13#if !defined(__i386__) && !defined(__x86_64__)
14#error "This header is only meant to be used on x86 and x64 architecture"
21typedef float __m128
__attribute__((__vector_size__(16), __aligned__(16)));
23typedef float __m128_u
__attribute__((__vector_size__(16), __aligned__(1)));
26typedef unsigned int __v4su
__attribute__((__vector_size__(16)));
35#define __DEFAULT_FN_ATTRS \
36 __attribute__((__always_inline__, __nodebug__, __target__("sse,no-evex512"), \
37 __min_vector_width__(128)))
38#define __DEFAULT_FN_ATTRS_MMX \
39 __attribute__((__always_inline__, __nodebug__, \
40 __target__("mmx,sse,no-evex512"), __min_vector_width__(64)))
80 return (__m128)((__v4sf)
__a + (__v4sf)
__b);
123 return (__m128)((__v4sf)
__a - (__v4sf)
__b);
165 return (__m128)((__v4sf)
__a * (__v4sf)
__b);
206 return (__m128)((__v4sf)
__a / (__v4sf)
__b);
224 return (__m128)__builtin_ia32_sqrtss((__v4sf)
__a);
241 return __builtin_ia32_sqrtps((__v4sf)
__a);
259 return (__m128)__builtin_ia32_rcpss((__v4sf)
__a);
276 return (__m128)__builtin_ia32_rcpps((__v4sf)
__a);
295 return __builtin_ia32_rsqrtss((__v4sf)
__a);
312 return __builtin_ia32_rsqrtps((__v4sf)
__a);
337 return __builtin_ia32_minss((__v4sf)
__a, (__v4sf)
__b);
358 return __builtin_ia32_minps((__v4sf)
__a, (__v4sf)
__b);
383 return __builtin_ia32_maxss((__v4sf)
__a, (__v4sf)
__b);
404 return __builtin_ia32_maxps((__v4sf)
__a, (__v4sf)
__b);
422 return (__m128)((__v4su)
__a & (__v4su)
__b);
444 return (__m128)(~(__v4su)
__a & (__v4su)
__b);
462 return (__m128)((__v4su)
__a | (__v4su)
__b);
481 return (__m128)((__v4su)
__a ^ (__v4su)
__b);
506 return (__m128)__builtin_ia32_cmpeqss((__v4sf)
__a, (__v4sf)
__b);
527 return (__m128)__builtin_ia32_cmpeqps((__v4sf)
__a, (__v4sf)
__b);
553 return (__m128)__builtin_ia32_cmpltss((__v4sf)
__a, (__v4sf)
__b);
575 return (__m128)__builtin_ia32_cmpltps((__v4sf)
__a, (__v4sf)
__b);
601 return (__m128)__builtin_ia32_cmpless((__v4sf)
__a, (__v4sf)
__b);
623 return (__m128)__builtin_ia32_cmpleps((__v4sf)
__a, (__v4sf)
__b);
649 return (__m128)__builtin_shufflevector((__v4sf)
__a,
650 (__v4sf)__builtin_ia32_cmpltss((__v4sf)
__b, (__v4sf)
__a),
673 return (__m128)__builtin_ia32_cmpltps((__v4sf)
__b, (__v4sf)
__a);
699 return (__m128)__builtin_shufflevector((__v4sf)
__a,
700 (__v4sf)__builtin_ia32_cmpless((__v4sf)
__b, (__v4sf)
__a),
723 return (__m128)__builtin_ia32_cmpleps((__v4sf)
__b, (__v4sf)
__a);
749 return (__m128)__builtin_ia32_cmpneqss((__v4sf)
__a, (__v4sf)
__b);
771 return (__m128)__builtin_ia32_cmpneqps((__v4sf)
__a, (__v4sf)
__b);
798 return (__m128)__builtin_ia32_cmpnltss((__v4sf)
__a, (__v4sf)
__b);
821 return (__m128)__builtin_ia32_cmpnltps((__v4sf)
__a, (__v4sf)
__b);
848 return (__m128)__builtin_ia32_cmpnless((__v4sf)
__a, (__v4sf)
__b);
871 return (__m128)__builtin_ia32_cmpnleps((__v4sf)
__a, (__v4sf)
__b);
898 return (__m128)__builtin_shufflevector((__v4sf)
__a,
899 (__v4sf)__builtin_ia32_cmpnltss((__v4sf)
__b, (__v4sf)
__a),
923 return (__m128)__builtin_ia32_cmpnltps((__v4sf)
__b, (__v4sf)
__a);
950 return (__m128)__builtin_shufflevector((__v4sf)
