11#error "Never use <avx2intrin.h> directly; include <immintrin.h> instead."
18#define __DEFAULT_FN_ATTRS256 \
19 __attribute__((__always_inline__, __nodebug__, __target__("avx2"), \
20 __min_vector_width__(256)))
21#define __DEFAULT_FN_ATTRS128 \
22 __attribute__((__always_inline__, __nodebug__, __target__("avx2"), \
23 __min_vector_width__(128)))
25#if defined(__cplusplus) && (__cplusplus >= 201103L)
26#define __DEFAULT_FN_ATTRS256_CONSTEXPR __DEFAULT_FN_ATTRS256 constexpr
27#define __DEFAULT_FN_ATTRS128_CONSTEXPR __DEFAULT_FN_ATTRS128 constexpr
29#define __DEFAULT_FN_ATTRS256_CONSTEXPR __DEFAULT_FN_ATTRS256
30#define __DEFAULT_FN_ATTRS128_CONSTEXPR __DEFAULT_FN_ATTRS128
91#define _mm256_mpsadbw_epu8(X, Y, M) \
92 ((__m256i)__builtin_ia32_mpsadbw256((__v32qi)(__m256i)(X), \
93 (__v32qi)(__m256i)(Y), (int)(M)))
108 return (__m256i)__builtin_elementwise_abs((__v32qs)
__a);
124 return (__m256i)__builtin_elementwise_abs((__v16hi)
__a);
140 return (__m256i)__builtin_elementwise_abs((__v8si)
__a);
171 return (__m256i)__builtin_ia32_packsswb256((__v16hi)
__a, (__v16hi)
__b);
203 return (__m256i)__builtin_ia32_packssdw256((__v8si)
__a, (__v8si)
__b);
234 return (__m256i)__builtin_ia32_packuswb256((__v16hi)
__a, (__v16hi)
__b);
266 return (__m256i) __builtin_ia32_packusdw256((__v8si)__V1, (__v8si)__V2);
285 return (__m256i)((__v32qu)
__a + (__v32qu)
__b);
304 return (__m256i)((__v16hu)
__a + (__v16hu)
__b);
323 return (__m256i)((__v8su)
__a + (__v8su)
__b);
342 return (__m256i)((__v4du)
__a + (__v4du)
__b);
360 return (__m256i)__builtin_elementwise_add_sat((__v32qs)
__a, (__v32qs)
__b);
377 return (__m256i)__builtin_elementwise_add_sat((__v16hi)
__a, (__v16hi)
__b);
395 return (__m256i)__builtin_elementwise_add_sat((__v32qu)
__a, (__v32qu)
__b);
412 return (__m256i)__builtin_elementwise_add_sat((__v16hu)
__a, (__v16hu)
__b);
439#define _mm256_alignr_epi8(a, b, n) \
440 ((__m256i)__builtin_ia32_palignr256((__v32qi)(__m256i)(a), \
441 (__v32qi)(__m256i)(b), (n)))
458 return (__m256i)((__v4du)
__a & (__v4du)
__b);
476 return (__m256i)(~(__v4du)
__a & (__v4du)
__b);
501 return (__m256i)__builtin_ia32_pavgb256((__v32qu)
__a, (__v32qu)
__b);
526 return (__m256i)__builtin_ia32_pavgw256((__v16hu)
__a, (__v16hu)
__b);
560 return (__m256i)__builtin_ia32_pblendvb256((__v32qi)__V1, (__v32qi)__V2,
601#define _mm256_blend_epi16(V1, V2, M) \
602 ((__m256i)__builtin_ia32_pblendw256((__v16hi)(__m256i)(V1), \
603 (__v16hi)(__m256i)(V2), (int)(M)))
628 return (__m256i)((__v32qi)
__a == (__v32qi)
__b);
654 return (__m256i)((__v16hi)
__a == (__v16hi)
__b);
680 return (__m256i)((__v8si)
__a == (__v8si)
__b);
706 return (__m256i)((__v4di)
__a == (__v4di)
__b);
734 return (__m256i)((__v32qs)
__a > (__v32qs)
__b);
760 return (__m256i)((__v16hi)
__a > (__v16hi)
__b);
786 return (__m256i)((__v8si)
__a > (__v8si)
__b);
812 return (__m256i)((__v4di)
__a > (__v4di)
__b);
848 return (__m256i)__builtin_ia32_phaddw256((__v16hi)
__a, (__v16hi)
__b);
880 return (__m256i)__builtin_ia32_phaddd256((__v8si)
__a, (__v8si)
__b);
915 return (__m256i)__builtin_ia32_phaddsw256((__v16hi)
__a, (__v16hi)
__b);
951 return (__m256i)__builtin_ia32_phsubw256((__v16hi)
__a, (__v16hi)
__b);
983 return (__m256i)__builtin_ia32_phsubd256((__v8si)
__a, (__v8si)
__b);
1019 return (__m256i)__builtin_ia32_phsubsw256((__v16hi)
__a, (__v16hi)
__b);
1049 return (__m256i)__builtin_ia32_pmaddubsw256((__v32qi)
__a, (__v32qi)
__b);
1081 return (__m256i)__builtin_ia32_pmaddwd256((__v16hi)
__a, (__v16hi)
__b);
1099 return (__m256i)__builtin_elementwise_max((__v32qs)
__a, (__v32qs)
__b);
1117 return (__m256i)__builtin_elementwise_max((__v16hi)
__a, (__v16hi)
__b);
1135 return (__m256i)__builtin_elementwise_max((__v8si)
__a, (__v8si)
__b);
1153 return (__m256i)__builtin_elementwise_max((__v32qu)
__a, (__v32qu)
__b);
1171 return (__m256i)__builtin_elementwise_max((__v16hu)
__a, (__v16hu)
__b);
1189 return (__m256i)__builtin_elementwise_max((__v8su)
__a, (__v8su)
__b);
1207 return (__m256i)__builtin_elementwise_min((__v32qs)
__a, (__v32qs)
__b);
1225 return (__m256i)__builtin_elementwise_min((__v16hi)
__a, (__v16hi)
__b);
1243 return (__m256i)__builtin_elementwise_min((__v8si)
__a, (__v8si)
__b);
1261 return (__m256i)__builtin_elementwise_min((__v32qu)
__a, (__v32qu)
__b);
1279 return (__m256i)__builtin_elementwise_min((__v16hu)
__a, (__v16hu)
__b);
1297 return (__m256i)__builtin_elementwise_min((__v8su)
