16 #define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("sse4.1"), __min_vector_width__(128))) 19 #define _MM_FROUND_TO_NEAREST_INT 0x00 20 #define _MM_FROUND_TO_NEG_INF 0x01 21 #define _MM_FROUND_TO_POS_INF 0x02 22 #define _MM_FROUND_TO_ZERO 0x03 23 #define _MM_FROUND_CUR_DIRECTION 0x04 25 #define _MM_FROUND_RAISE_EXC 0x00 26 #define _MM_FROUND_NO_EXC 0x08 28 #define _MM_FROUND_NINT (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEAREST_INT) 29 #define _MM_FROUND_FLOOR (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEG_INF) 30 #define _MM_FROUND_CEIL (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_POS_INF) 31 #define _MM_FROUND_TRUNC (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_ZERO) 32 #define _MM_FROUND_RINT (_MM_FROUND_RAISE_EXC | _MM_FROUND_CUR_DIRECTION) 33 #define _MM_FROUND_NEARBYINT (_MM_FROUND_NO_EXC | _MM_FROUND_CUR_DIRECTION) 50 #define _mm_ceil_ps(X) _mm_round_ps((X), _MM_FROUND_CEIL) 67 #define _mm_ceil_pd(X) _mm_round_pd((X), _MM_FROUND_CEIL) 92 #define _mm_ceil_ss(X, Y) _mm_round_ss((X), (Y), _MM_FROUND_CEIL) 117 #define _mm_ceil_sd(X, Y) _mm_round_sd((X), (Y), _MM_FROUND_CEIL) 134 #define _mm_floor_ps(X) _mm_round_ps((X), _MM_FROUND_FLOOR) 151 #define _mm_floor_pd(X) _mm_round_pd((X), _MM_FROUND_FLOOR) 176 #define _mm_floor_ss(X, Y) _mm_round_ss((X), (Y), _MM_FROUND_FLOOR) 201 #define _mm_floor_sd(X, Y) _mm_round_sd((X), (Y), _MM_FROUND_FLOOR) 233 #define _mm_round_ps(X, M) \ 234 (__m128)__builtin_ia32_roundps((__v4sf)(__m128)(X), (M)) 274 #define _mm_round_ss(X, Y, M) \ 275 (__m128)__builtin_ia32_roundss((__v4sf)(__m128)(X), \ 276 (__v4sf)(__m128)(Y), (M)) 308 #define _mm_round_pd(X, M) \ 309 (__m128d)__builtin_ia32_roundpd((__v2df)(__m128d)(X), (M)) 349 #define _mm_round_sd(X, Y, M) \ 350 (__m128d)__builtin_ia32_roundsd((__v2df)(__m128d)(X), \ 351 (__v2df)(__m128d)(Y), (M)) 378 #define _mm_blend_pd(V1, V2, M) \ 379 (__m128d) __builtin_ia32_blendpd ((__v2df)(__m128d)(V1), \ 380 (__v2df)(__m128d)(V2), (int)(M)) 406 #define _mm_blend_ps(V1, V2, M) \ 407 (__m128) __builtin_ia32_blendps ((__v4sf)(__m128)(V1), \ 408 (__v4sf)(__m128)(V2), (int)(M)) 433 return (__m128d) __builtin_ia32_blendvpd ((
__v2df)__V1, (
__v2df)__V2,
460 return (__m128) __builtin_ia32_blendvps ((__v4sf)__V1, (__v4sf)__V2,
487 return (__m128i) __builtin_ia32_pblendvb128 ((
__v16qi)__V1, (
__v16qi)__V2,
515 #define _mm_blend_epi16(V1, V2, M) \ 516 (__m128i) __builtin_ia32_pblendw128 ((__v8hi)(__m128i)(V1), \ 517 (__v8hi)(__m128i)(V2), (int)(M)) 556 return (__m128i) __builtin_ia32_pmuldq128 ((
__v4si)__V1, (
__v4si)__V2);
592 #define _mm_dp_ps(X, Y, M) \ 593 (__m128) __builtin_ia32_dpps((__v4sf)(__m128)(X), \ 594 (__v4sf)(__m128)(Y), (M)) 627 #define _mm_dp_pd(X, Y, M) \ 628 (__m128d) __builtin_ia32_dppd((__v2df)(__m128d)(X), \ 629 (__v2df)(__m128d)(Y), (M)) 647 return (__m128i) __builtin_nontemporal_load ((
const __v2di *) __V);
