsi64x_sd (__m128d __A, long long __B)
-: 938:{
-: 939: return (__m128d)__builtin_ia32_cvtsi642sd ((__v2df) __A, __B);
-: 940:}
-: 941:#endif
-: 942:
-: 943:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 944:_mm_cvtss_sd (__m128d __A, __m128 __B)
-: 945:{
-: 946: return (__m128d)__builtin_ia32_cvtss2sd ((__v2df) __A, (__v4sf)__B);
-: 947:}
-: 948:
-: 949:#ifdef __OPTIMIZE__
-: 950:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 951:_mm_shuffle_pd(__m128d __A, __m128d __B, const int __mask)
-: 952:{
-: 953: return (__m128d)__builtin_ia32_shufpd ((__v2df)__A, (__v2df)__B, __mask);
-: 954:}
-: 955:#else
-: 956:#define _mm_shuffle_pd(A, B, N) \
-: 957: ((__m128d)__builtin_ia32_shufpd ((__v2df)(__m128d)(A), \
-: 958: (__v2df)(__m128d)(B), (int)(N)))
-: 959:#endif
-: 960:
-: 961:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 962:_mm_unpackhi_pd (__m128d __A, __m128d __B)
-: 963:{
-: 964: return (__m128d)__builtin_ia32_unpckhpd ((__v2df)__A, (__v2df)__B);
-: 965:}
-: 966:
-: 967:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 968:_mm_unpacklo_pd (__m128d __A, __m128d __B)
-: 969:{
-: 970: return (__m128d)__builtin_ia32_unpcklpd ((__v2df)__A, (__v2df)__B);
-: 971:}
-: 972:
-: 973:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 974:_mm_loadh_pd (__m128d __A, double const *__B)
-: 975:{
-: 976: return (__m128d)__builtin_ia32_loadhpd ((__v2df)__A, __B);
-: 977:}
-: 978:
-: 979:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 980:_mm_loadl_pd (__m128d __A, double const *__B)
-: 981:{
-: 982: return (__m128d)__builtin_ia32_loadlpd ((__v2df)__A, __B);
-: 983:}
-: 984:
-: 985:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 986:_mm_movemask_pd (__m128d __A)
-: 987:{
-: 988: return __builtin_ia32_movmskpd ((__v2df)__A);
-: 989:}
-: 990:
-: 991:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 992:_mm_packs_epi16 (__m128i __A, __m128i __B)
-: 993:{
-: 994: return (__m128i)__builtin_ia32_packsswb128 ((__v8hi)__A, (__v8hi)__B);
-: 995:}
-: 996:
-: 997:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 998:_mm_packs_epi32 (__m128i __A, __m128i __B)
-: 999:{
-: 1000: return (__m128i)__builtin_ia32_packssdw128 ((__v4si)__A, (__v4si)__B);
-: 1001:}
-: 1002:
-: 1003:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1004:_mm_packus_epi16 (__m128i __A, __m128i __B)
-: 1005:{
-: 1006: return (__m128i)__builtin_ia32_packuswb128 ((__v8hi)__A, (__v8hi)__B);
-: 1007:}
-: 1008:
-: 1009:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1010:_mm_unpackhi_epi8 (__m128i __A, __m128i __B)
-: 1011:{
-: 1012: return (__m128i)__builtin_ia32_punpckhbw128 ((__v16qi)__A, (__v16qi)__B);
-: 1013:}
-: 1014:
-: 1015:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1016:_mm_unpackhi_epi16 (__m128i __A, __m128i __B)
-: 1017:{
-: 1018: return (__m128i)__builtin_ia32_punpckhwd128 ((__v8hi)__A, (__v8hi)__B);
-: 1019:}
-: 1020:
-: 1021:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1022:_mm_unpackhi_epi32 (__m128i __A, __m128i __B)
-: 1023:{
-: 1024: return (__m128i)__builtin_ia32_punpckhdq128 ((__v4si)__A, (__v4si)__B);
-: 1025:}
-: 1026:
-: 1027:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1028:_mm_unpackhi_epi64 (__m128i __A, __m128i __B)
-: 1029:{