__a,
951 (__v4sf)__builtin_ia32_cmpnless((__v4sf)
__b, (__v4sf)
__a),
975 return (__m128)__builtin_ia32_cmpnleps((__v4sf)
__b, (__v4sf)
__a);
1002 return (__m128)__builtin_ia32_cmpordss((__v4sf)
__a, (__v4sf)
__b);
1026 return (__m128)__builtin_ia32_cmpordps((__v4sf)
__a, (__v4sf)
__b);
1053 return (__m128)__builtin_ia32_cmpunordss((__v4sf)
__a, (__v4sf)
__b);
1077 return (__m128)__builtin_ia32_cmpunordps((__v4sf)
__a, (__v4sf)
__b);
1101 return __builtin_ia32_comieq((__v4sf)
__a, (__v4sf)
__b);
1126 return __builtin_ia32_comilt((__v4sf)
__a, (__v4sf)
__b);
1150 return __builtin_ia32_comile((__v4sf)
__a, (__v4sf)
__b);
1174 return __builtin_ia32_comigt((__v4sf)
__a, (__v4sf)
__b);
1198 return __builtin_ia32_comige((__v4sf)
__a, (__v4sf)
__b);
1222 return __builtin_ia32_comineq((__v4sf)
__a, (__v4sf)
__b);
1245 return __builtin_ia32_ucomieq((__v4sf)
__a, (__v4sf)
__b);
1269 return __builtin_ia32_ucomilt((__v4sf)
__a, (__v4sf)
__b);
1293 return __builtin_ia32_ucomile((__v4sf)
__a, (__v4sf)
__b);
1317 return __builtin_ia32_ucomigt((__v4sf)
__a, (__v4sf)
__b);
1341 return __builtin_ia32_ucomige((__v4sf)
__a, (__v4sf)
__b);
1364 return __builtin_ia32_ucomineq((__v4sf)
__a, (__v4sf)
__b);
1386 return __builtin_ia32_cvtss2si((__v4sf)
__a);
1430_mm_cvtss_si64(__m128
__a)
1432 return __builtin_ia32_cvtss2si64((__v4sf)
__a);
1454 return (__m64)__builtin_ia32_cvtps2pi((__v4sf)
__a);
1496 return __builtin_ia32_cvttss2si((__v4sf)
__a);
1539_mm_cvttss_si64(__m128
__a)
1541 return __builtin_ia32_cvttss2si64((__v4sf)
__a);
1564 return (__m64)__builtin_ia32_cvttps2pi((__v4sf)
__a);
1652_mm_cvtsi64_ss(__m128
__a,
long long __b)
1680 return __builtin_ia32_cvtpi2ps((__v4sf)
__a, (__v2si)
__b);
1741 typedef float __mm_loadh_pi_v2f32
__attribute__((__vector_size__(8)));
1742 struct __mm_loadh_pi_struct {
1743 __mm_loadh_pi_v2f32 __u;
1745 __mm_loadh_pi_v2f32
__b = ((
const struct __mm_loadh_pi_struct*)
__p)->__u;
1746 __m128 __bb = __builtin_shufflevector(
__b,
__b, 0, 1, 0, 1);
1747 return __builtin_shufflevector(
__a, __bb, 0, 1, 4, 5);
1768 typedef float __mm_loadl_pi_v2f32
__attribute__((__vector_size__(8)));
1769 struct __mm_loadl_pi_struct {
1770 __mm_loadl_pi_v2f32 __u;
1772 __mm_loadl_pi_v2f32
__b = ((
const struct __mm_loadl_pi_struct*)
__p)->__u;
1773 __m128 __bb = __builtin_shufflevector(
__b,
__b, 0, 1, 0, 1);
1774 return __builtin_shufflevector(
__a, __bb, 4, 5, 2, 3);
1795 struct __mm_load_ss_struct {
1798 float __u = ((
const struct __mm_load_ss_struct*)
__p)->__u;
1799 return __extension__ (__m128){ __u, 0, 0, 0 };
1817 struct __mm_load1_ps_struct {
1820 float __u = ((
const struct __mm_load1_ps_struct*)
__p)->__u;
1821 return __extension__ (__m128){ __u, __u, __u, __u };
1824#define _mm_load_ps1(p) _mm_load1_ps(p)
1840 return *(
const __m128*)
__p;
1860 return ((
const struct __loadu_ps*)
__p)->__v;
1880 return __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__a, 3, 2, 1, 0);
1893 return (__m128)__builtin_ia32_undef128();
1913 return __extension__ (__m128){ __w, 0, 0, 0 };
