smmintrin.h File Reference

#include <tmmintrin.h>

Go to the source code of this file.

Macros
#define	__DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("sse4.1")))
#define	_MM_FROUND_TO_NEAREST_INT   0x00
#define	_MM_FROUND_TO_NEG_INF   0x01
#define	_MM_FROUND_TO_POS_INF   0x02
#define	_MM_FROUND_TO_ZERO   0x03
#define	_MM_FROUND_CUR_DIRECTION   0x04
#define	_MM_FROUND_RAISE_EXC   0x00
#define	_MM_FROUND_NO_EXC   0x08
#define	_MM_FROUND_NINT   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEAREST_INT)
#define	_MM_FROUND_FLOOR   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEG_INF)
#define	_MM_FROUND_CEIL   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_POS_INF)
#define	_MM_FROUND_TRUNC   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_ZERO)
#define	_MM_FROUND_RINT   (_MM_FROUND_RAISE_EXC | _MM_FROUND_CUR_DIRECTION)
#define	_MM_FROUND_NEARBYINT   (_MM_FROUND_NO_EXC | _MM_FROUND_CUR_DIRECTION)
#define	_mm_ceil_ps(X)   _mm_round_ps((X), _MM_FROUND_CEIL)
#define	_mm_ceil_pd(X)   _mm_round_pd((X), _MM_FROUND_CEIL)
#define	_mm_ceil_ss(X, Y)   _mm_round_ss((X), (Y), _MM_FROUND_CEIL)
#define	_mm_ceil_sd(X, Y)   _mm_round_sd((X), (Y), _MM_FROUND_CEIL)
#define	_mm_floor_ps(X)   _mm_round_ps((X), _MM_FROUND_FLOOR)
#define	_mm_floor_pd(X)   _mm_round_pd((X), _MM_FROUND_FLOOR)
#define	_mm_floor_ss(X, Y)   _mm_round_ss((X), (Y), _MM_FROUND_FLOOR)
#define	_mm_floor_sd(X, Y)   _mm_round_sd((X), (Y), _MM_FROUND_FLOOR)
#define	_mm_round_ps(X, M)
#define	_mm_round_ss(X, Y, M)
#define	_mm_round_pd(X, M)
#define	_mm_round_sd(X, Y, M)
#define	_mm_blend_pd(V1, V2, M)
#define	_mm_blend_ps(V1, V2, M)
#define	_mm_blend_epi16(V1, V2, M)
#define	_mm_dp_ps(X, Y, M)
#define	_mm_dp_pd(X, Y, M)
#define	_mm_insert_ps(X, Y, N)   __builtin_ia32_insertps128((X), (Y), (N))
#define	_mm_extract_ps(X, N)
#define	_MM_EXTRACT_FLOAT(D, X, N)
#define	_MM_MK_INSERTPS_NDX(X, Y, Z)   (((X) << 6) | ((Y) << 4) | (Z))
#define	_MM_PICK_OUT_PS(X, N)
#define	_mm_insert_epi8(X, I, N)
#define	_mm_insert_epi32(X, I, N)
#define	_mm_extract_epi8(X, N)
#define	_mm_extract_epi32(X, N)
#define	_mm_test_all_ones(V)   _mm_testc_si128((V), _mm_cmpeq_epi32((V), (V)))
#define	_mm_test_mix_ones_zeros(M, V)   _mm_testnzc_si128((M), (V))
#define	_mm_test_all_zeros(M, V)   _mm_testz_si128 ((M), (V))
#define	_mm_mpsadbw_epu8(X, Y, M)
#define	__DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("sse4.2")))
#define	_SIDD_UBYTE_OPS   0x00
#define	_SIDD_UWORD_OPS   0x01
#define	_SIDD_SBYTE_OPS   0x02
#define	_SIDD_SWORD_OPS   0x03
#define	_SIDD_CMP_EQUAL_ANY   0x00
#define	_SIDD_CMP_RANGES   0x04
#define	_SIDD_CMP_EQUAL_EACH   0x08
#define	_SIDD_CMP_EQUAL_ORDERED   0x0c
#define	_SIDD_POSITIVE_POLARITY   0x00
#define	_SIDD_NEGATIVE_POLARITY   0x10
#define	_SIDD_MASKED_POSITIVE_POLARITY   0x20
#define	_SIDD_MASKED_NEGATIVE_POLARITY   0x30
#define	_SIDD_LEAST_SIGNIFICANT   0x00
#define	_SIDD_MOST_SIGNIFICANT   0x40
#define	_SIDD_BIT_MASK   0x00
#define	_SIDD_UNIT_MASK   0x40
#define	_mm_cmpistrm(A, B, M)
#define	_mm_cmpistri(A, B, M)
#define	_mm_cmpestrm(A, LA, B, LB, M)
#define	_mm_cmpestri(A, LA, B, LB, M)
#define	_mm_cmpistra(A, B, M)
#define	_mm_cmpistrc(A, B, M)
#define	_mm_cmpistro(A, B, M)
#define	_mm_cmpistrs(A, B, M)
#define	_mm_cmpistrz(A, B, M)
#define	_mm_cmpestra(A, LA, B, LB, M)
#define	_mm_cmpestrc(A, LA, B, LB, M)
#define	_mm_cmpestro(A, LA, B, LB, M)
#define	_mm_cmpestrs(A, LA, B, LB, M)
#define	_mm_cmpestrz(A, LA, B, LB, M)

