avxintrin.h File Reference

Go to the source code of this file.

Macros
#define	__DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("avx")))
#define	_mm256_round_pd(V, M)
#define	_mm256_round_ps(V, M)
#define	_mm256_ceil_pd(V)   _mm256_round_pd((V), _MM_FROUND_CEIL)
#define	_mm256_floor_pd(V)   _mm256_round_pd((V), _MM_FROUND_FLOOR)
#define	_mm256_ceil_ps(V)   _mm256_round_ps((V), _MM_FROUND_CEIL)
#define	_mm256_floor_ps(V)   _mm256_round_ps((V), _MM_FROUND_FLOOR)
#define	_mm_permute_pd(A, C)
#define	_mm256_permute_pd(A, C)
#define	_mm_permute_ps(A, C)
#define	_mm256_permute_ps(A, C)
#define	_mm256_permute2f128_pd(V1, V2, M)
#define	_mm256_permute2f128_ps(V1, V2, M)
#define	_mm256_permute2f128_si256(V1, V2, M)
#define	_mm256_blend_pd(V1, V2, M)
#define	_mm256_blend_ps(V1, V2, M)
#define	_mm256_dp_ps(V1, V2, M)
#define	_mm256_shuffle_ps(a, b, mask)
#define	_mm256_shuffle_pd(a, b, mask)
#define	_CMP_EQ_OQ   0x00 /* Equal (ordered, non-signaling) */
#define	_CMP_LT_OS   0x01 /* Less-than (ordered, signaling) */
#define	_CMP_LE_OS   0x02 /* Less-than-or-equal (ordered, signaling) */
#define	_CMP_UNORD_Q   0x03 /* Unordered (non-signaling) */
#define	_CMP_NEQ_UQ   0x04 /* Not-equal (unordered, non-signaling) */
#define	_CMP_NLT_US   0x05 /* Not-less-than (unordered, signaling) */
#define	_CMP_NLE_US   0x06 /* Not-less-than-or-equal (unordered, signaling) */
#define	_CMP_ORD_Q   0x07 /* Ordered (nonsignaling) */
#define	_CMP_EQ_UQ   0x08 /* Equal (unordered, non-signaling) */
#define	_CMP_NGE_US   0x09 /* Not-greater-than-or-equal (unord, signaling) */
#define	_CMP_NGT_US   0x0a /* Not-greater-than (unordered, signaling) */
#define	_CMP_FALSE_OQ   0x0b /* False (ordered, non-signaling) */
#define	_CMP_NEQ_OQ   0x0c /* Not-equal (ordered, non-signaling) */
#define	_CMP_GE_OS   0x0d /* Greater-than-or-equal (ordered, signaling) */
#define	_CMP_GT_OS   0x0e /* Greater-than (ordered, signaling) */
#define	_CMP_TRUE_UQ   0x0f /* True (unordered, non-signaling) */
#define	_CMP_EQ_OS   0x10 /* Equal (ordered, signaling) */
#define	_CMP_LT_OQ   0x11 /* Less-than (ordered, non-signaling) */
#define	_CMP_LE_OQ   0x12 /* Less-than-or-equal (ordered, non-signaling) */
#define	_CMP_UNORD_S   0x13 /* Unordered (signaling) */
#define	_CMP_NEQ_US   0x14 /* Not-equal (unordered, signaling) */
#define	_CMP_NLT_UQ   0x15 /* Not-less-than (unordered, non-signaling) */
#define	_CMP_NLE_UQ   0x16 /* Not-less-than-or-equal (unord, non-signaling) */
#define	_CMP_ORD_S   0x17 /* Ordered (signaling) */
#define	_CMP_EQ_US   0x18 /* Equal (unordered, signaling) */
#define	_CMP_NGE_UQ   0x19 /* Not-greater-than-or-equal (unord, non-sign) */
#define	_CMP_NGT_UQ   0x1a /* Not-greater-than (unordered, non-signaling) */
#define	_CMP_FALSE_OS   0x1b /* False (ordered, signaling) */
#define	_CMP_NEQ_OS   0x1c /* Not-equal (ordered, signaling) */
#define	_CMP_GE_OQ   0x1d /* Greater-than-or-equal (ordered, non-signaling) */
#define	_CMP_GT_OQ   0x1e /* Greater-than (ordered, non-signaling) */
#define	_CMP_TRUE_US   0x1f /* True (unordered, signaling) */
#define	_mm_cmp_pd(a, b, c)
#define	_mm_cmp_ps(a, b, c)
#define	_mm256_cmp_pd(a, b, c)
#define	_mm256_cmp_ps(a, b, c)
#define	_mm_cmp_sd(a, b, c)
#define	_mm_cmp_ss(a, b, c)
#define	_mm256_insertf128_ps(V1, V2, M)
#define	_mm256_insertf128_pd(V1, V2, M)
#define	_mm256_insertf128_si256(V1, V2, M)
#define	_mm256_extractf128_ps(V, M)
#define	_mm256_extractf128_pd(V, M)
#define	_mm256_extractf128_si256(V, M)