1931 return __extension__ (__m128){ __w, __w, __w, __w };
1977 return __extension__ (__m128){ __w, __x,
__y, __z };
2005 return __extension__ (__m128){ __z,
__y, __x, __w };
2020 return __extension__ (__m128){ 0.0f, 0.0f, 0.0f, 0.0f };
2037 typedef float __mm_storeh_pi_v2f32
__attribute__((__vector_size__(8)));
2038 struct __mm_storeh_pi_struct {
2039 __mm_storeh_pi_v2f32 __u;
2041 ((
struct __mm_storeh_pi_struct*)
__p)->__u = __builtin_shufflevector(
__a,
__a, 2, 3);
2058 typedef float __mm_storeh_pi_v2f32
__attribute__((__vector_size__(8)));
2059 struct __mm_storeh_pi_struct {
2060 __mm_storeh_pi_v2f32 __u;
2062 ((
struct __mm_storeh_pi_struct*)
__p)->__u = __builtin_shufflevector(
__a,
__a, 0, 1);
2079 struct __mm_store_ss_struct {
2082 ((
struct __mm_store_ss_struct*)
__p)->__u =
__a[0];
2100 struct __storeu_ps {
2103 ((
struct __storeu_ps*)
__p)->__v =
__a;
2140 __a = __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__a, 0, 0, 0, 0);
2179 __a = __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__a, 3, 2, 1, 0);
2183#define _MM_HINT_ET0 7
2184#define _MM_HINT_ET1 6
2185#define _MM_HINT_T0 3
2186#define _MM_HINT_T1 2
2187#define _MM_HINT_T2 1
2188#define _MM_HINT_NTA 0
2218#define _mm_prefetch(a, sel) (__builtin_prefetch((const void *)(a), \
2219 ((sel) >> 2) & 1, (sel) & 0x3))
2237 __builtin_ia32_movntq((__m64 *)
__p,
__a);
2256 __builtin_nontemporal_store((__v4sf)
__a, (__v4sf*)
__p);
2259#if defined(__cplusplus)
2274#if defined(__cplusplus)
2298#define _mm_extract_pi16(a, n) \
2299 ((int)__builtin_ia32_vec_ext_v4hi((__v4hi)a, (int)n))
2329#define _mm_insert_pi16(a, d, n) \
2330 ((__m64)__builtin_ia32_vec_set_v4hi((__v4hi)a, (int)d, (int)n))
2348 return (__m64)__builtin_ia32_pmaxsw((__v4hi)
__a, (__v4hi)
__b);
2367 return (__m64)__builtin_ia32_pmaxub((__v8qi)
__a, (__v8qi)
__b);
2386 return (__m64)__builtin_ia32_pminsw((__v4hi)
__a, (__v4hi)
__b);
2405 return (__m64)__builtin_ia32_pminub((__v8qi)
__a, (__v8qi)
__b);
2423 return __builtin_ia32_pmovmskb((__v8qi)
__a);
2442 return (__m64)__builtin_ia32_pmulhuw((__v4hi)
__a, (__v4hi)
__b);
2479#define _mm_shuffle_pi16(a, n) \
2480 ((__m64)__builtin_ia32_pshufw((__v4hi)(__m64)(a), (n)))
2508 __builtin_ia32_maskmovq((__v8qi)__d, (__v8qi)__n,
__p);
2527 return (__m64)__builtin_ia32_pavgb((__v8qi)
__a, (__v8qi)
__b);
2546 return (__m64)__builtin_ia32_pavgw((__v4hi)
__a, (__v4hi)
__b);
2568 return (__m64)__builtin_ia32_psadbw((__v8qi)
__a, (__v8qi)
__b);
2571#if defined(__cplusplus)
2681#if defined(__cplusplus)
2724#define _mm_shuffle_ps(a, b, mask) \
2725 ((__m128)__builtin_ia32_shufps((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), \
2747 return __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__b, 2, 6, 3, 7);
2769 return __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__b, 0, 4, 1, 5);
2813 return __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__b, 6, 7, 2, 3);
2834 return __builtin_shufflevector((__v4sf)
__a, (__v4sf)
__b, 0, 1, 4, 5);
3047 return __builtin_ia32_movmskps((__v4sf)
__a);
3051#define _CMP_EQ_OQ 0x00