-: 1030: return (__m128i)__builtin_ia32_punpckhqdq128 ((__v2di)__A, (__v2di)__B);
-: 1031:}
-: 1032:
-: 1033:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1034:_mm_unpacklo_epi8 (__m128i __A, __m128i __B)
-: 1035:{
-: 1036: return (__m128i)__builtin_ia32_punpcklbw128 ((__v16qi)__A, (__v16qi)__B);
-: 1037:}
-: 1038:
-: 1039:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1040:_mm_unpacklo_epi16 (__m128i __A, __m128i __B)
-: 1041:{
-: 1042: return (__m128i)__builtin_ia32_punpcklwd128 ((__v8hi)__A, (__v8hi)__B);
-: 1043:}
-: 1044:
-: 1045:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1046:_mm_unpacklo_epi32 (__m128i __A, __m128i __B)
-: 1047:{
-: 1048: return (__m128i)__builtin_ia32_punpckldq128 ((__v4si)__A, (__v4si)__B);
-: 1049:}
-: 1050:
-: 1051:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1052:_mm_unpacklo_epi64 (__m128i __A, __m128i __B)
-: 1053:{
-: 1054: return (__m128i)__builtin_ia32_punpcklqdq128 ((__v2di)__A, (__v2di)__B);
-: 1055:}
-: 1056:
-: 1057:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1058:_mm_add_epi8 (__m128i __A, __m128i __B)
-: 1059:{
-: 1060: return (__m128i) ((__v16qu)__A + (__v16qu)__B);
-: 1061:}
-: 1062:
-: 1063:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1064:_mm_add_epi16 (__m128i __A, __m128i __B)
-: 1065:{
-: 1066: return (__m128i) ((__v8hu)__A + (__v8hu)__B);
-: 1067:}
-: 1068:
-: 1069:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1070:_mm_add_epi32 (__m128i __A, __m128i __B)
-: 1071:{
-: 1072: return (__m128i) ((__v4su)__A + (__v4su)__B);
-: 1073:}
-: 1074:
-: 1075:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1076:_mm_add_epi64 (__m128i __A, __m128i __B)
-: 1077:{
-: 1078: return (__m128i) ((__v2du)__A + (__v2du)__B);
-: 1079:}
-: 1080:
-: 1081:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1082:_mm_adds_epi8 (__m128i __A, __m128i __B)
-: 1083:{
-: 1084: return (__m128i)__builtin_ia32_paddsb128 ((__v16qi)__A, (__v16qi)__B);
-: 1085:}
-: 1086:
-: 1087:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1088:_mm_adds_epi16 (__m128i __A, __m128i __B)
-: 1089:{
-: 1090: return (__m128i)__builtin_ia32_paddsw128 ((__v8hi)__A, (__v8hi)__B);
-: 1091:}
-: 1092:
-: 1093:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1094:_mm_adds_epu8 (__m128i __A, __m128i __B)
-: 1095:{
-: 1096: return (__m128i)__builtin_ia32_paddusb128 ((__v16qi)__A, (__v16qi)__B);
-: 1097:}
-: 1098:
-: 1099:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1100:_mm_adds_epu16 (__m128i __A, __m128i __B)
-: 1101:{
-: 1102: return (__m128i)__builtin_ia32_paddusw128 ((__v8hi)__A, (__v8hi)__B);
-: 1103:}
-: 1104:
-: 1105:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1106:_mm_sub_epi8 (__m128i __A, __m128i __B)
-: 1107:{
-: 1108: return (__m128i) ((__v16qu)__A - (__v16qu)__B);
-: 1109:}
-: 1110:
-: 1111:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1112:_mm_sub_epi16 (__m128i __A, __m128i __B)
-: 1113:{
-: 1114: return (__m128i) ((__v8hu)__A - (__v8hu)__B);
-: 1115:}
-: 1116:
-: 1117:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1118:_mm_sub_epi32 (__m128i __A, __m128i __B)
-: 1119:{
-: 1120: return (__m128i) ((__v4su)__A - (__v4su)__B);
-: 1121:}
-: 1122:
-: 1123:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1124:_mm_sub_epi64 (__m128i __A, __m128i __B)
-: 1125:{