Macro Definition Documentation

#define __DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("sse4.1")))

#define __DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("sse4.2")))

#define _mm_blend_epi16	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m128i)__builtin_shufflevector((__v8hi)(__m128i)(V1), \
                                   (__v8hi)(__m128i)(V2), \
                                   (((M) & 0x01) ?  8 : 0), \
                                   (((M) & 0x02) ?  9 : 1), \
                                   (((M) & 0x04) ? 10 : 2), \
                                   (((M) & 0x08) ? 11 : 3), \
                                   (((M) & 0x10) ? 12 : 4), \
                                   (((M) & 0x20) ? 13 : 5), \
                                   (((M) & 0x40) ? 14 : 6), \
                                   (((M) & 0x80) ? 15 : 7)); })

#define _mm_blend_pd	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_shufflevector((__v2df)(__m128d)(V1), \
                                   (__v2df)(__m128d)(V2), \
                                   (((M) & 0x01) ? 2 : 0), \
                                   (((M) & 0x02) ? 3 : 1)); })

#define _mm_blend_ps	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m128)__builtin_shufflevector((__v4sf)(__m128)(V1), (__v4sf)(__m128)(V2), \
                                  (((M) & 0x01) ? 4 : 0), \
                                  (((M) & 0x02) ? 5 : 1), \
                                  (((M) & 0x04) ? 6 : 2), \
                                  (((M) & 0x08) ? 7 : 3)); })

#define _mm_ceil_pd

(

X

)

_mm_round_pd((X), _MM_FROUND_CEIL)

#define _mm_ceil_ps

(

X

)

_mm_round_ps((X), _MM_FROUND_CEIL)

#define _mm_ceil_sd	(		X,
			Y
	)		_mm_round_sd((X), (Y), _MM_FROUND_CEIL)

#define _mm_ceil_ss	(		X,
			Y
	)		_mm_round_ss((X), (Y), _MM_FROUND_CEIL)

#define _mm_cmpestra	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestria128((__v16qi)(__m128i)(A), (int)(LA), \
                                    (__v16qi)(__m128i)(B), (int)(LB), \
                                    (int)(M))

#define _mm_cmpestrc	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestric128((__v16qi)(__m128i)(A), (int)(LA), \
                                    (__v16qi)(__m128i)(B), (int)(LB), \
                                    (int)(M))

#define _mm_cmpestri	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestri128((__v16qi)(__m128i)(A), (int)(LA), \
                                   (__v16qi)(__m128i)(B), (int)(LB), \
                                   (int)(M))

#define _mm_cmpestrm	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(__m128i)__builtin_ia32_pcmpestrm128((__v16qi)(__m128i)(A), (int)(LA), \
                                       (__v16qi)(__m128i)(B), (int)(LB), \
                                       (int)(M))