3052#define _CMP_LT_OS 0x01
3053#define _CMP_LE_OS 0x02
3054#define _CMP_UNORD_Q 0x03
3055#define _CMP_NEQ_UQ 0x04
3056#define _CMP_NLT_US 0x05
3057#define _CMP_NLE_US 0x06
3058#define _CMP_ORD_Q 0x07
3092#define _mm_cmp_ps(a, b, c) \
3093 ((__m128)__builtin_ia32_cmpps((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), (c)))
3127#define _mm_cmp_ss(a, b, c) \
3128 ((__m128)__builtin_ia32_cmpss((__v4sf)(__m128)(a), (__v4sf)(__m128)(b), (c)))
3130#define _MM_ALIGN16 __attribute__((aligned(16)))
3132#define _MM_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w))
3134#define _MM_EXCEPT_INVALID (0x0001U)
3135#define _MM_EXCEPT_DENORM (0x0002U)
3136#define _MM_EXCEPT_DIV_ZERO (0x0004U)
3137#define _MM_EXCEPT_OVERFLOW (0x0008U)
3138#define _MM_EXCEPT_UNDERFLOW (0x0010U)
3139#define _MM_EXCEPT_INEXACT (0x0020U)
3140#define _MM_EXCEPT_MASK (0x003fU)
3142#define _MM_MASK_INVALID (0x0080U)
3143#define _MM_MASK_DENORM (0x0100U)
3144#define _MM_MASK_DIV_ZERO (0x0200U)
3145#define _MM_MASK_OVERFLOW (0x0400U)
3146#define _MM_MASK_UNDERFLOW (0x0800U)
3147#define _MM_MASK_INEXACT (0x1000U)
3148#define _MM_MASK_MASK (0x1f80U)
3150#define _MM_ROUND_NEAREST (0x0000U)
3151#define _MM_ROUND_DOWN (0x2000U)
3152#define _MM_ROUND_UP (0x4000U)
3153#define _MM_ROUND_TOWARD_ZERO (0x6000U)
3154#define _MM_ROUND_MASK (0x6000U)
3156#define _MM_FLUSH_ZERO_MASK (0x8000U)
3157#define _MM_FLUSH_ZERO_ON (0x8000U)
3158#define _MM_FLUSH_ZERO_OFF (0x0000U)
3160#define _MM_GET_EXCEPTION_MASK() (_mm_getcsr() & _MM_MASK_MASK)
3161#define _MM_GET_EXCEPTION_STATE() (_mm_getcsr() & _MM_EXCEPT_MASK)
3162#define _MM_GET_FLUSH_ZERO_MODE() (_mm_getcsr() & _MM_FLUSH_ZERO_MASK)
3163#define _MM_GET_ROUNDING_MODE() (_mm_getcsr() & _MM_ROUND_MASK)
3165#define _MM_SET_EXCEPTION_MASK(x) (_mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | (x)))
3166#define _MM_SET_EXCEPTION_STATE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | (x)))
3167#define _MM_SET_FLUSH_ZERO_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | (x)))
3168#define _MM_SET_ROUNDING_MODE(x) (_mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | (x)))
3170#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3) \
3172 __m128 tmp3, tmp2, tmp1, tmp0; \
3173 tmp0 = _mm_unpacklo_ps((row0), (row1)); \
3174 tmp2 = _mm_unpacklo_ps((row2), (row3)); \
3175 tmp1 = _mm_unpackhi_ps((row0), (row1)); \
3176 tmp3 = _mm_unpackhi_ps((row2), (row3)); \
3177 (row0) = _mm_movelh_ps(tmp0, tmp2); \
3178 (row1) = _mm_movehl_ps(tmp2, tmp0); \
3179 (row2) = _mm_movelh_ps(tmp1, tmp3); \
3180 (row3) = _mm_movehl_ps(tmp3, tmp1); \
3184#define _m_pextrw _mm_extract_pi16
3185#define _m_pinsrw _mm_insert_pi16
3186#define _m_pmaxsw _mm_max_pi16
3187#define _m_pmaxub _mm_max_pu8
3188#define _m_pminsw _mm_min_pi16
3189#define _m_pminub _mm_min_pu8
3190#define _m_pmovmskb _mm_movemask_pi8
3191#define _m_pmulhuw _mm_mulhi_pu16
3192#define _m_pshufw _mm_shuffle_pi16
3193#define _m_maskmovq _mm_maskmove_si64
3194#define _m_pavgb _mm_avg_pu8
3195#define _m_pavgw _mm_avg_pu16
3196#define _m_psadbw _mm_sad_pu8