__a, (__v8su)
__b);
1320 return __builtin_ia32_pmovmskb256((__v32qi)
__a);
1347 return (__m256i)__builtin_convertvector((__v16qs)__V, __v16hi);
1374 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3, 4, 5, 6, 7), __v8si);
1400 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3), __v4di);
1425 return (__m256i)__builtin_convertvector((__v8hi)__V, __v8si);
1449 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v8hi)__V, (__v8hi)__V, 0, 1, 2, 3), __v4di);
1473 return (__m256i)__builtin_convertvector((__v4si)__V, __v4di);
1498 return (__m256i)__builtin_convertvector((__v16qu)__V, __v16hi);
1523 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v16qu)__V, (__v16qu)__V, 0, 1, 2, 3, 4, 5, 6, 7), __v8si);
1547 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v16qu)__V, (__v16qu)__V, 0, 1, 2, 3), __v4di);
1572 return (__m256i)__builtin_convertvector((__v8hu)__V, __v8si);
1596 return (__m256i)__builtin_convertvector(__builtin_shufflevector((__v8hu)__V, (__v8hu)__V, 0, 1, 2, 3), __v4di);
1620 return (__m256i)__builtin_convertvector((__v4su)__V, __v4di);
1645 return (__m256i)__builtin_ia32_pmuldq256((__v8si)
__a, (__v8si)
__b);
1672 return (__m256i)__builtin_ia32_pmulhrsw256((__v16hi)
__a, (__v16hi)
__b);
1691 return (__m256i)__builtin_ia32_pmulhuw256((__v16hu)
__a, (__v16hu)
__b);
1710 return (__m256i)__builtin_ia32_pmulhw256((__v16hi)
__a, (__v16hi)
__b);
1729 return (__m256i)((__v16hu)
__a * (__v16hu)
__b);
1747 return (__m256i)((__v8su)
__a * (__v8su)
__b);
1772 return __builtin_ia32_pmuludq256((__v8si)
__a, (__v8si)
__b);
1790 return (__m256i)((__v4du)
__a | (__v4du)
__b);
1835 return __builtin_ia32_psadbw256((__v32qi)
__a, (__v32qi)
__b);
1874 return (__m256i)__builtin_ia32_pshufb256((__v32qi)
__a, (__v32qi)
__b);
1907#define _mm256_shuffle_epi32(a, imm) \
1908 ((__m256i)__builtin_ia32_pshufd256((__v8si)(__m256i)(a), (int)(imm)))
1943#define _mm256_shufflehi_epi16(a, imm) \
1944 ((__m256i)__builtin_ia32_pshufhw256((__v16hi)(__m256i)(a), (int)(imm)))
1980#define _mm256_shufflelo_epi16(a, imm) \
1981 ((__m256i)__builtin_ia32_pshuflw256((__v16hi)(__m256i)(a), (int)(imm)))
2001 return (__m256i)__builtin_ia32_psignb256((__v32qi)
__a, (__v32qi)
__b);
2022 return (__m256i)__builtin_ia32_psignw256((__v16hi)
__a, (__v16hi)
__b);
2043 return (__m256i)__builtin_ia32_psignd256((__v8si)
__a, (__v8si)
__b);
2063#define _mm256_slli_si256(a, imm) \
2064 ((__m256i)__builtin_ia32_pslldqi256_byteshift((__v32qi)(__m256i)(a), \
2084#define _mm256_bslli_epi128(a, imm) \
2085 ((__m256i)__builtin_ia32_pslldqi256_byteshift((__v32qi)(__m256i)(a), \
2103 return (__m256i)__builtin_ia32_psllwi256((__v16hi)
__a, __count);
2124 return (__m256i)__builtin_ia32_psllw256((__v16hi)
__a, (__v8hi)__count);
2142 return (__m256i)__builtin_ia32_pslldi256((__v8si)
__a, __count);
2163 return (__m256i)__builtin_ia32_pslld256((__v8si)
__a, (__v4si)__count);
2181 return __builtin_ia32_psllqi256((__v4di)
__a, __count);
2202 return __builtin_ia32_psllq256((__v4di)
__a, __count);
2221 return (__m256i)__builtin_ia32_psrawi256((__v16hi)
__a, __count);
2243 return (__m256i)__builtin_ia32_psraw256((__v16hi)
__a, (__v8hi)__count);
2262 return (__m256i)__builtin_ia32_psradi256((__v8si)
__a, __count);
2284 return (__m256i)__builtin_ia32_psrad256((__v8si)
__a, (__v4si)__count);
2304#define _mm256_srli_si256(a, imm) \
2305 ((__m256i)__builtin_ia32_psrldqi256_byteshift((__v32qi)(__m256i)(a), \
2325#define _mm256_bsrli_epi128(a, imm) \
2326 ((__m256i)__builtin_ia32_psrldqi256_byteshift((__v32qi)(__m256i)(a), \
2344 return (__m256i)__builtin_ia32_psrlwi256((__v16hi)
__a, __count);
2365 return (__m256i)__builtin_ia32_psrlw256((__v16hi)
__a, (__v8hi)__count);
2383 return (__m256i)__builtin_ia32_psrldi256((__v8si)
__a, __count);
2404 return (__m256i)__builtin_ia32_psrld256((__v8si)
__a, (__v4si)__count);
2422 return __builtin_ia32_psrlqi256((__v4di)
__a, __count);
2443 return __builtin_ia32_psrlq256((__v4di)
__a, __count);
2470 return (__m256i)((__v32qu)
__a - (__v32qu)
__b);
2497 return (__m256i)((__v16hu)
__a - (__v16hu)
__b);
2523 return (__m256i)((__v8su)
__a - (__v8su)
__b);
2549 return (__m256i)((__v4du)
__a - (__v4du)
__b);
2574 return (__m256i)__builtin_elementwise_sub_sat((__v32qs)
__a, (__v32qs)
__b);
2599 return (__m256i)__builtin_elementwise_sub_sat((__v16hi)
__a, (__v16hi)
__b);
2625 return (__m256i)__builtin_elementwise_sub_sat((__v32qu)
__a, (__v32qu)
__b);
2650 return (__m256i)__builtin_elementwise_sub_sat((__v16hu)