-: 1126: return (__m128i) ((__v2du)__A - (__v2du)__B);
-: 1127:}
-: 1128:
-: 1129:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1130:_mm_subs_epi8 (__m128i __A, __m128i __B)
-: 1131:{
-: 1132: return (__m128i)__builtin_ia32_psubsb128 ((__v16qi)__A, (__v16qi)__B);
-: 1133:}
-: 1134:
-: 1135:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1136:_mm_subs_epi16 (__m128i __A, __m128i __B)
-: 1137:{
-: 1138: return (__m128i)__builtin_ia32_psubsw128 ((__v8hi)__A, (__v8hi)__B);
-: 1139:}
-: 1140:
-: 1141:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1142:_mm_subs_epu8 (__m128i __A, __m128i __B)
-: 1143:{
-: 1144: return (__m128i)__builtin_ia32_psubusb128 ((__v16qi)__A, (__v16qi)__B);
-: 1145:}
-: 1146:
-: 1147:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1148:_mm_subs_epu16 (__m128i __A, __m128i __B)
-: 1149:{
-: 1150: return (__m128i)__builtin_ia32_psubusw128 ((__v8hi)__A, (__v8hi)__B);
-: 1151:}
-: 1152:
-: 1153:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1154:_mm_madd_epi16 (__m128i __A, __m128i __B)
-: 1155:{
-: 1156: return (__m128i)__builtin_ia32_pmaddwd128 ((__v8hi)__A, (__v8hi)__B);
-: 1157:}
-: 1158:
-: 1159:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1160:_mm_mulhi_epi16 (__m128i __A, __m128i __B)
-: 1161:{
-: 1162: return (__m128i)__builtin_ia32_pmulhw128 ((__v8hi)__A, (__v8hi)__B);
-: 1163:}
-: 1164:
-: 1165:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1166:_mm_mullo_epi16 (__m128i __A, __m128i __B)
-: 1167:{
-: 1168: return (__m128i) ((__v8hu)__A * (__v8hu)__B);
-: 1169:}
-: 1170:
-: 1171:extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1172:_mm_mul_su32 (__m64 __A, __m64 __B)
-: 1173:{
-: 1174: return (__m64)__builtin_ia32_pmuludq ((__v2si)__A, (__v2si)__B);
-: 1175:}
-: 1176:
-: 1177:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1178:_mm_mul_epu32 (__m128i __A, __m128i __B)
-: 1179:{
-: 1180: return (__m128i)__builtin_ia32_pmuludq128 ((__v4si)__A, (__v4si)__B);
-: 1181:}
-: 1182:
-: 1183:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1184:_mm_slli_epi16 (__m128i __A, int __B)
-: 1185:{
-: 1186: return (__m128i)__builtin_ia32_psllwi128 ((__v8hi)__A, __B);
-: 1187:}
-: 1188:
-: 1189:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1190:_mm_slli_epi32 (__m128i __A, int __B)
-: 1191:{
-: 1192: return (__m128i)__builtin_ia32_pslldi128 ((__v4si)__A, __B);
-: 1193:}
-: 1194:
-: 1195:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1196:_mm_slli_epi64 (__m128i __A, int __B)
-: 1197:{
-: 1198: return (__m128i)__builtin_ia32_psllqi128 ((__v2di)__A, __B);
-: 1199:}
-: 1200:
-: 1201:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1202:_mm_srai_epi16 (__m128i __A, int __B)
-: 1203:{
-: 1204: return (__m128i)__builtin_ia32_psrawi128 ((__v8hi)__A, __B);
-: 1205:}
-: 1206:
-: 1207:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1208:_mm_srai_epi32 (__m128i __A, int __B)
-: 1209:{
-: 1210: return (__m128i)__builtin_ia32_psradi128 ((__v4si)__A, __B);
-: 1211:}
-: 1212:
-: 1213:#ifdef __OPTIMIZE__
-: 1214:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1215:_mm_bsrli_si128 (__m128i __A, const int __N)
-: 1216:{
-: 1217: return (__m128i)__builtin_ia32_psrldqi128 (__A, __N * 8);
-: 1218:}
-: 1219:
-: 1220:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1221:_mm_bslli_si128 (__m128i __A, const int __N)
-: 1222:{
-: 1223: return (__m128i)__builtin_ia32_pslldqi128 (__A, __N * 8);
-: 1224:}
-: 1225:
-: 1226:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1227:_mm_srli_si128 (__m128i __A, const int __N)
-: 1228:{
-: 1229: return (__m128i)__builtin_ia32_psrldqi128 (__A, __N * 8);
-: 1230:}
-: 1231:
-: 1232:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1233:_mm_slli_si128 (__m128i __A, const int __N)
-: 1234:{
-: 1235: return (__m128i)__builtin_ia32_pslldqi128 (__A, __N * 8);
-: 1236:}
-: 1237:#else
-: 1238:#define _mm_bsrli_si128(A, N) \
-: 1239: ((__m128i)__builtin_ia32_psrldqi128 ((__m128i)(A), (int)(N) * 8))
-: 1240:#define _mm_bslli_si128(A, N) \
-: 1241: ((__m128i)__builtin_ia32_pslldqi128 ((__m128i)(A), (int)(N) * 8))
-: 1242:#define _mm_srli_si128(A, N) \
-: 1243: ((__m128i)__builtin_ia32_psrldqi128 ((__m128i)(A), (int)(N) * 8))
-: 1244:#define _mm_slli_si128(A, N) \
-: 1245: ((__m128i)__builtin_ia32_pslldqi128 ((__m128i)(A), (int)(N) * 8))
-: 1246:#endif
-: 1247:
-: 1248:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1249:_mm_srli_epi16 (__m128i __A, int __B)
-: 1250:{
-: 1251: return (__m128i)__builtin_ia32_psrlwi128 ((__v8hi)__A, __B);
-: 1252:}
-: 1253:
-: 1254:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1255:_mm_srli_epi32 (__m128i __A, int __B)
-: 1256:{
-: 1257: return (__m128i)__builtin_ia32_psrldi128 ((__v4si)__A, __B);
-: 1258:}
-: 1259:
-: 1260:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1261:_mm_srli_epi64 (__m128i __A, int __B)
-: 1262:{
-: 1263: return (__m128i)__builtin_ia32_psrlqi128 ((__v2di)__A, __B);
-: 1264:}
-: 1265:
-: 1266:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1267:_mm_sll_epi16 (__m128i __A, __m128i __B)
-: 1268:{
-: 1269: return (__m128i)__builtin_ia32_psllw128((__v8hi)__A, (__v8hi)__B);
-: 1270:}
-: 1271:
-: 1272:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1273:_mm_sll_epi32 (__m128i __A, __m128i __B)
-: 1274:{
-: 1275: return (__m128i)__builtin_ia32_pslld128((__v4si)__A, (__v4si)__B);
-: 1276:}
-: 1277:
-: 1278:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1279:_mm_sll_epi64 (__m128i __A, __m128i __B)
-: 1280:{
-: 1281: return (__m128i)__builtin_ia32_psllq128((__v2di)__A, (__v2di)__B);
-: 1282:}
-: 1283:
-: 1284:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1285:_mm_sra_epi16 (__m128i __A, __m128i __B)
-: 1286:{
-: 1287: return (__m128i)__builtin_ia32_psraw128 ((__v8hi)__A, (__v8hi)__B);
-: 1288:}
-: 1289:
-: 1290:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1291:_mm_sra_epi32 (__m128i __A, __m128i __B)
-: 1292:{
-: 1293: return (__m128i)__builtin_ia32_psrad128 ((__v4si)__A, (__v4si)__B);
-: 1294:}
-: 1295:
-: 1296:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1297:_mm_srl_epi16 (__m128i __A, __m128i __B)
-: 1298:{
-: 1299: return (__m128i)__builtin_ia32_psrlw128 ((__v8hi)__A, (__v8hi)__B);
-: 1300:}
-: 1301:
-: 1302:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1303:_mm_srl_epi32 (__m128i __A, __m128i __B)
-: 1304:{
-: 1305: return (__m128i)__builtin_ia32_psrld128 ((__v4si)__A, (__v4si)__B);
-: 1306:}
-: 1307:
-: 1308:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1309:_mm_srl_epi64 (__m128i __A, __m128i __B)