667 return (__m128i) __builtin_ia32_pminsb128 ((
__v16qi) __V1, (
__v16qi) __V2);
686 return (__m128i) __builtin_ia32_pmaxsb128 ((
__v16qi) __V1, (
__v16qi) __V2);
705 return (__m128i) __builtin_ia32_pminuw128 ((
__v8hi) __V1, (
__v8hi) __V2);
724 return (__m128i) __builtin_ia32_pmaxuw128 ((
__v8hi) __V1, (
__v8hi) __V2);
743 return (__m128i) __builtin_ia32_pminsd128 ((
__v4si) __V1, (
__v4si) __V2);
762 return (__m128i) __builtin_ia32_pmaxsd128 ((
__v4si) __V1, (
__v4si) __V2);
781 return (__m128i) __builtin_ia32_pminud128((
__v4si) __V1, (
__v4si) __V2);
800 return (__m128i) __builtin_ia32_pmaxud128((
__v4si) __V1, (
__v4si) __V2);
844 #define _mm_insert_ps(X, Y, N) __builtin_ia32_insertps128((X), (Y), (N)) 868 #define _mm_extract_ps(X, N) (__extension__ \ 869 ({ union { int __i; float __f; } __t; \ 870 __t.__f = __builtin_ia32_vec_ext_v4sf((__v4sf)(__m128)(X), (int)(N)); \ 875 #define _MM_EXTRACT_FLOAT(D, X, N) \ 876 { (D) = __builtin_ia32_vec_ext_v4sf((__v4sf)(__m128)(X), (int)(N)); } 880 #define _MM_MK_INSERTPS_NDX(X, Y, Z) (((X) << 6) | ((Y) << 4) | (Z)) 883 #define _MM_PICK_OUT_PS(X, N) _mm_insert_ps (_mm_setzero_ps(), (X), \ 884 _MM_MK_INSERTPS_NDX((N), 0, 0x0e)) 927 #define _mm_insert_epi8(X, I, N) \ 928 (__m128i)__builtin_ia32_vec_set_v16qi((__v16qi)(__m128i)(X), \ 959 #define _mm_insert_epi32(X, I, N) \ 960 (__m128i)__builtin_ia32_vec_set_v4si((__v4si)(__m128i)(X), \ 990 #define _mm_insert_epi64(X, I, N) \ 991 (__m128i)__builtin_ia32_vec_set_v2di((__v2di)(__m128i)(X), \ 992 (long long)(I), (int)(N)) 1033 #define _mm_extract_epi8(X, N) \ 1034 (int)(unsigned char)__builtin_ia32_vec_ext_v16qi((__v16qi)(__m128i)(X), \ 1059 #define _mm_extract_epi32(X, N) \ 1060 (int)__builtin_ia32_vec_ext_v4si((__v4si)(__m128i)(X), (int)(N)) 1082 #define _mm_extract_epi64(X, N) \ 1083 (long long)__builtin_ia32_vec_ext_v2di((__v2di)(__m128i)(X), (int)(N)) 1102 return __builtin_ia32_ptestz128((
__v2di)__M, (
__v2di)__V);
1120 return __builtin_ia32_ptestc128((
__v2di)__M, (
__v2di)__V);
1139 return __builtin_ia32_ptestnzc128((
__v2di)__M, (
__v2di)__V);
1157 #define _mm_test_all_ones(V) _mm_testc_si128((V), _mm_cmpeq_epi32((V), (V))) 1176 #define _mm_test_mix_ones_zeros(M, V) _mm_testnzc_si128((M), (V)) 1194 #define _mm_test_all_zeros(M, V) _mm_testz_si128 ((M), (V)) 1234 return (__m128i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3, 4, 5, 6, 7),
__v8hi);
1255 return (__m128i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1, 2, 3),
__v4si);
1276 return (__m128i)__builtin_convertvector(__builtin_shufflevector((__v16qs)__V, (__v16qs)__V, 0, 1),
__v2di);
1295 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v8hi)__V, (
__v8hi)__V, 0, 1, 2, 3),
__v4si);
1314 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v8hi)__V, (
__v8hi)__V, 0, 1),
__v2di);
1333 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v4si)__V, (