__a, (__v16hu)
__b);
2683 return (__m256i)__builtin_shufflevector((__v32qi)
__a, (__v32qi)
__b, 8, 32+8, 9, 32+9, 10, 32+10, 11, 32+11, 12, 32+12, 13, 32+13, 14, 32+14, 15, 32+15, 24, 32+24, 25, 32+25, 26, 32+26, 27, 32+27, 28, 32+28, 29, 32+29, 30, 32+30, 31, 32+31);
2717 return (__m256i)__builtin_shufflevector((__v16hi)
__a, (__v16hi)
__b, 4, 16+4, 5, 16+5, 6, 16+6, 7, 16+7, 12, 16+12, 13, 16+13, 14, 16+14, 15, 16+15);
2750 return (__m256i)__builtin_shufflevector((__v8si)
__a, (__v8si)
__b, 2, 8+2, 3, 8+3, 6, 8+6, 7, 8+7);
2779 return (__m256i)__builtin_shufflevector((__v4di)
__a, (__v4di)
__b, 1, 4+1, 3, 4+3);
2812 return (__m256i)__builtin_shufflevector((__v32qi)
__a, (__v32qi)
__b, 0, 32+0, 1, 32+1, 2, 32+2, 3, 32+3, 4, 32+4, 5, 32+5, 6, 32+6, 7, 32+7, 16, 32+16, 17, 32+17, 18, 32+18, 19, 32+19, 20, 32+20, 21, 32+21, 22, 32+22, 23, 32+23);
2846 return (__m256i)__builtin_shufflevector((__v16hi)
__a, (__v16hi)
__b, 0, 16+0, 1, 16+1, 2, 16+2, 3, 16+3, 8, 16+8, 9, 16+9, 10, 16+10, 11, 16+11);
2879 return (__m256i)__builtin_shufflevector((__v8si)
__a, (__v8si)
__b, 0, 8+0, 1, 8+1, 4, 8+4, 5, 8+5);
2908 return (__m256i)__builtin_shufflevector((__v4di)
__a, (__v4di)
__b, 0, 4+0, 2, 4+2);
2926 return (__m256i)((__v4du)
__a ^ (__v4du)
__b);
2944 return (__m256i)__builtin_nontemporal_load((
const __v4di_aligned *)__V);
2960 return (__m128)__builtin_shufflevector((__v4sf)__X, (__v4sf)__X, 0, 0, 0, 0);
2976 return __builtin_shufflevector((__v2df)
__a, (__v2df)
__a, 0, 0);
2992 return (__m256)__builtin_shufflevector((__v4sf)__X, (__v4sf)__X, 0, 0, 0, 0, 0, 0, 0, 0);
3008 return (__m256d)__builtin_shufflevector((__v2df)__X, (__v2df)__X, 0, 0, 0, 0);
3023 return (__m256i)__builtin_shufflevector((__v2di)__X, (__v2di)__X, 0, 1, 0, 1);
3026#define _mm_broadcastsi128_si256(X) _mm256_broadcastsi128_si256(X)
3061#define _mm_blend_epi32(V1, V2, M) \
3062 ((__m128i)__builtin_ia32_pblendd128((__v4si)(__m128i)(V1), \
3063 (__v4si)(__m128i)(V2), (int)(M)))
3098#define _mm256_blend_epi32(V1, V2, M) \
3099 ((__m256i)__builtin_ia32_pblendd256((__v8si)(__m256i)(V1), \
3100 (__v8si)(__m256i)(V2), (int)(M)))
3114 return (__m256i)__builtin_shufflevector((__v16qi)__X, (__v16qi)__X, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
3129 return (__m256i)__builtin_shufflevector((__v8hi)__X, (__v8hi)__X, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
3144 return (__m256i)__builtin_shufflevector((__v4si)__X, (__v4si)__X, 0, 0, 0, 0, 0, 0, 0, 0);
3159 return (__m256i)__builtin_shufflevector((__v2di)__X, (__v2di)__X, 0, 0, 0, 0);
3174 return (__m128i)__builtin_shufflevector((__v16qi)__X, (__v16qi)__X, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
3189 return (__m128i)__builtin_shufflevector((__v8hi)__X, (__v8hi)__X, 0, 0, 0, 0, 0, 0, 0, 0);
3204 return (__m128i)__builtin_shufflevector((__v4si)__X, (__v4si)__X, 0, 0, 0, 0);
3219 return (__m128i)__builtin_shufflevector((__v2di)__X, (__v2di)__X, 0, 0);
3247 return (__m256i)__builtin_ia32_permvarsi256((__v8si)
__a, (__v8si)
__b);
3277#define _mm256_permute4x64_pd(V, M) \
3278 ((__m256d)__builtin_ia32_permdf256((__v4df)(__m256d)(V), (int)(M)))
3305 return (__m256)__builtin_ia32_permvarsf256((__v8sf)
__a, (__v8si)
__b);
3335#define _mm256_permute4x64_epi64(V, M) \
3336 ((__m256i)__builtin_ia32_permdi256((__v4di)(__m256i)(V), (int)(M)))
3381#define _mm256_permute2x128_si256(V1, V2, M) \
3382 ((__m256i)__builtin_ia32_permti256((__m256i)(V1), (__m256i)(V2), (int)(M)))
3401#define _mm256_extracti128_si256(V, M) \
3402 ((__m128i)__builtin_ia32_extract128i256((__v4di)(__m256i)(V), (int)(M)))
3424#define _mm256_inserti128_si256(V1, V2, M) \
3425 ((__m256i)__builtin_ia32_insert128i256((__v4di)(__m256i)(V1), \
3426 (__v2di)(__m128i)(V2), (int)(M)))
3457 return (__m256i)__builtin_ia32_maskloadd256((
const __v8si *)__X, (__v8si)__M);
3489 return (__m256i)__builtin_ia32_maskloadq256((
const __v4di *)__X, (__v4di)__M);
3521 return (__m128i)__builtin_ia32_maskloadd((
const __v4si *)__X, (__v4si)__M);
3553 return (__m128i)__builtin_ia32_maskloadq((
const __v2di *)__X, (__v2di)__M);
3583 __builtin_ia32_maskstored256((__v8si *)__X, (__v8si)__M, (__v8si)
__Y);
3613 __builtin_ia32_maskstoreq256((__v4di *)__X, (__v4di)__M, (__v4di)
__Y);
3643 __builtin_ia32_maskstored((__v4si *)__X, (__v4si)__M, (__v4si)
__Y);
3673 __builtin_ia32_maskstoreq(( __v2di *)__X, (__v2di)__M, (__v2di)
__Y);
3695 return (__m256i)__builtin_ia32_psllv8si((__v8si)__X, (__v8si)