#define _mm_cmpestro	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestrio128((__v16qi)(__m128i)(A), (int)(LA), \
                                    (__v16qi)(__m128i)(B), (int)(LB), \
                                    (int)(M))

#define _mm_cmpestrs	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestris128((__v16qi)(__m128i)(A), (int)(LA), \
                                    (__v16qi)(__m128i)(B), (int)(LB), \
                                    (int)(M))

#define _mm_cmpestrz	(		A,
			LA,
			B,
			LB,
			M
	)

Value:

(int)__builtin_ia32_pcmpestriz128((__v16qi)(__m128i)(A), (int)(LA), \
                                    (__v16qi)(__m128i)(B), (int)(LB), \
                                    (int)(M))

#define _mm_cmpistra	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistria128((__v16qi)(__m128i)(A), \
                                    (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistrc	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistric128((__v16qi)(__m128i)(A), \
                                    (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistri	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistri128((__v16qi)(__m128i)(A), \
                                   (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistrm	(		A,
			B,
			M
	)

Value:

(__m128i)__builtin_ia32_pcmpistrm128((__v16qi)(__m128i)(A), \
                                       (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistro	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistrio128((__v16qi)(__m128i)(A), \
                                    (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistrs	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistris128((__v16qi)(__m128i)(A), \
                                    (__v16qi)(__m128i)(B), (int)(M))

#define _mm_cmpistrz	(		A,
			B,
			M
	)

Value:

(int)__builtin_ia32_pcmpistriz128((__v16qi)(__m128i)(A), \
                                    (__v16qi)(__m128i)(B), (int)(M))

#define _mm_dp_pd	(		X,
			Y,
			M
	)

Value:

__extension__ ({\
  (__m128d) __builtin_ia32_dppd((__v2df)(__m128d)(X), \
                                (__v2df)(__m128d)(Y), (M)); })

#define _mm_dp_ps	(		X,
			Y,
			M
	)

Value:

__extension__ ({ \
  (__m128) __builtin_ia32_dpps((__v4sf)(__m128)(X), \
                               (__v4sf)(__m128)(Y), (M)); })

#define _mm_extract_epi32	(		X,
			N
	)

Value:

(__extension__                         \
                                 ({ __v4si __a = (__v4si)(__m128i)(X); \
                                    (int)__a[(N) & 3];}))

#define _mm_extract_epi8	(		X,
			N
	)

Value:

(__extension__                           \
                                ({ __v16qi __a = (__v16qi)(__m128i)(X); \
                                   (int)(unsigned char) __a[(N) & 15];}))

#define _MM_EXTRACT_FLOAT	(		D,
			X,
			N
	)

Value:

(__extension__ ({ __v4sf __a = (__v4sf)(X); \
                                                    (D) = __a[N]; }))

#define _mm_extract_ps	(		X,
			N
	)

Value:

(__extension__                      \
                              ({ union { int __i; float __f; } __t;  \
                                 __v4sf __a = (__v4sf)(__m128)(X);       \
                                 __t.__f = __a[(N) & 3];                 \
                                 __t.__i;}))

#define _mm_floor_pd

(

X

)

_mm_round_pd((X), _MM_FROUND_FLOOR)

#define _mm_floor_ps

(

X

)

_mm_round_ps((X), _MM_FROUND_FLOOR)

#define _mm_floor_sd	(		X,
			Y
	)		_mm_round_sd((X), (Y), _MM_FROUND_FLOOR)

#define _mm_floor_ss	(		X,
			Y
	)		_mm_round_ss((X), (Y), _MM_FROUND_FLOOR)

#define _MM_FROUND_CEIL   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_POS_INF)

#define _MM_FROUND_CUR_DIRECTION   0x04

#define _MM_FROUND_FLOOR   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEG_INF)

#define _MM_FROUND_NEARBYINT   (_MM_FROUND_NO_EXC | _MM_FROUND_CUR_DIRECTION)

#define _MM_FROUND_NINT   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_NEAREST_INT)