3199#undef __DEFAULT_FN_ATTRS
3200#undef __DEFAULT_FN_ATTRS_MMX
3203#if defined(__SSE2__) && !__building_module(_Builtin_intrinsics)
3204#include <emmintrin.h>
_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 vector float __c
static __inline__ vector float vector float __b
static __inline__ uint32_t volatile uint32_t * __p
static __inline__ uint32_t uint32_t __y
static __inline__ void int __a
struct __storeu_i16 *__P __v
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_unpackhi_pi16(__m64 __m1, __m64 __m2)
Unpacks the upper 32 bits from two 64-bit integer vectors of [4 x i16] and interleaves them into a 64...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_unpacklo_pi8(__m64 __m1, __m64 __m2)
Unpacks the lower 32 bits from two 64-bit integer vectors of [8 x i8] and interleaves them into a 64-...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_cmpgt_pi16(__m64 __m1, __m64 __m2)
Compares the 16-bit integer elements of two 64-bit integer vectors of [4 x i16] to determine if the e...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_cmpgt_pi8(__m64 __m1, __m64 __m2)
Compares the 8-bit integer elements of two 64-bit integer vectors of [8 x i8] to determine if the ele...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_packs_pi32(__m64 __m1, __m64 __m2)
Converts, with saturation, 32-bit signed integers from both 64-bit integer vector parameters of [2 x ...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_unpacklo_pi16(__m64 __m1, __m64 __m2)
Unpacks the lower 32 bits from two 64-bit integer vectors of [4 x i16] and interleaves them into a 64...
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_setzero_si64(void)
Constructs a 64-bit integer vector initialized to zero.
static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_packs_pi16(__m64 __m1, __m64 __m2)
Converts, with saturation, 16-bit signed integers from both 64-bit integer vector parameters of [4 x ...
static __inline__ int __DEFAULT_FN_ATTRS _mm_comigt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the first ope...
static __inline__ int __DEFAULT_FN_ATTRS _mm_cvttss_si32(__m128 __a)
Converts the lower (first) element of a vector of [4 x float] into a signed truncated (rounded toward...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_rcp_ss(__m128 __a)
Calculates the approximate reciprocal of the value stored in the low-order bits of a 128-bit vector o...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmplt_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_sqrt_ss(__m128 __a)
Calculates the square root of the value stored in the low-order bits of a 128-bit vector of [4 x floa...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_div_ps(__m128 __a, __m128 __b)
Divides two 128-bit vectors of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnge_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpeq_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] for equa...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_setzero_ps(void)
Constructs a 128-bit floating-point vector of [4 x float] initialized to zero.