__Y);
3717 return (__m128i)__builtin_ia32_psllv4si((__v4si)__X, (__v4si)
__Y);
3739 return (__m256i)__builtin_ia32_psllv4di((__v4di)__X, (__v4di)
__Y);
3761 return (__m128i)__builtin_ia32_psllv2di((__v2di)__X, (__v2di)
__Y);
3784 return (__m256i)__builtin_ia32_psrav8si((__v8si)__X, (__v8si)
__Y);
3807 return (__m128i)__builtin_ia32_psrav4si((__v4si)__X, (__v4si)
__Y);
3829 return (__m256i)__builtin_ia32_psrlv8si((__v8si)__X, (__v8si)
__Y);
3851 return (__m128i)__builtin_ia32_psrlv4si((__v4si)__X, (__v4si)
__Y);
3873 return (__m256i)__builtin_ia32_psrlv4di((__v4di)__X, (__v4di)
__Y);
3895 return (__m128i)__builtin_ia32_psrlv2di((__v2di)__X, (__v2di)
__Y);
3941#define _mm_mask_i32gather_pd(a, m, i, mask, s) \
3942 ((__m128d)__builtin_ia32_gatherd_pd((__v2df)(__m128i)(a), \
3943 (double const *)(m), \
3944 (__v4si)(__m128i)(i), \
3945 (__v2df)(__m128d)(mask), (s)))
3989#define _mm256_mask_i32gather_pd(a, m, i, mask, s) \
3990 ((__m256d)__builtin_ia32_gatherd_pd256((__v4df)(__m256d)(a), \
3991 (double const *)(m), \
3992 (__v4si)(__m128i)(i), \
3993 (__v4df)(__m256d)(mask), (s)))
4037#define _mm_mask_i64gather_pd(a, m, i, mask, s) \
4038 ((__m128d)__builtin_ia32_gatherq_pd((__v2df)(__m128d)(a), \
4039 (double const *)(m), \
4040 (__v2di)(__m128i)(i), \
4041 (__v2df)(__m128d)(mask), (s)))
4085#define _mm256_mask_i64gather_pd(a, m, i, mask, s) \
4086 ((__m256d)__builtin_ia32_gatherq_pd256((__v4df)(__m256d)(a), \
4087 (double const *)(m), \
4088 (__v4di)(__m256i)(i), \
4089 (__v4df)(__m256d)(mask), (s)))
4133#define _mm_mask_i32gather_ps(a, m, i, mask, s) \
4134 ((__m128)__builtin_ia32_gatherd_ps((__v4sf)(__m128)(a), \
4135 (float const *)(m), \
4136 (__v4si)(__m128i)(i), \
4137 (__v4sf)(__m128)(mask), (s)))
4181#define _mm256_mask_i32gather_ps(a, m, i, mask, s) \
4182 ((__m256)__builtin_ia32_gatherd_ps256((__v8sf)(__m256)(a), \
4183 (float const *)(m), \
4184 (__v8si)(__m256i)(i), \
4185 (__v8sf)(__m256)(mask), (s)))
4232#define _mm_mask_i64gather_ps(a, m, i, mask, s) \
4233 ((__m128)__builtin_ia32_gatherq_ps((__v4sf)(__m128)(a), \
4234 (float const *)(m), \
4235 (__v2di)(__m128i)(i), \
4236 (__v4sf)(__m128)(mask), (s)))
4280#define _mm256_mask_i64gather_ps(a, m, i, mask, s) \
4281 ((__m128)__builtin_ia32_gatherq_ps256((__v4sf)(__m128)(a), \
4282 (float const *)(m), \
4283 (__v4di)(__m256i)(i), \
4284 (__v4sf)(__m128)(mask), (s)))
4328#define _mm_mask_i32gather_epi32(a, m, i, mask, s) \
4329 ((__m128i)__builtin_ia32_gatherd_d((__v4si)(__m128i)(a), \
4331 (__v4si)(__m128i)(i), \
4332 (__v4si)(__m128i)(mask), (s)))
4376#define _mm256_mask_i32gather_epi32(a, m, i, mask, s) \
4377 ((__m256i)__builtin_ia32_gatherd_d256((__v8si)(__m256i)(a), \
4379 (__v8si)(__m256i)(i), \
4380 (__v8si)(__m256i)(mask), (s)))
4427#define _mm_mask_i64gather_epi32(a, m, i, mask, s) \
4428 ((__m128i)__builtin_ia32_gatherq_d((__v4si)(__m128i)(a), \
4430 (__v2di)(__m128i)(i), \
4431 (__v4si)(__m128i)(mask), (s)))
4475#define _mm256_mask_i64gather_epi32(a, m, i, mask, s) \
4476 ((__m128i)__builtin_ia32_gatherq_d256((__v4si)(__m128i)(a), \
4478 (__v4di)(__m256i)(i), \
4479 (__v4si)(__m128i)(mask), (s)))
4524#define _mm_mask_i32gather_epi64(a, m, i, mask, s) \
4525 ((__m128i)__builtin_ia32_gatherd_q((__v2di)(__m128i)(a), \
4526 (long long const *)(m), \
4527 (__v4si)(__m128i)(i), \
4528 (__v2di)(__m128i)(mask), (s)))
4572#define _mm256_mask_i32gather_epi64(a, m, i, mask, s) \
4573 ((__m256i)__builtin_ia32_gatherd_q256((__v4di)(__m256i)(a), \
4574 (long long const *)(m), \
4575 (__v4si)(__m128i)(i), \
4576 (__v4di)(__m256i)(mask), (s)))
4620#define _mm_mask_i64gather_epi64(a, m, i, mask, s) \
4621 ((__m128i)__builtin_ia32_gatherq_q((__v2di)(__m128i)(a), \
4622 (long long const *)(m), \
4623 (__v2di)(__m128i)(i), \
4624 (__v2di)(__m128i)(mask), (s)))
4668#define _mm256_mask_i64gather_epi64(a, m, i, mask, s) \
4669 ((__m256i)__builtin_ia32_gatherq_q256((__v4di)(__m256i)(a), \
4670 (long long const *)(m), \
4671 (__v4di)(__m256i)(i), \
4672 (__v4di)(__m256i)(mask), (s)))
4702#define _mm_i32gather_pd(m, i, s) \
4703 ((__m128d)__builtin_ia32_gatherd_pd((__v2df)_mm_undefined_pd(), \
4704 (double const *)(m), \
4705 (__v4si)(__m128i)(i), \
4706 (__v2df)_mm_cmpeq_pd(_mm_setzero_pd(), \
4707 _mm_setzero_pd()), \
4737#define _mm256_i32gather_pd(m, i, s) \
4738 ((__m256d)__builtin_ia32_gatherd_pd256((__v4df)_mm256_undefined_pd(), \
4739 (double const *)(m), \
4740 (__v4si)(__m128i)(i), \