__v4si)__V, 0, 1),
__v2di);
1353 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v16qu)__V, (
__v16qu)__V, 0, 1, 2, 3, 4, 5, 6, 7),
__v8hi);
1372 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v16qu)__V, (
__v16qu)__V, 0, 1, 2, 3),
__v4si);
1391 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v16qu)__V, (
__v16qu)__V, 0, 1),
__v2di);
1410 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v8hu)__V, (
__v8hu)__V, 0, 1, 2, 3),
__v4si);
1429 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v8hu)__V, (
__v8hu)__V, 0, 1),
__v2di);
1448 return (__m128i)__builtin_convertvector(__builtin_shufflevector((
__v4su)__V, (
__v4su)__V, 0, 1),
__v2di);
1477 return (__m128i) __builtin_ia32_packusdw128((
__v4si)__V1, (
__v4si)__V2);
1516 #define _mm_mpsadbw_epu8(X, Y, M) \ 1517 (__m128i) __builtin_ia32_mpsadbw128((__v16qi)(__m128i)(X), \ 1518 (__v16qi)(__m128i)(Y), (M)) 1536 return (__m128i) __builtin_ia32_phminposuw128((
__v8hi)__V);
1544 #undef __DEFAULT_FN_ATTRS 1545 #define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("sse4.2"))) 1548 #define _SIDD_UBYTE_OPS 0x00 1549 #define _SIDD_UWORD_OPS 0x01 1550 #define _SIDD_SBYTE_OPS 0x02 1551 #define _SIDD_SWORD_OPS 0x03 1554 #define _SIDD_CMP_EQUAL_ANY 0x00 1555 #define _SIDD_CMP_RANGES 0x04 1556 #define _SIDD_CMP_EQUAL_EACH 0x08 1557 #define _SIDD_CMP_EQUAL_ORDERED 0x0c 1560 #define _SIDD_POSITIVE_POLARITY 0x00 1561 #define _SIDD_NEGATIVE_POLARITY 0x10 1562 #define _SIDD_MASKED_POSITIVE_POLARITY 0x20 1563 #define _SIDD_MASKED_NEGATIVE_POLARITY 0x30 1566 #define _SIDD_LEAST_SIGNIFICANT 0x00 1567 #define _SIDD_MOST_SIGNIFICANT 0x40 1570 #define _SIDD_BIT_MASK 0x00 1571 #define _SIDD_UNIT_MASK 0x40 1626 #define _mm_cmpistrm(A, B, M) \ 1627 (__m128i)__builtin_ia32_pcmpistrm128((__v16qi)(__m128i)(A), \ 1628 (__v16qi)(__m128i)(B), (int)(M)) 1680 #define _mm_cmpistri(A, B, M) \ 1681 (int)__builtin_ia32_pcmpistri128((__v16qi)(__m128i)(A), \ 1682 (__v16qi)(__m128i)(B), (int)(M)) 1740 #define _mm_cmpestrm(A, LA, B, LB, M) \ 1741 (__m128i)__builtin_ia32_pcmpestrm128((__v16qi)(__m128i)(A), (int)(LA), \ 1742 (__v16qi)(__m128i)(B), (int)(LB), \ 1799 #define _mm_cmpestri(A, LA, B, LB, M) \ 1800 (int)__builtin_ia32_pcmpestri128((__v16qi)(__m128i)(A), (int)(LA), \ 1801 (__v16qi)(__m128i)(B), (int)(LB), \ 1851 #define _mm_cmpistra(A, B, M) \ 1852 (int)__builtin_ia32_pcmpistria128((__v16qi)(__m128i)(A), \ 1853 (__v16qi)(__m128i)(B), (int)(M)) 1900 #define _mm_cmpistrc(A, B, M) \ 1901 (int)__builtin_ia32_pcmpistric128((__v16qi)(__m128i)(A), \ 1902 (__v16qi)(__m128i)(B), (int)(M)) 1948 #define _mm_cmpistro(A, B, M) \ 1949 (int)__builtin_ia32_pcmpistrio128((__v16qi)(__m128i)(A), \ 1950 (__v16qi)(__m128i)(B), (int)(M)) 1998 #define _mm_cmpistrs(A, B, M) \ 1999 (int)__builtin_ia32_pcmpistris128((__v16qi)(__m128i)(A), \ 2000 (__v16qi)(__m128i)(B), (int)(M)) 2048 #define _mm_cmpistrz(A, B, M) \ 2049 (int)__builtin_ia32_pcmpistriz128((__v16qi)(__m128i)(A), \ 