-: 1310:{
-: 1311: return (__m128i)__builtin_ia32_psrlq128 ((__v2di)__A, (__v2di)__B);
-: 1312:}
-: 1313:
-: 1314:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1315:_mm_and_si128 (__m128i __A, __m128i __B)
-: 1316:{
-: 1317: return (__m128i) ((__v2du)__A & (__v2du)__B);
-: 1318:}
-: 1319:
-: 1320:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1321:_mm_andnot_si128 (__m128i __A, __m128i __B)
-: 1322:{
-: 1323: return (__m128i)__builtin_ia32_pandn128 ((__v2di)__A, (__v2di)__B);
-: 1324:}
-: 1325:
-: 1326:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1327:_mm_or_si128 (__m128i __A, __m128i __B)
-: 1328:{
-: 1329: return (__m128i) ((__v2du)__A | (__v2du)__B);
-: 1330:}
-: 1331:
-: 1332:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1333:_mm_xor_si128 (__m128i __A, __m128i __B)
-: 1334:{
-: 1335: return (__m128i) ((__v2du)__A ^ (__v2du)__B);
-: 1336:}
-: 1337:
-: 1338:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1339:_mm_cmpeq_epi8 (__m128i __A, __m128i __B)
-: 1340:{
-: 1341: return (__m128i) ((__v16qi)__A == (__v16qi)__B);
-: 1342:}
-: 1343:
-: 1344:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1345:_mm_cmpeq_epi16 (__m128i __A, __m128i __B)
-: 1346:{
-: 1347: return (__m128i) ((__v8hi)__A == (__v8hi)__B);
-: 1348:}
-: 1349:
-: 1350:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1351:_mm_cmpeq_epi32 (__m128i __A, __m128i __B)
-: 1352:{
-: 1353: return (__m128i) ((__v4si)__A == (__v4si)__B);
-: 1354:}
-: 1355:
-: 1356:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1357:_mm_cmplt_epi8 (__m128i __A, __m128i __B)
-: 1358:{
-: 1359: return (__m128i) ((__v16qs)__A < (__v16qs)__B);
-: 1360:}
-: 1361:
-: 1362:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1363:_mm_cmplt_epi16 (__m128i __A, __m128i __B)
-: 1364:{
-: 1365: return (__m128i) ((__v8hi)__A < (__v8hi)__B);
-: 1366:}
-: 1367:
-: 1368:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1369:_mm_cmplt_epi32 (__m128i __A, __m128i __B)
-: 1370:{
-: 1371: return (__m128i) ((__v4si)__A < (__v4si)__B);
-: 1372:}
-: 1373:
-: 1374:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1375:_mm_cmpgt_epi8 (__m128i __A, __m128i __B)
-: 1376:{
-: 1377: return (__m128i) ((__v16qs)__A > (__v16qs)__B);
-: 1378:}
-: 1379:
-: 1380:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1381:_mm_cmpgt_epi16 (__m128i __A, __m128i __B)
-: 1382:{
-: 1383: return (__m128i) ((__v8hi)__A > (__v8hi)__B);
-: 1384:}
-: 1385:
-: 1386:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1387:_mm_cmpgt_epi32 (__m128i __A, __m128i __B)
-: 1388:{
-: 1389: return (__m128i) ((__v4si)__A > (__v4si)__B);
-: 1390:}
-: 1391:
-: 1392:#ifdef __OPTIMIZE__
-: 1393:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1394:_mm_extract_epi16 (__m128i const __A, int const __N)
-: 1395:{
-: 1396: return (unsigned short) __builtin_ia32_vec_ext_v8hi ((__v8hi)__A, __N);
-: 1397:}
-: 1398:
-: 1399:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1400:_mm_insert_epi16 (__m128i const __A, int const __D, int const __N)
-: 1401:{
-: 1402: return (__m128i) __builtin_ia32_vec_set_v8hi ((__v8hi)__A, __D, __N);
-: 1403:}
-: 1404:#else
-: 1405:#define _mm_extract_epi16(A, N) \
-: 1406: ((int) (unsigned short) __builtin_ia32_vec_ext_v8hi ((__v8hi)(__m128i)(A), (int)(N)))