4741 (__v4df)_mm256_cmp_pd(_mm256_setzero_pd(), \
4742 _mm256_setzero_pd(), \
4773#define _mm_i64gather_pd(m, i, s) \
4774 ((__m128d)__builtin_ia32_gatherq_pd((__v2df)_mm_undefined_pd(), \
4775 (double const *)(m), \
4776 (__v2di)(__m128i)(i), \
4777 (__v2df)_mm_cmpeq_pd(_mm_setzero_pd(), \
4778 _mm_setzero_pd()), \
4808#define _mm256_i64gather_pd(m, i, s) \
4809 ((__m256d)__builtin_ia32_gatherq_pd256((__v4df)_mm256_undefined_pd(), \
4810 (double const *)(m), \
4811 (__v4di)(__m256i)(i), \
4812 (__v4df)_mm256_cmp_pd(_mm256_setzero_pd(), \
4813 _mm256_setzero_pd(), \
4844#define _mm_i32gather_ps(m, i, s) \
4845 ((__m128)__builtin_ia32_gatherd_ps((__v4sf)_mm_undefined_ps(), \
4846 (float const *)(m), \
4847 (__v4si)(__m128i)(i), \
4848 (__v4sf)_mm_cmpeq_ps(_mm_setzero_ps(), \
4849 _mm_setzero_ps()), \
4879#define _mm256_i32gather_ps(m, i, s) \
4880 ((__m256)__builtin_ia32_gatherd_ps256((__v8sf)_mm256_undefined_ps(), \
4881 (float const *)(m), \
4882 (__v8si)(__m256i)(i), \
4883 (__v8sf)_mm256_cmp_ps(_mm256_setzero_ps(), \
4884 _mm256_setzero_ps(), \
4917#define _mm_i64gather_ps(m, i, s) \
4918 ((__m128)__builtin_ia32_gatherq_ps((__v4sf)_mm_undefined_ps(), \
4919 (float const *)(m), \
4920 (__v2di)(__m128i)(i), \
4921 (__v4sf)_mm_cmpeq_ps(_mm_setzero_ps(), \
4922 _mm_setzero_ps()), \
4952#define _mm256_i64gather_ps(m, i, s) \
4953 ((__m128)__builtin_ia32_gatherq_ps256((__v4sf)_mm_undefined_ps(), \
4954 (float const *)(m), \
4955 (__v4di)(__m256i)(i), \
4956 (__v4sf)_mm_cmpeq_ps(_mm_setzero_ps(), \
4957 _mm_setzero_ps()), \
4987#define _mm_i32gather_epi32(m, i, s) \
4988 ((__m128i)__builtin_ia32_gatherd_d((__v4si)_mm_undefined_si128(), \
4989 (int const *)(m), (__v4si)(__m128i)(i), \
4990 (__v4si)_mm_set1_epi32(-1), (s)))
5019#define _mm256_i32gather_epi32(m, i, s) \
5020 ((__m256i)__builtin_ia32_gatherd_d256((__v8si)_mm256_undefined_si256(), \
5021 (int const *)(m), (__v8si)(__m256i)(i), \
5022 (__v8si)_mm256_set1_epi32(-1), (s)))
5053#define _mm_i64gather_epi32(m, i, s) \
5054 ((__m128i)__builtin_ia32_gatherq_d((__v4si)_mm_undefined_si128(), \
5055 (int const *)(m), (__v2di)(__m128i)(i), \
5056 (__v4si)_mm_set1_epi32(-1), (s)))
5085#define _mm256_i64gather_epi32(m, i, s) \
5086 ((__m128i)__builtin_ia32_gatherq_d256((__v4si)_mm_undefined_si128(), \
5087 (int const *)(m), (__v4di)(__m256i)(i), \
5088 (__v4si)_mm_set1_epi32(-1), (s)))
5118#define _mm_i32gather_epi64(m, i, s) \
5119 ((__m128i)__builtin_ia32_gatherd_q((__v2di)_mm_undefined_si128(), \
5120 (long long const *)(m), \
5121 (__v4si)(__m128i)(i), \
5122 (__v2di)_mm_set1_epi64x(-1), (s)))
5151#define _mm256_i32gather_epi64(m, i, s) \
5152 ((__m256i)__builtin_ia32_gatherd_q256((__v4di)_mm256_undefined_si256(), \
5153 (long long const *)(m), \
5154 (__v4si)(__m128i)(i), \
5155 (__v4di)_mm256_set1_epi64x(-1), (s)))
5184#define _mm_i64gather_epi64(m, i, s) \
5185 ((__m128i)__builtin_ia32_gatherq_q((__v2di)_mm_undefined_si128(), \
5186 (long long const *)(m), \
5187 (__v2di)(__m128i)(i), \
5188 (__v2di)_mm_set1_epi64x(-1), (s)))
5217#define _mm256_i64gather_epi64(m, i, s) \
5218 ((__m256i)__builtin_ia32_gatherq_q256((__v4di)_mm256_undefined_si256(), \
5219 (long long const *)(m), \
5220 (__v4di)(__m256i)(i), \
5221 (__v4di)_mm256_set1_epi64x(-1), (s)))
5223#undef __DEFAULT_FN_ATTRS256
5224#undef __DEFAULT_FN_ATTRS128
5225#undef __DEFAULT_FN_ATTRS256_CONSTEXPR
5226#undef __DEFAULT_FN_ATTRS128_CONSTEXPR
_Float16 __2f16 __attribute__((ext_vector_type(2)))
Zeroes the upper 128 bits (bits 255:128) of all YMM registers.
static __inline__ vector float vector float __b
#define __DEFAULT_FN_ATTRS128
#define __DEFAULT_FN_ATTRS256
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_srl_epi32(__m256i __a, __m128i __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_sllv_epi32(__m256i __X, __m256i __Y)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __X left by the number of bits given...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi32_epi64(__m128i __V)
Sign-extends 32-bit elements from the 128-bit vector of [4 x i32] in __V and returns the 64-bit value...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_srl_epi16(__m256i __a, __m128i __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a right by the number of bits giv...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hadd_epi16(__m256i __a, __m256i __b)
Horizontally adds the adjacent pairs of 16-bit integers from two 256-bit vectors of [16 x i16] and re...