Typedefs
typedef double __v4df	__attribute__ ((__vector_size__(32)))

Macro Definition Documentation

#define __DEFAULT_FN_ATTRS   __attribute__((__always_inline__, __nodebug__, __target__("avx")))

#define _CMP_EQ_OQ   0x00 /* Equal (ordered, non-signaling) */

#define _CMP_EQ_OS   0x10 /* Equal (ordered, signaling) */

#define _CMP_EQ_UQ   0x08 /* Equal (unordered, non-signaling) */

#define _CMP_EQ_US   0x18 /* Equal (unordered, signaling) */

#define _CMP_FALSE_OQ   0x0b /* False (ordered, non-signaling) */

#define _CMP_FALSE_OS   0x1b /* False (ordered, signaling) */

#define _CMP_GE_OQ   0x1d /* Greater-than-or-equal (ordered, non-signaling) */

#define _CMP_GE_OS   0x0d /* Greater-than-or-equal (ordered, signaling) */

#define _CMP_GT_OQ   0x1e /* Greater-than (ordered, non-signaling) */

#define _CMP_GT_OS   0x0e /* Greater-than (ordered, signaling) */

#define _CMP_LE_OQ   0x12 /* Less-than-or-equal (ordered, non-signaling) */

#define _CMP_LE_OS   0x02 /* Less-than-or-equal (ordered, signaling) */

#define _CMP_LT_OQ   0x11 /* Less-than (ordered, non-signaling) */

#define _CMP_LT_OS   0x01 /* Less-than (ordered, signaling) */

#define _CMP_NEQ_OQ   0x0c /* Not-equal (ordered, non-signaling) */

#define _CMP_NEQ_OS   0x1c /* Not-equal (ordered, signaling) */

#define _CMP_NEQ_UQ   0x04 /* Not-equal (unordered, non-signaling) */

#define _CMP_NEQ_US   0x14 /* Not-equal (unordered, signaling) */

#define _CMP_NGE_UQ   0x19 /* Not-greater-than-or-equal (unord, non-sign) */

#define _CMP_NGE_US   0x09 /* Not-greater-than-or-equal (unord, signaling) */

#define _CMP_NGT_UQ   0x1a /* Not-greater-than (unordered, non-signaling) */

#define _CMP_NGT_US   0x0a /* Not-greater-than (unordered, signaling) */

#define _CMP_NLE_UQ   0x16 /* Not-less-than-or-equal (unord, non-signaling) */

#define _CMP_NLE_US   0x06 /* Not-less-than-or-equal (unordered, signaling) */

#define _CMP_NLT_UQ   0x15 /* Not-less-than (unordered, non-signaling) */

#define _CMP_NLT_US   0x05 /* Not-less-than (unordered, signaling) */

#define _CMP_ORD_Q   0x07 /* Ordered (nonsignaling) */

#define _CMP_ORD_S   0x17 /* Ordered (signaling) */

#define _CMP_TRUE_UQ   0x0f /* True (unordered, non-signaling) */

#define _CMP_TRUE_US   0x1f /* True (unordered, signaling) */

#define _CMP_UNORD_Q   0x03 /* Unordered (non-signaling) */

#define _CMP_UNORD_S   0x13 /* Unordered (signaling) */

#define _mm256_blend_pd	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256d)__builtin_shufflevector((__v4df)(__m256d)(V1), \
                                   (__v4df)(__m256d)(V2), \
                                   (((M) & 0x01) ? 4 : 0), \
                                   (((M) & 0x02) ? 5 : 1), \
                                   (((M) & 0x04) ? 6 : 2), \
                                   (((M) & 0x08) ? 7 : 3)); })