-: 1407:#define _mm_insert_epi16(A, D, N) \
-: 1408: ((__m128i) __builtin_ia32_vec_set_v8hi ((__v8hi)(__m128i)(A), \
-: 1409: (int)(D), (int)(N)))
-: 1410:#endif
-: 1411:
-: 1412:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1413:_mm_max_epi16 (__m128i __A, __m128i __B)
-: 1414:{
-: 1415: return (__m128i)__builtin_ia32_pmaxsw128 ((__v8hi)__A, (__v8hi)__B);
-: 1416:}
-: 1417:
-: 1418:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1419:_mm_max_epu8 (__m128i __A, __m128i __B)
-: 1420:{
-: 1421: return (__m128i)__builtin_ia32_pmaxub128 ((__v16qi)__A, (__v16qi)__B);
-: 1422:}
-: 1423:
-: 1424:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1425:_mm_min_epi16 (__m128i __A, __m128i __B)
-: 1426:{
-: 1427: return (__m128i)__builtin_ia32_pminsw128 ((__v8hi)__A, (__v8hi)__B);
-: 1428:}
-: 1429:
-: 1430:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1431:_mm_min_epu8 (__m128i __A, __m128i __B)
-: 1432:{
-: 1433: return (__m128i)__builtin_ia32_pminub128 ((__v16qi)__A, (__v16qi)__B);
-: 1434:}
-: 1435:
-: 1436:extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1437:_mm_movemask_epi8 (__m128i __A)
-: 1438:{
-: 1439: return __builtin_ia32_pmovmskb128 ((__v16qi)__A);
-: 1440:}
-: 1441:
-: 1442:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1443:_mm_mulhi_epu16 (__m128i __A, __m128i __B)
-: 1444:{
-: 1445: return (__m128i)__builtin_ia32_pmulhuw128 ((__v8hi)__A, (__v8hi)__B);
-: 1446:}
-: 1447:
-: 1448:#ifdef __OPTIMIZE__
-: 1449:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1450:_mm_shufflehi_epi16 (__m128i __A, const int __mask)
-: 1451:{
-: 1452: return (__m128i)__builtin_ia32_pshufhw ((__v8hi)__A, __mask);
-: 1453:}
-: 1454:
-: 1455:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1456:_mm_shufflelo_epi16 (__m128i __A, const int __mask)
-: 1457:{
-: 1458: return (__m128i)__builtin_ia32_pshuflw ((__v8hi)__A, __mask);
-: 1459:}
-: 1460:
-: 1461:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1462:_mm_shuffle_epi32 (__m128i __A, const int __mask)
-: 1463:{
-: 1464: return (__m128i)__builtin_ia32_pshufd ((__v4si)__A, __mask);
-: 1465:}
-: 1466:#else
-: 1467:#define _mm_shufflehi_epi16(A, N) \
-: 1468: ((__m128i)__builtin_ia32_pshufhw ((__v8hi)(__m128i)(A), (int)(N)))
-: 1469:#define _mm_shufflelo_epi16(A, N) \
-: 1470: ((__m128i)__builtin_ia32_pshuflw ((__v8hi)(__m128i)(A), (int)(N)))
-: 1471:#define _mm_shuffle_epi32(A, N) \
-: 1472: ((__m128i)__builtin_ia32_pshufd ((__v4si)(__m128i)(A), (int)(N)))
-: 1473:#endif
-: 1474:
-: 1475:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1476:_mm_maskmoveu_si128 (__m128i __A, __m128i __B, char *__C)
-: 1477:{
-: 1478: __builtin_ia32_maskmovdqu ((__v16qi)__A, (__v16qi)__B, __C);
-: 1479:}
-: 1480:
-: 1481:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1482:_mm_avg_epu8 (__m128i __A, __m128i __B)
-: 1483:{
-: 1484: return (__m128i)__builtin_ia32_pavgb128 ((__v16qi)__A, (__v16qi)__B);
-: 1485:}
-: 1486:
-: 1487:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1488:_mm_avg_epu16 (__m128i __A, __m128i __B)
-: 1489:{
-: 1490: return (__m128i)__builtin_ia32_pavgw128 ((__v8hi)__A, (__v8hi)__B);
-: 1491:}
-: 1492:
-: 1493:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1494:_mm_sad_epu8 (__m128i __A, __m128i __B)