static __inline__ void __DEFAULT_FN_ATTRS256 _mm256_maskstore_epi64(long long *__X, __m256i __M, __m256i __Y)
Conditionally stores four 64-bit integer elements from the 256-bit vector of [4 x i64] in __Y to memo...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_stream_load_si256(const void *__V)
Loads the 256-bit integer vector from memory __V using a non-temporal memory hint and returns the vec...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpgt_epi8(__m256i __a, __m256i __b)
Compares corresponding signed bytes in the 256-bit integer vectors in __a and __b for greater-than an...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epi32(__m256i __a, __m256i __b)
Compares the corresponding signed 32-bit integers in the two 256-bit vectors of [8 x i32] in __a and ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mul_epu32(__m256i __a, __m256i __b)
Multiplies unsigned 32-bit integers from even-numered elements of two 256-bit vectors of [8 x i32] an...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epu16(__m256i __a, __m256i __b)
Compares the corresponding unsigned 16-bit integers in the two 256-bit vectors of [16 x i16] in __a a...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sll_epi32(__m256i __a, __m128i __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a left by the number of bits given...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_srav_epi32(__m128i __X, __m128i __Y)
Shifts each 32-bit element of the 128-bit vector of [4 x i32] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpackhi_epi32(__m256i __a, __m256i __b)
Unpacks and interleaves 32-bit integers from parts of the 256-bit vectors of [8 x i32] in __a and __b...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_andnot_si256(__m256i __a, __m256i __b)
Computes the bitwise AND of the 256-bit integer vector in __b with the bitwise NOT of the 256-bit int...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mulhi_epi16(__m256i __a, __m256i __b)
Multiplies signed 16-bit integer elements of two 256-bit vectors of [16 x i16], and returns the upper...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_adds_epu8(__m256i __a, __m256i __b)
Adds 8-bit integers from corresponding bytes of two 256-bit integer vectors using unsigned saturation...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sll_epi64(__m256i __a, __m128i __count)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __a left by the number of bits given...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_packs_epi32(__m256i __a, __m256i __b)
Converts the elements of two 256-bit vectors of [8 x i32] to 16-bit integers using signed saturation,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskload_epi64(long long const *__X, __m256i __M)
Conditionally loads four 64-bit integer elements from memory __X, if the most significant bit of the ...
static __inline__ void __DEFAULT_FN_ATTRS256 _mm256_maskstore_epi32(int *__X, __m256i __M, __m256i __Y)
Conditionally stores eight 32-bit integer elements from the 256-bit vector of [8 x i32] in __Y to mem...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mulhi_epu16(__m256i __a, __m256i __b)
Multiplies unsigned 16-bit integer elements of two 256-bit vectors of [16 x i16], and returns the upp...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_shuffle_epi8(__m256i __a, __m256i __b)
Shuffles 8-bit integers in the 256-bit integer vector __a according to control information in the 256...
#define __DEFAULT_FN_ATTRS128_CONSTEXPR
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastb_epi8(__m128i __X)
Broadcasts the low byte from the 128-bit integer vector in __X to all bytes of the 128-bit result.
static __inline__ __m128d __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastsd_pd(__m128d __a)
Broadcasts the 64-bit floating-point value from the low element of the 128-bit vector of [2 x double]...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskload_epi32(int const *__X, __m256i __M)
Conditionally loads eight 32-bit integer elements from memory __X, if the most significant bit of the...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu8_epi16(__m128i __V)
Zero-extends bytes from the 128-bit integer vector in __V and returns the 16-bit values in the corres...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_subs_epu16(__m256i __a, __m256i __b)
Subtracts 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16] using unsi...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastb_epi8(__m128i __X)
Broadcasts the low byte from the 128-bit integer vector in __X to all bytes of the 256-bit result.
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sub_epi16(__m256i __a, __m256i __b)
Subtracts 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16].
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_permutevar8x32_epi32(__m256i __a, __m256i __b)
Sets the result's 256-bit vector of [8 x i32] to copies of elements of the 256-bit vector of [8 x i32...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastq_epi64(__m128i __X)
Broadcasts the low element from the 128-bit vector of [2 x i64] in __X to both elements of the result...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_slli_epi64(__m256i __a, int __count)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __a left by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpacklo_epi16(__m256i __a, __m256i __b)
Unpacks and interleaves 16-bit integers from parts of the 256-bit vectors of [16 x i16] in __a and __...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_abs_epi32(__m256i __a)
Computes the absolute value of each signed 32-bit element in the 256-bit vector of [8 x i32] in __a a...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sub_epi64(__m256i __a, __m256i __b)
Subtracts 64-bit integers from corresponding elements of two 256-bit vectors of [4 x i64].
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maddubs_epi16(__m256i __a, __m256i __b)
Multiplies each unsigned byte from the 256-bit integer vector in __a with the corresponding signed by...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hsub_epi16(__m256i __a, __m256i __b)
Horizontally subtracts adjacent pairs of 16-bit integers from two 256-bit vectors of [16 x i16] and r...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srav_epi32(__m256i __X, __m256i __Y)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mullo_epi16(__m256i __a, __m256i __b)
Multiplies signed 16-bit integer elements of two 256-bit vectors of [16 x i16], and returns the lower...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu8_epi64(__m128i __V)
Zero-extends the first four bytes from the 128-bit integer vector in __V and returns the 64-bit value...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastw_epi16(__m128i __X)
Broadcasts the low element from the 128-bit vector of [8 x i16] in __X to all elements of the result'...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpeq_epi32(__m256i __a, __m256i __b)
Compares corresponding elements in the 256-bit vectors of [8 x i32] in __a and __b for equality and r...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_adds_epi16(__m256i __a, __m256i __b)
Adds 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16] using signed sa...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mulhrs_epi16(__m256i __a, __m256i __b)
Multiplies signed 16-bit integer elements of two 256-bit vectors of [16 x i16], truncates the 32-bit ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epi8(__m256i __a, __m256i __b)
Compares the corresponding signed bytes in the two 256-bit integer vectors in __a and __b and returns...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sll_epi16(__m256i __a, __m128i __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a left by the number of bits spec...
static __inline__ void __DEFAULT_FN_ATTRS128 _mm_maskstore_epi64(long long *__X, __m128i __M, __m128i __Y)
Conditionally stores two 64-bit integer elements from the 128-bit vector of [2 x i64] in __Y to memor...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_add_epi8(__m256i __a, __m256i __b)
Adds 8-bit integers from corresponding bytes of two 256-bit integer vectors and returns the lower 8 b...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu32_epi64(__m128i __V)
Zero-extends 32-bit elements from the 128-bit vector of [4 x i32] in __V and returns the 64-bit value...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu8_epi32(__m128i __V)
Zero-extends bytes from the lower half of the 128-bit integer vector in __V and returns the 32-bit va...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srli_epi16(__m256i __a, int __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a right by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_add_epi16(__m256i __a, __m256i __b)
Adds 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16] and returns the...