2050 (__v16qi)(__m128i)(B), (int)(M)) 2102 #define _mm_cmpestra(A, LA, B, LB, M) \ 2103 (int)__builtin_ia32_pcmpestria128((__v16qi)(__m128i)(A), (int)(LA), \ 2104 (__v16qi)(__m128i)(B), (int)(LB), \ 2156 #define _mm_cmpestrc(A, LA, B, LB, M) \ 2157 (int)__builtin_ia32_pcmpestric128((__v16qi)(__m128i)(A), (int)(LA), \ 2158 (__v16qi)(__m128i)(B), (int)(LB), \ 2209 #define _mm_cmpestro(A, LA, B, LB, M) \ 2210 (int)__builtin_ia32_pcmpestrio128((__v16qi)(__m128i)(A), (int)(LA), \ 2211 (__v16qi)(__m128i)(B), (int)(LB), \ 2264 #define _mm_cmpestrs(A, LA, B, LB, M) \ 2265 (int)__builtin_ia32_pcmpestris128((__v16qi)(__m128i)(A), (int)(LA), \ 2266 (__v16qi)(__m128i)(B), (int)(LB), \ 2318 #define _mm_cmpestrz(A, LA, B, LB, M) \ 2319 (int)__builtin_ia32_pcmpestriz128((__v16qi)(__m128i)(A), (int)(LA), \ 2320 (__v16qi)(__m128i)(B), (int)(LB), \ 2361 return __builtin_ia32_crc32qi(__C, __D);
2381 return __builtin_ia32_crc32hi(__C, __D);
2401 return __builtin_ia32_crc32si(__C, __D);
2420 _mm_crc32_u64(
unsigned long long __C,
unsigned long long __D)
2422 return __builtin_ia32_crc32di(
__C, __D);
2426 #undef __DEFAULT_FN_ATTRS static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epi32(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [4 x i32] and returns a 128-bit vector ...
__vector signed char __v16qi
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi16_epi64(__m128i __V)
Sign-extends each of the lower two 16-bit integer elements of a 128-bit integer vector of [8 x i16] t...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epu32(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [4 x u32] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi64(__m128i __V1, __m128i __V2)
Compares each of the corresponding 64-bit values of the 128-bit integer vectors for equality...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packus_epi32(__m128i __V1, __m128i __V2)
Converts 32-bit signed integers from both 128-bit integer vector operands into 16-bit unsigned intege...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epi32(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [4 x i32] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_stream_load_si128(__m128i const *__V)
Loads integer values from a 128-bit aligned memory location to a 128-bit integer vector.
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu16_epi64(__m128i __V)
Zero-extends each of the lower two 16-bit integer elements of a 128-bit integer vector of [8 x i16] t...
static __inline__ unsigned int __DEFAULT_FN_ATTRS _mm_crc32_u16(unsigned int __C, unsigned short __D)
Adds the unsigned integer operand to the CRC-32C checksum of the unsigned short operand.
static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_blendv_pd(__m128d __V1, __m128d __V2, __m128d __M)
Returns a 128-bit vector of [2 x double] where the values are selected from either the first or secon...