-: 1495:{
-: 1496: return (__m128i)__builtin_ia32_psadbw128 ((__v16qi)__A, (__v16qi)__B);
-: 1497:}
-: 1498:
-: 1499:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1500:_mm_stream_si32 (int *__A, int __B)
-: 1501:{
-: 1502: __builtin_ia32_movnti (__A, __B);
-: 1503:}
-: 1504:
-: 1505:#ifdef __x86_64__
-: 1506:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1507:_mm_stream_si64 (long long int *__A, long long int __B)
-: 1508:{
-: 1509: __builtin_ia32_movnti64 (__A, __B);
-: 1510:}
-: 1511:#endif
-: 1512:
-: 1513:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1514:_mm_stream_si128 (__m128i *__A, __m128i __B)
-: 1515:{
-: 1516: __builtin_ia32_movntdq ((__v2di *)__A, (__v2di)__B);
-: 1517:}
-: 1518:
-: 1519:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1520:_mm_stream_pd (double *__A, __m128d __B)
-: 1521:{
-: 1522: __builtin_ia32_movntpd (__A, (__v2df)__B);
-: 1523:}
-: 1524:
-: 1525:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1526:_mm_clflush (void const *__A)
-: 1527:{
-: 1528: __builtin_ia32_clflush (__A);
-: 1529:}
-: 1530:
-: 1531:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1532:_mm_lfence (void)
-: 1533:{
-: 1534: __builtin_ia32_lfence ();
-: 1535:}
-: 1536:
-: 1537:extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1538:_mm_mfence (void)
-: 1539:{
-: 1540: __builtin_ia32_mfence ();
-: 1541:}
-: 1542:
-: 1543:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1544:_mm_cvtsi32_si128 (int __A)
-: 1545:{
-: 1546: return _mm_set_epi32 (0, 0, 0, __A);
-: 1547:}
-: 1548:
-: 1549:#ifdef __x86_64__
-: 1550:/* Intel intrinsic. */
-: 1551:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1552:_mm_cvtsi64_si128 (long long __A)
-: 1553:{
-: 1554: return _mm_set_epi64x (0, __A);
-: 1555:}
-: 1556:
-: 1557:/* Microsoft intrinsic. */
-: 1558:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1559:_mm_cvtsi64x_si128 (long long __A)
-: 1560:{
-: 1561: return _mm_set_epi64x (0, __A);
-: 1562:}
-: 1563:#endif
-: 1564:
-: 1565:/* Casts between various SP, DP, INT vector types. Note that these do no
-: 1566: conversion of values, they just change the type. */
-: 1567:extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1568:_mm_castpd_ps(__m128d __A)
-: 1569:{
-: 1570: return (__m128) __A;
-: 1571:}
-: 1572:
-: 1573:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1574:_mm_castpd_si128(__m128d __A)
-: 1575:{
-: 1576: return (__m128i) __A;
-: 1577:}
-: 1578:
-: 1579:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1580:_mm_castps_pd(__m128 __A)
-: 1581:{
-: 1582: return (__m128d) __A;
-: 1583:}
-: 1584:
-: 1585:extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1586:_mm_castps_si128(__m128 __A)
-: 1587:{
-: 1588: return (__m128i) __A;
-: 1589:}
-: 1590:
-: 1591:extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1592:_mm_castsi128_ps(__m128i __A)
-: 1593:{
-: 1594: return (__m128) __A;
-: 1595:}
-: 1596:
-: 1597:extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-: 1598:_mm_castsi128_pd(__m128i __A)
-: 1599:{
-: 1600: return (__m128d) __A;
-: 1601:}
-: 1602:
-: 1603:#ifdef __DISABLE_SSE2__
-: 1604:#undef __DISABLE_SSE2__
-: 1605:#pragma GCC pop_options
-: 1606:#endif /* __DISABLE_SSE2__ */
-: 1607:
-: 1608:#endif /* _EMMINTRIN_H_INCLUDED */