static __inline__ __m256 __DEFAULT_FN_ATTRS256 _mm256_permutevar8x32_ps(__m256 __a, __m256i __b)
Sets the result's 256-bit vector of [8 x float] to copies of elements of the 256-bit vector of [8 x f...
static __inline__ int __DEFAULT_FN_ATTRS256 _mm256_movemask_epi8(__m256i __a)
Creates a 32-bit integer mask from the most significant bit of each byte in the 256-bit integer vecto...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epi16(__m256i __a, __m256i __b)
Compares the corresponding signed 16-bit integers in the two 256-bit vectors of [16 x i16] in __a and...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi8_epi16(__m128i __V)
Sign-extends bytes from the 128-bit integer vector in __V and returns the 16-bit values in the corres...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_srlv_epi32(__m128i __X, __m128i __Y)
Shifts each 32-bit element of the 128-bit vector of [4 x i32] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_blendv_epi8(__m256i __V1, __m256i __V2, __m256i __M)
Merges 8-bit integer values from either of the two 256-bit vectors __V1 or __V2, as specified by the ...
static __inline__ __m256 __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastss_ps(__m128 __X)
Broadcasts the 32-bit floating-point value from the low element of the 128-bit vector of [4 x float] ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpeq_epi64(__m256i __a, __m256i __b)
Compares corresponding elements in the 256-bit vectors of [4 x i64] in __a and __b for equality and r...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_add_epi64(__m256i __a, __m256i __b)
Adds 64-bit integers from corresponding elements of two 256-bit vectors of [4 x i64] and returns the ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sign_epi32(__m256i __a, __m256i __b)
Sets each element of the result to the corresponding element of the 256-bit vector of [8 x i32] in __...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpacklo_epi8(__m256i __a, __m256i __b)
Unpacks and interleaves 8-bit integers from parts of the 256-bit integer vectors in __a and __b to fo...
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskload_epi32(int const *__X, __m128i __M)
Conditionally loads four 32-bit integer elements from memory __X, if the most significant bit of the ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srli_epi64(__m256i __a, int __count)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __a right by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hadds_epi16(__m256i __a, __m256i __b)
Horizontally adds the adjacent pairs of 16-bit integers from two 256-bit vectors of [16 x i16] using ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastw_epi16(__m128i __X)
Broadcasts the low element from the 128-bit vector of [8 x i16] in __X to all elements of the result'...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hsub_epi32(__m256i __a, __m256i __b)
Horizontally subtracts adjacent pairs of 32-bit integers from two 256-bit vectors of [8 x i32] and re...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpacklo_epi64(__m256i __a, __m256i __b)
Unpacks and interleaves 64-bit integers from parts of the 256-bit vectors of [4 x i64] in __a and __b...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sign_epi8(__m256i __a, __m256i __b)
Sets each byte of the result to the corresponding byte of the 256-bit integer vector in __a,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpgt_epi16(__m256i __a, __m256i __b)
Compares corresponding signed elements in the 256-bit vectors of [16 x i16] in __a and __b for greate...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srai_epi16(__m256i __a, int __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a right by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi8_epi32(__m128i __V)
Sign-extends bytes from the lower half of the 128-bit integer vector in __V and returns the 32-bit va...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastq_epi64(__m128i __X)
Broadcasts the low element from the 128-bit vector of [2 x i64] in __X to all elements of the result'...
#define __DEFAULT_FN_ATTRS256_CONSTEXPR
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mul_epi32(__m256i __a, __m256i __b)
Multiplies signed 32-bit integers from even-numbered elements of two 256-bit vectors of [8 x i32] and...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epu16(__m256i __a, __m256i __b)
Compares the corresponding unsigned 16-bit integers in the two 256-bit vectors of [16 x i16] in __a a...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_packs_epi16(__m256i __a, __m256i __b)
Converts the elements of two 256-bit vectors of [16 x i16] to 8-bit integers using signed saturation,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_mullo_epi32(__m256i __a, __m256i __b)
Multiplies signed 32-bit integer elements of two 256-bit vectors of [8 x i32], and returns the lower ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epi32(__m256i __a, __m256i __b)
Compares the corresponding signed 32-bit integers in the two 256-bit vectors of [8 x i32] in __a and ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hsubs_epi16(__m256i __a, __m256i __b)
Horizontally subtracts adjacent pairs of 16-bit integers from two 256-bit vectors of [16 x i16] using...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpgt_epi32(__m256i __a, __m256i __b)
Compares corresponding signed elements in the 256-bit vectors of [8 x i32] in __a and __b for greater...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpackhi_epi64(__m256i __a, __m256i __b)
Unpacks and interleaves 64-bit integers from parts of the 256-bit vectors of [4 x i64] in __a and __b...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_subs_epu8(__m256i __a, __m256i __b)
Subtracts 8-bit integers from corresponding bytes of two 256-bit integer vectors using unsigned satur...
static __inline__ __m128 __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastss_ps(__m128 __X)
Broadcasts the 32-bit floating-point value from the low element of the 128-bit vector of [4 x float] ...
static __inline__ void __DEFAULT_FN_ATTRS128 _mm_maskstore_epi32(int *__X, __m128i __M, __m128i __Y)
Conditionally stores four 32-bit integer elements from the 128-bit vector of [4 x i32] in __Y to memo...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sub_epi8(__m256i __a, __m256i __b)
Subtracts 8-bit integers from corresponding bytes of two 256-bit integer vectors.
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_subs_epi16(__m256i __a, __m256i __b)
Subtracts 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16] using sign...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sra_epi32(__m256i __a, __m128i __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpeq_epi8(__m256i __a, __m256i __b)
Compares corresponding bytes in the 256-bit integer vectors in __a and __b for equality and returns t...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_srlv_epi64(__m128i __X, __m128i __Y)
Shifts each 64-bit element of the 128-bit vector of [2 x i64] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srlv_epi32(__m256i __X, __m256i __Y)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_abs_epi8(__m256i __a)
Computes the absolute value of each signed byte in the 256-bit integer vector __a and returns each va...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu16_epi64(__m128i __V)
Zero-extends 16-bit elements from the lower half of the 128-bit vector of [8 x i16] in __V and return...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_madd_epi16(__m256i __a, __m256i __b)
Multiplies corresponding 16-bit elements of two 256-bit vectors of [16 x i16], forming 32-bit interme...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpacklo_epi32(__m256i __a, __m256i __b)
Unpacks and interleaves 32-bit integers from parts of the 256-bit vectors of [8 x i32] in __a and __b...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi8_epi64(__m128i __V)
Sign-extends the first four bytes from the 128-bit integer vector in __V and returns the 64-bit value...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srli_epi32(__m256i __a, int __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a right by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_packus_epi32(__m256i __V1, __m256i __V2)
Converts elements from two 256-bit vectors of [8 x i32] to 16-bit integers using unsigned saturation,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi16_epi64(__m128i __V)
Sign-extends 16-bit elements from the lower half of the 128-bit vector of [8 x i16] in __V and return...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpackhi_epi8(__m256i __a, __m256i __b)
Unpacks and interleaves 8-bit integers from parts of the 256-bit integer vectors in __a and __b to fo...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_and_si256(__m256i __a, __m256i __b)
Computes the bitwise AND of the 256-bit integer vectors in __a and __b.