#define _mm256_blend_ps	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256)__builtin_shufflevector((__v8sf)(__m256)(V1), \
                                  (__v8sf)(__m256)(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 _mm256_ceil_pd

(

V

)

_mm256_round_pd((V), _MM_FROUND_CEIL)

#define _mm256_ceil_ps

(

V

)

_mm256_round_ps((V), _MM_FROUND_CEIL)

#define _mm256_cmp_pd	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m256d)__builtin_ia32_cmppd256((__v4df)(__m256d)(a), \
                                   (__v4df)(__m256d)(b), (c)); })

#define _mm256_cmp_ps	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m256)__builtin_ia32_cmpps256((__v8sf)(__m256)(a), \
                                  (__v8sf)(__m256)(b), (c)); })

#define _mm256_dp_ps	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256)__builtin_ia32_dpps256((__v8sf)(__m256)(V1), \
                                 (__v8sf)(__m256)(V2), (M)); })

#define _mm256_extractf128_pd	(		V,
			M
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_shufflevector( \
    (__v4df)(__m256d)(V), \
    (__v4df)(_mm256_setzero_pd()), \
    (((M) & 1) ? 2 : 0), \
    (((M) & 1) ? 3 : 1) );})

#define _mm256_extractf128_ps	(		V,
			M
	)

Value:

__extension__ ({ \
  (__m128)__builtin_shufflevector( \
    (__v8sf)(__m256)(V), \
    (__v8sf)(_mm256_setzero_ps()), \
    (((M) & 1) ? 4 : 0), \
    (((M) & 1) ? 5 : 1), \
    (((M) & 1) ? 6 : 2), \
    (((M) & 1) ? 7 : 3) );})

#define _mm256_extractf128_si256	(		V,
			M
	)

Value:

__extension__ ({ \
  (__m128i)__builtin_shufflevector( \
    (__v4di)(__m256i)(V), \
    (__v4di)(_mm256_setzero_si256()), \
    (((M) & 1) ? 2 : 0), \
    (((M) & 1) ? 3 : 1) );})

#define _mm256_floor_pd

(

V

)

_mm256_round_pd((V), _MM_FROUND_FLOOR)

#define _mm256_floor_ps

(

V

)

_mm256_round_ps((V), _MM_FROUND_FLOOR)

#define _mm256_insertf128_pd	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256d)__builtin_shufflevector( \
    (__v4df)(__m256d)(V1), \
    (__v4df)_mm256_castpd128_pd256((__m128d)(V2)), \
    (((M) & 1) ? 0 : 4), \
    (((M) & 1) ? 1 : 5), \
    (((M) & 1) ? 4 : 2), \
    (((M) & 1) ? 5 : 3) );})

#define _mm256_insertf128_ps	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256)__builtin_shufflevector( \
    (__v8sf)(__m256)(V1), \
    (__v8sf)_mm256_castps128_ps256((__m128)(V2)), \
    (((M) & 1) ?  0 :  8), \
    (((M) & 1) ?  1 :  9), \
    (((M) & 1) ?  2 : 10), \
    (((M) & 1) ?  3 : 11), \
    (((M) & 1) ?  8 :  4), \
    (((M) & 1) ?  9 :  5), \
    (((M) & 1) ? 10 :  6), \
    (((M) & 1) ? 11 :  7) );})

#define _mm256_insertf128_si256	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256i)__builtin_shufflevector( \
    (__v4di)(__m256i)(V1), \
    (__v4di)_mm256_castsi128_si256((__m128i)(V2)), \
    (((M) & 1) ? 0 : 4), \
    (((M) & 1) ? 1 : 5), \
    (((M) & 1) ? 4 : 2), \
    (((M) & 1) ? 5 : 3) );})

#define _mm256_permute2f128_pd	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256d)__builtin_ia32_vperm2f128_pd256((__v4df)(__m256d)(V1), \
                                           (__v4df)(__m256d)(V2), (M)); })

#define _mm256_permute2f128_ps	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256)__builtin_ia32_vperm2f128_ps256((__v8sf)(__m256)(V1), \
                                          (__v8sf)(__m256)(V2), (M)); })

#define _mm256_permute2f128_si256	(		V1,
			V2,
			M
	)

Value:

__extension__ ({ \
  (__m256i)__builtin_ia32_vperm2f128_si256((__v8si)(__m256i)(V1), \
                                           (__v8si)(__m256i)(V2), (M)); })

#define _mm256_permute_pd	(		A,
			C
	)

Value:

__extension__ ({ \
  (__m256d)__builtin_shufflevector((__v4df)(__m256d)(A), \
                                   (__v4df)_mm256_setzero_pd(), \
                                   (C) & 0x1, ((C) & 0x2) >> 1, \
                                   2 + (((C) & 0x4) >> 2), \
                                   2 + (((C) & 0x8) >> 3)); })

#define _mm256_permute_ps	(		A,
			C
	)

Value:

__extension__ ({ \
  (__m256)__builtin_shufflevector((__v8sf)(__m256)(A), \
                                  (__v8sf)_mm256_setzero_ps(), \
                                  (C) & 0x3, ((C) & 0xc) >> 2, \
                                  ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6, \
                                  4 + (((C) & 0x03) >> 0), \
                                  4 + (((C) & 0x0c) >> 2), \
                                  4 + (((C) & 0x30) >> 4), \
                                  4 + (((C) & 0xc0) >> 6)); })

#define _mm256_round_pd	(		V,
			M
	)

Value:

__extension__ ({ \
    (__m256d)__builtin_ia32_roundpd256((__v4df)(__m256d)(V), (M)); })

#define _mm256_round_ps	(		V,
			M
	)

Value:

__extension__ ({ \
  (__m256)__builtin_ia32_roundps256((__v8sf)(__m256)(V), (M)); })

#define _mm256_shuffle_pd	(		a,
			b,
			mask
	)

Value:

__extension__ ({ \
        (__m256d)__builtin_shufflevector((__v4df)(__m256d)(a), \
                                         (__v4df)(__m256d)(b), \
                                         (mask) & 0x1, \
                                         (((mask) & 0x2) >> 1) + 4, \
                                         (((mask) & 0x4) >> 2) + 2, \
                                         (((mask) & 0x8) >> 3) + 6); })

#define _mm256_shuffle_ps	(		a,
			b,
			mask
	)

Value:

__extension__ ({ \
        (__m256)__builtin_shufflevector((__v8sf)(__m256)(a), \
                                        (__v8sf)(__m256)(b), \
                                        (mask) & 0x3, \
                                        ((mask) & 0xc) >> 2, \
                                        (((mask) & 0x30) >> 4) + 8, \
                                        (((mask) & 0xc0) >> 6) + 8, \
                                        ((mask) & 0x3) + 4, \
                                        (((mask) & 0xc) >> 2) + 4, \
                                        (((mask) & 0x30) >> 4) + 12, \
                                        (((mask) & 0xc0) >> 6) + 12); })

#define _mm_cmp_pd	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_ia32_cmppd((__v2df)(__m128d)(a), \
                                (__v2df)(__m128d)(b), (c)); })

#define _mm_cmp_ps	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m128)__builtin_ia32_cmpps((__v4sf)(__m128)(a), \
                               (__v4sf)(__m128)(b), (c)); })

#define _mm_cmp_sd	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_ia32_cmpsd((__v2df)(__m128d)(a), \
                                (__v2df)(__m128d)(b), (c)); })

#define _mm_cmp_ss	(		a,
			b,
			c
	)

Value:

__extension__ ({ \
  (__m128)__builtin_ia32_cmpss((__v4sf)(__m128)(a), \
                               (__v4sf)(__m128)(b), (c)); })

#define _mm_permute_pd	(		A,
			C
	)

Value:

__extension__ ({ \
  (__m128d)__builtin_shufflevector((__v2df)(__m128d)(A), \
                                   (__v2df)_mm_setzero_pd(), \
                                   (C) & 0x1, ((C) & 0x2) >> 1); })

#define _mm_permute_ps	(		A,
			C
	)

Value:

__extension__ ({ \
  (__m128)__builtin_shufflevector((__v4sf)(__m128)(A), \
                                  (__v4sf)_mm_setzero_ps(), \
                                   (C) & 0x3, ((C) & 0xc) >> 2, \
                                   ((C) & 0x30) >> 4, ((C) & 0xc0) >> 6); })

Typedef Documentation

typedef long long __m256i __attribute__((__vector_size__(32)))