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtps_pi32(__m128 __a)
Converts two low-order float values in a 128-bit vector of [4 x float] into a 64-bit vector of [2 x i...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_set_ss(float __w)
Constructs a 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_and_ps(__m128 __a, __m128 __b)
Performs a bitwise AND of two 128-bit vectors of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmplt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ int __DEFAULT_FN_ATTRS _mm_cvt_ss2si(__m128 __a)
Converts a float value contained in the lower 32 bits of a vector of [4 x float] into a 32-bit intege...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpeq_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands for equality.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_add_ps(__m128 __a, __m128 __b)
Adds two 128-bit vectors of [4 x float], and returns the results of the addition.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_load_ps(const float *__p)
Loads a 128-bit floating-point vector of [4 x float] from an aligned memory location.
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_mulhi_pu16(__m64 __a, __m64 __b)
Multiplies packed 16-bit unsigned integer values and writes the high-order 16 bits of each 32-bit pro...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpneq_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] for ineq...
static __inline__ int __DEFAULT_FN_ATTRS _mm_comile_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the first ope...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvt_ps2pi(__m128 __a)
Converts two low-order float values in a 128-bit vector of [4 x float] into a 64-bit vector of [2 x i...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_andnot_ps(__m128 __a, __m128 __b)
Performs a bitwise AND of two 128-bit vectors of [4 x float], using the one's complement of the value...
static __inline__ void __DEFAULT_FN_ATTRS _mm_storer_ps(float *__p, __m128 __a)
Stores float values from a 128-bit vector of [4 x float] to an aligned memory location in reverse ord...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_undefined_ps(void)
Create a 128-bit vector of [4 x float] with undefined values.
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvttps_pi32(__m128 __a)
Converts the lower (first) two elements of a 128-bit vector of [4 x float] into two signed truncated ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnle_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomilt_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmple_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvt_si2ss(__m128 __a, int __b)
Converts a 32-bit signed integer value into a floating point value and writes it to the lower 32 bits...
static __inline__ void __DEFAULT_FN_ATTRS_MMX _mm_stream_pi(void *__p, __m64 __a)
Stores a 64-bit integer in the specified aligned memory location.
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpi16_ps(__m64 __a)
Converts a 64-bit vector of [4 x i16] into a 128-bit vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_rsqrt_ps(__m128 __a)
Calculates the approximate reciprocals of the square roots of the values stored in a 128-bit vector o...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtps_pi8(__m128 __a)
Converts each single-precision floating-point element of a 128-bit floating-point vector of [4 x floa...
static __inline__ void __DEFAULT_FN_ATTRS _mm_storel_pi(__m64 *__p, __m128 __a)
Stores the lower 64 bits of a 128-bit vector of [4 x float] to a memory location.
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomile_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpge_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ int __DEFAULT_FN_ATTRS _mm_comieq_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands for equality.
static __inline__ void __DEFAULT_FN_ATTRS _mm_store1_ps(float *__p, __m128 __a)
Stores the lower 32 bits of a 128-bit vector of [4 x float] into four contiguous elements in an align...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpu16_ps(__m64 __a)
Converts a 64-bit vector of 16-bit unsigned integer values into a 128-bit vector of [4 x float].
void _mm_sfence(void)
Forces strong memory ordering (serialization) between store instructions preceding this instruction a...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_set_ps1(float __w)
Constructs a 128-bit floating-point vector of [4 x float], with each of the four single-precision flo...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_mul_ps(__m128 __a, __m128 __b)
Multiplies two 128-bit vectors of [4 x float] and returns the results of the multiplication.
#define __DEFAULT_FN_ATTRS
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_max_ps(__m128 __a, __m128 __b)
Compares two 128-bit vectors of [4 x float] and returns the greater of each pair of values.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_rsqrt_ss(__m128 __a)
Calculates the approximate reciprocal of the square root of the value stored in the low-order bits of...