__vector unsigned char __v16qu
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epi8(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [16 x i8] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi32_epi64(__m128i __V)
Sign-extends each of the lower two 32-bit integer elements of a 128-bit integer vector of [4 x i32] t...
static __inline__ unsigned char int __C
static __inline__ unsigned int __DEFAULT_FN_ATTRS _mm_crc32_u32(unsigned int __C, unsigned int __D)
Adds the first unsigned integer operand to the CRC-32C checksum of the second unsigned integer operan...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mullo_epi32(__m128i __V1, __m128i __V2)
Multiples corresponding elements of two 128-bit vectors of [4 x i32] and returns the lower 32 bits of...
static __inline__ int __DEFAULT_FN_ATTRS _mm_testz_si128(__m128i __M, __m128i __V)
Tests whether the specified bits in a 128-bit integer vector are all zeros.
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_blendv_epi8(__m128i __V1, __m128i __V2, __m128i __M)
Returns a 128-bit vector of [16 x i8] where the values are selected from either of the first or secon...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi64(__m128i __V1, __m128i __V2)
Compares each of the corresponding 64-bit values of the 128-bit integer vectors to determine if the v...
static __inline__ int __DEFAULT_FN_ATTRS _mm_testnzc_si128(__m128i __M, __m128i __V)
Tests whether the specified bits in a 128-bit integer vector are neither all zeros nor all ones...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu8_epi16(__m128i __V)
Zero-extends each of the lower eight 8-bit integer elements of a 128-bit vector of [16 x i8] to 16-bi...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epu16(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [8 x u16] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epu16(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [8 x u16] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mul_epi32(__m128i __V1, __m128i __V2)
Multiplies corresponding even-indexed elements of two 128-bit vectors of [4 x i32] and returns a 128-...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu8_epi64(__m128i __V)
Zero-extends each of the lower two 8-bit integer elements of a 128-bit integer vector of [16 x i8] to...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_blendv_ps(__m128 __V1, __m128 __V2, __m128 __M)
Returns a 128-bit vector of [4 x float] where the values are selected from either the first or second...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epi8(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [16 x i8] and returns a 128-bit vector ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu16_epi32(__m128i __V)
Zero-extends each of the lower four 16-bit integer elements of a 128-bit integer vector of [8 x i16] ...
static __inline__ unsigned int __DEFAULT_FN_ATTRS _mm_crc32_u8(unsigned int __C, unsigned char __D)
Adds the unsigned integer operand to the CRC-32C checksum of the unsigned char operand.
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi8_epi64(__m128i __V)
Sign-extends each of the lower two 8-bit integer elements of a 128-bit integer vector of [16 x i8] to...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_minpos_epu16(__m128i __V)
Finds the minimum unsigned 16-bit element in the input 128-bit vector of [8 x u16] and returns it and...
static __inline__ unsigned int unsigned char __D
__vector unsigned int __v4su
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epu32(__m128i __V1, __m128i __V2)
Compares the corresponding elements of two 128-bit vectors of [4 x u32] and returns a 128-bit vector ...
#define __DEFAULT_FN_ATTRS
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi16_epi32(__m128i __V)
Sign-extends each of the lower four 16-bit integer elements of a 128-bit integer vector of [8 x i16] ...
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi8_epi32(__m128i __V)
Sign-extends each of the lower four 8-bit integer elements of a 128-bit vector of [16 x i8] to 32-bit...
__vector long long __v2di
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu32_epi64(__m128i __V)
Zero-extends each of the lower two 32-bit integer elements of a 128-bit integer vector of [4 x i32] t...
static __inline__ int __DEFAULT_FN_ATTRS _mm_testc_si128(__m128i __M, __m128i __V)
Tests whether the specified bits in a 128-bit integer vector are all ones.
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepu8_epi32(__m128i __V)
Zero-extends each of the lower four 8-bit integer elements of a 128-bit vector of [16 x i8] to 32-bit...
__vector unsigned short __v8hu
static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtepi8_epi16(__m128i __V)
Sign-extends each of the lower eight 8-bit integer elements of a 128-bit vector of [16 x i8] to 16-bi...