#define _MM_FROUND_NO_EXC   0x08

#define _MM_FROUND_RAISE_EXC   0x00

#define _MM_FROUND_RINT   (_MM_FROUND_RAISE_EXC | _MM_FROUND_CUR_DIRECTION)

#define _MM_FROUND_TO_NEAREST_INT   0x00

#define _MM_FROUND_TO_NEG_INF   0x01

#define _MM_FROUND_TO_POS_INF   0x02

#define _MM_FROUND_TO_ZERO   0x03

#define _MM_FROUND_TRUNC   (_MM_FROUND_RAISE_EXC | _MM_FROUND_TO_ZERO)

#define _mm_insert_epi32	(		X,
			I,
			N
	)

Value:

(__extension__                         \
                                   ({ __v4si __a = (__v4si)(__m128i)(X); \
                                      __a[(N) & 3] = (I);                \
                                      __a;}))

#define _mm_insert_epi8	(		X,
			I,
			N
	)

Value:

(__extension__                           \
                                  ({ __v16qi __a = (__v16qi)(__m128i)(X); \
                                     __a[(N) & 15] = (I);                 \
                                     __a;}))

#define _mm_insert_ps	(		X,
			Y,
			N
	)		__builtin_ia32_insertps128((X), (Y), (N))

#define _MM_MK_INSERTPS_NDX	(		X,
			Y,
			Z
	)		(((X) << 6) | ((Y) << 4) | (Z))

#define _mm_mpsadbw_epu8	(		X,
			Y,
			M
	)

Value:

__extension__ ({ \
  (__m128i) __builtin_ia32_mpsadbw128((__v16qi)(__m128i)(X), \
                                      (__v16qi)(__m128i)(Y), (M)); })

#define _MM_PICK_OUT_PS	(		X,
			N
	)

Value:

_mm_insert_ps (_mm_setzero_ps(), (X),   \
                                             _MM_MK_INSERTPS_NDX((N), 0, 0x0e))

#define _mm_round_pd	(		X,
			M
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_ia32_roundpd((__v2df)(__m128d)(X), (M)); })

#define _mm_round_ps	(		X,
			M
	)

Value:

__extension__ ({ \
  (__m128)__builtin_ia32_roundps((__v4sf)(__m128)(X), (M)); })

#define _mm_round_sd	(		X,
			Y,
			M
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_ia32_roundsd((__v2df)(__m128d)(X), \
                                  (__v2df)(__m128d)(Y), (M)); })

#define _mm_round_ss	(		X,
			Y,
			M
	)

Value:

__extension__ ({ \
  (__m128)__builtin_ia32_roundss((__v4sf)(__m128)(X), \
                                 (__v4sf)(__m128)(Y), (M)); })

#define _mm_test_all_ones

(

V

)

_mm_testc_si128((V), _mm_cmpeq_epi32((V), (V)))

#define _mm_test_all_zeros	(		M,
			V
	)		_mm_testz_si128 ((M), (V))

#define _mm_test_mix_ones_zeros	(		M,
			V
	)		_mm_testnzc_si128((M), (V))

#define _SIDD_BIT_MASK   0x00

#define _SIDD_CMP_EQUAL_ANY   0x00

#define _SIDD_CMP_EQUAL_EACH   0x08

#define _SIDD_CMP_EQUAL_ORDERED   0x0c

#define _SIDD_CMP_RANGES   0x04

#define _SIDD_LEAST_SIGNIFICANT   0x00

#define _SIDD_MASKED_NEGATIVE_POLARITY   0x30

#define _SIDD_MASKED_POSITIVE_POLARITY   0x20

#define _SIDD_MOST_SIGNIFICANT   0x40

#define _SIDD_NEGATIVE_POLARITY   0x10

#define _SIDD_POSITIVE_POLARITY   0x00

#define _SIDD_SBYTE_OPS   0x02

#define _SIDD_SWORD_OPS   0x03

#define _SIDD_UBYTE_OPS   0x00

#define _SIDD_UNIT_MASK   0x40

#define _SIDD_UWORD_OPS   0x01