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomige_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_avg_pu16(__m64 __a, __m64 __b)
Computes the rounded averages of the packed unsigned 16-bit integer values and writes the averages to...
static __inline__ int __DEFAULT_FN_ATTRS _mm_comilt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the first ope...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_loadl_pi(__m128 __a, const __m64 *__p)
Loads two packed float values from the address __p into the low-order bits of a 128-bit vector of [4 ...
static __inline__ void __DEFAULT_FN_ATTRS _mm_storeu_ps(float *__p, __m128 __a)
Stores a 128-bit vector of [4 x float] to an unaligned memory location.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_sub_ss(__m128 __a, __m128 __b)
Subtracts the 32-bit float value in the low-order bits of the second operand from the corresponding v...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_sub_ps(__m128 __a, __m128 __b)
Subtracts each of the values of the second operand from the first operand, both of which are 128-bit ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_load1_ps(const float *__p)
Loads a 32-bit float value and duplicates it to all four vector elements of a 128-bit vector of [4 x ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_movelh_ps(__m128 __a, __m128 __b)
Constructs a 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_min_ps(__m128 __a, __m128 __b)
Compares two 128-bit vectors of [4 x float] and returns the lesser of each pair of values.
static __inline__ void __DEFAULT_FN_ATTRS _mm_stream_ps(void *__p, __m128 __a)
Moves packed float values from a 128-bit vector of [4 x float] to a 128-bit aligned memory location.
static __inline__ int __DEFAULT_FN_ATTRS _mm_comige_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the first ope...
static __inline__ int __DEFAULT_FN_ATTRS _mm_cvtss_si32(__m128 __a)
Converts a float value contained in the lower 32 bits of a vector of [4 x float] into a 32-bit intege...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpgt_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_setr_ps(float __z, float __y, float __x, float __w)
Constructs a 128-bit floating-point vector of [4 x float], initialized in reverse order with the spec...
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomigt_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_unpackhi_ps(__m128 __a, __m128 __b)
Unpacks the high-order (index 2,3) values from two 128-bit vectors of [4 x float] and interleaves the...
static __inline__ void __DEFAULT_FN_ATTRS _mm_store_ss(float *__p, __m128 __a)
Stores the lower 32 bits of a 128-bit vector of [4 x float] to a memory location.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpngt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_loadh_pi(__m128 __a, const __m64 *__p)
Loads two packed float values from the address __p into the high-order bits of a 128-bit vector of [4...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_xor_ps(__m128 __a, __m128 __b)
Performs a bitwise exclusive OR of two 128-bit vectors of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_rcp_ps(__m128 __a)
Calculates the approximate reciprocals of the values stored in a 128-bit vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_move_ss(__m128 __a, __m128 __b)
Constructs a 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_set1_ps(float __w)
Constructs a 128-bit floating-point vector of [4 x float], with each of the four single-precision flo...
static __inline__ void __DEFAULT_FN_ATTRS _mm_store_ps(float *__p, __m128 __a)
Stores a 128-bit vector of [4 x float] into an aligned memory location.
void _mm_setcsr(unsigned int __i)
Sets the MXCSR register with the 32-bit unsigned integer value.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_or_ps(__m128 __a, __m128 __b)
Performs a bitwise OR of two 128-bit vectors of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_sqrt_ps(__m128 __a)
Calculates the square roots of the values stored in a 128-bit vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpneq_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands for inequality.
static __inline__ float __DEFAULT_FN_ATTRS _mm_cvtss_f32(__m128 __a)
Extracts a float value contained in the lower 32 bits of a vector of [4 x float].
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_max_pu8(__m64 __a, __m64 __b)
Compares each of the corresponding packed 8-bit unsigned integer values of the 64-bit integer vectors...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_mul_ss(__m128 __a, __m128 __b)
Multiplies two 32-bit float values in the low-order bits of the operands.
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_min_pi16(__m64 __a, __m64 __b)
Compares each of the corresponding packed 16-bit integer values of the 64-bit integer vectors,...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtsi32_ss(__m128 __a, int __b)
Converts a 32-bit signed integer value into a floating point value and writes it to the lower 32 bits...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtt_ps2pi(__m128 __a)
Converts the lower (first) two elements of a 128-bit vector of [4 x float] into two signed truncated ...
static __inline__ int __DEFAULT_FN_ATTRS _mm_cvtt_ss2si(__m128 __a)
Converts the lower (first) element of a vector of [4 x float] into a signed truncated (rounded toward...