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_slli_epi32(__m256i __a, int __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a left by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastsi128_si256(__m128i __X)
Broadcasts the 128-bit integer data from __X to both the lower and upper halves of the 256-bit result...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srai_epi32(__m256i __a, int __count)
Shifts each 32-bit element of the 256-bit vector of [8 x i32] in __a right by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sub_epi32(__m256i __a, __m256i __b)
Subtracts 32-bit integers from corresponding elements of two 256-bit vectors of [8 x i32].
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_abs_epi16(__m256i __a)
Computes the absolute value of each signed 16-bit element in the 256-bit vector of [16 x i16] in __a ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epu8(__m256i __a, __m256i __b)
Compares the corresponding unsigned bytes in the two 256-bit integer vectors in __a and __b and retur...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_or_si256(__m256i __a, __m256i __b)
Computes the bitwise OR of the 256-bit integer vectors in __a and __b.
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_packus_epi16(__m256i __a, __m256i __b)
Converts elements from two 256-bit vectors of [16 x i16] to 8-bit integers using unsigned saturation,...
static __inline__ __m256d __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastsd_pd(__m128d __X)
Broadcasts the 64-bit floating-point value from the low element of the 128-bit vector of [2 x double]...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epu8(__m256i __a, __m256i __b)
Compares the corresponding unsigned bytes in the two 256-bit integer vectors in __a and __b and retur...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepu16_epi32(__m128i __V)
Zero-extends 16-bit elements from the 128-bit vector of [8 x i16] in __V and returns the 32-bit value...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epu32(__m256i __a, __m256i __b)
Compares the corresponding unsigned 32-bit integers in the two 256-bit vectors of [8 x i32] in __a an...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_sllv_epi64(__m128i __X, __m128i __Y)
Shifts each 64-bit element of the 128-bit vector of [2 x i64] in __X left by the number of bits given...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_sllv_epi32(__m128i __X, __m128i __Y)
Shifts each 32-bit element of the 128-bit vector of [4 x i32] in __X left by the number of bits given...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_srl_epi64(__m256i __a, __m128i __count)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __a right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epi16(__m256i __a, __m256i __b)
Compares the corresponding signed 16-bit integers in the two 256-bit vectors of [16 x i16] in __a and...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_avg_epu16(__m256i __a, __m256i __b)
Computes the averages of the corresponding unsigned 16-bit integers in the two 256-bit vectors of [16...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_xor_si256(__m256i __a, __m256i __b)
Computes the bitwise XOR of the 256-bit integer vectors in __a and __b.
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_min_epu32(__m256i __a, __m256i __b)
Compares the corresponding unsigned 32-bit integers in the two 256-bit vectors of [8 x i32] in __a an...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sra_epi16(__m256i __a, __m128i __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a right by the number of bits giv...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cvtepi16_epi32(__m128i __V)
Sign-extends 16-bit elements from the 128-bit vector of [8 x i16] in __V and returns the 32-bit value...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_adds_epu16(__m256i __a, __m256i __b)
Adds 16-bit integers from corresponding elements of two 256-bit vectors of [16 x i16] using unsigned ...
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskload_epi64(long long const *__X, __m128i __M)
Conditionally loads two 64-bit integer elements from memory __X, if the most significant bit of the c...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sign_epi16(__m256i __a, __m256i __b)
Sets each element of the result to the corresponding element of the 256-bit vector of [16 x i16] in _...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_hadd_epi32(__m256i __a, __m256i __b)
Horizontally adds the adjacent pairs of 32-bit integers from two 256-bit vectors of [8 x i32] and ret...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_max_epi8(__m256i __a, __m256i __b)
Compares the corresponding signed bytes in the two 256-bit integer vectors in __a and __b and returns...
static __inline__ __m128i __DEFAULT_FN_ATTRS128_CONSTEXPR _mm_broadcastd_epi32(__m128i __X)
Broadcasts the low element from the 128-bit vector of [4 x i32] in __X to all elements of the result'...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_broadcastd_epi32(__m128i __X)
Broadcasts the low element from the 128-bit vector of [4 x i32] in __X to all elements of the result'...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_unpackhi_epi16(__m256i __a, __m256i __b)
Unpacks and interleaves 16-bit integers from parts of the 256-bit vectors of [16 x i16] in __a and __...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_srlv_epi64(__m256i __X, __m256i __Y)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __X right by the number of bits give...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_subs_epi8(__m256i __a, __m256i __b)
Subtracts 8-bit integers from corresponding bytes of two 256-bit integer vectors using signed saturat...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpgt_epi64(__m256i __a, __m256i __b)
Compares corresponding signed elements in the 256-bit vectors of [4 x i64] in __a and __b for greater...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_sad_epu8(__m256i __a, __m256i __b)
Computes four sum of absolute difference (SAD) operations on sets of eight unsigned 8-bit integers fr...
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_add_epi32(__m256i __a, __m256i __b)
Adds 32-bit integers from corresponding elements of two 256-bit vectors of [8 x i32] and returns the ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_adds_epi8(__m256i __a, __m256i __b)
Adds 8-bit integers from corresponding bytes of two 256-bit integer vectors using signed saturation,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_sllv_epi64(__m256i __X, __m256i __Y)
Shifts each 64-bit element of the 256-bit vector of [4 x i64] in __X left by the number of bits given...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_slli_epi16(__m256i __a, int __count)
Shifts each 16-bit element of the 256-bit vector of [16 x i16] in __a left by __count bits,...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_cmpeq_epi16(__m256i __a, __m256i __b)
Compares corresponding elements in the 256-bit vectors of [16 x i16] in __a and __b for equality and ...
static __inline__ __m256i __DEFAULT_FN_ATTRS256_CONSTEXPR _mm256_avg_epu8(__m256i __a, __m256i __b)
Computes the averages of the corresponding unsigned bytes in the two 256-bit integer vectors in __a a...
static __inline__ void int __a
__inline unsigned int unsigned int __Y