static __inline__ int __DEFAULT_FN_ATTRS _mm_movemask_ps(__m128 __a)
Extracts the sign bits from each single-precision floating-point element of a 128-bit floating-point ...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
Converts the two 32-bit signed integer values from each 64-bit vector operand of [2 x i32] into a 128...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_movehl_ps(__m128 __a, __m128 __b)
Constructs a 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_loadr_ps(const float *__p)
Loads four packed float values, in reverse order, from an aligned memory location to 32-bit elements ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpord_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnlt_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ void __DEFAULT_FN_ATTRS _mm_storeh_pi(__m64 *__p, __m128 __a)
Stores the upper 64 bits of a 128-bit vector of [4 x float] to a memory location.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpngt_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnge_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpord_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpgt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvt_pi2ps(__m128 __a, __m64 __b)
Converts two elements of a 64-bit vector of [2 x i32] into two floating point values and writes them ...
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomieq_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_add_ss(__m128 __a, __m128 __b)
Adds the 32-bit float values in the low-order bits of the operands.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_set_ps(float __z, float __y, float __x, float __w)
Constructs a 128-bit floating-point vector of [4 x float] initialized with the specified single-preci...
static __inline__ int __DEFAULT_FN_ATTRS_MMX _mm_movemask_pi8(__m64 __a)
Takes the most significant bit from each 8-bit element in a 64-bit integer vector to create an 8-bit ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnlt_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ void __DEFAULT_FN_ATTRS _mm_store_ps1(float *__p, __m128 __a)
Stores the lower 32 bits of a 128-bit vector of [4 x float] into four contiguous elements in an align...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_min_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands and returns the lesser value ...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpu8_ps(__m64 __a)
Converts the lower four unsigned 8-bit integer values from a 64-bit vector of [8 x u8] into a 128-bit...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmple_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
unsigned int _mm_getcsr(void)
Returns the contents of the MXCSR register as a 32-bit unsigned integer value.
#define __DEFAULT_FN_ATTRS_MMX
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_avg_pu8(__m64 __a, __m64 __b)
Computes the rounded averages of the packed unsigned 8-bit integer values and writes the averages to ...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpi8_ps(__m64 __a)
Converts the lower four 8-bit values from a 64-bit vector of [8 x i8] into a 128-bit vector of [4 x f...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_unpacklo_ps(__m128 __a, __m128 __b)
Unpacks the low-order (index 0,1) values from two 128-bit vectors of [4 x float] and interleaves them...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_max_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands and returns the greater value...
static __inline__ void __DEFAULT_FN_ATTRS_MMX _mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
Conditionally copies the values from each 8-bit element in the first 64-bit integer vector operand to...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_max_pi16(__m64 __a, __m64 __b)
Compares each of the corresponding packed 16-bit integer values of the 64-bit integer vectors,...
static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomineq_ss(__m128 __a, __m128 __b)
Performs an unordered comparison of two 32-bit float values using the low-order bits of both operands...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtps_pi16(__m128 __a)
Converts each single-precision floating-point element of a 128-bit floating-point vector of [4 x floa...
static __inline__ int __DEFAULT_FN_ATTRS _mm_comineq_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the first ope...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_min_pu8(__m64 __a, __m64 __b)
Compares each of the corresponding packed 8-bit unsigned integer values of the 64-bit integer vectors...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpnle_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_sad_pu8(__m64 __a, __m64 __b)
Subtracts the corresponding 8-bit unsigned integer values of the two 64-bit vector operands and compu...
static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX _mm_cvtpi32_ps(__m128 __a, __m64 __b)
Converts two elements of a 64-bit vector of [2 x i32] into two floating point values and writes them ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpunord_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpge_ss(__m128 __a, __m128 __b)
Compares two 32-bit float values in the low-order bits of both operands to determine if the value in ...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cmpunord_ps(__m128 __a, __m128 __b)
Compares each of the corresponding 32-bit float values of the 128-bit vectors of [4 x float] to deter...
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_loadu_ps(const float *__p)
Loads a 128-bit floating-point vector of [4 x float] from an unaligned memory location.
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_load_ss(const float *__p)
Constructs a 128-bit floating-point vector of [4 x float].
static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_div_ss(__m128 __a, __m128 __b)
Divides the value in the low-order 32 bits of the first operand by the corresponding value in the sec...