402 /// __m256d _mm256_round_pd(__m256d V, const int M);
407 /// \param V
424 #define _mm256_round_pd(V, M) \  argument
425   ((__m256d)__builtin_ia32_roundpd256((__v4df)(__m256d)(V), (M)))
434 /// __m256 _mm256_round_ps(__m256 V, const int M);
439 /// \param V
456 #define _mm256_round_ps(V, M) \  argument
457   ((__m256)__builtin_ia32_roundps256((__v8sf)(__m256)(V), (M)))
466 /// __m256d _mm256_ceil_pd(__m256d V);
471 /// \param V
474 #define _mm256_ceil_pd(V)  _mm256_round_pd((V), _MM_FROUND_CEIL)  argument
483 /// __m256d _mm256_floor_pd(__m256d V);
488 /// \param V
492 #define _mm256_floor_pd(V) _mm256_round_pd((V), _MM_FROUND_FLOOR)  argument
501 /// __m256 _mm256_ceil_ps(__m256 V);
506 /// \param V
509 #define _mm256_ceil_ps(V)  _mm256_round_ps((V), _MM_FROUND_CEIL)  argument
518 /// __m256 _mm256_floor_ps(__m256 V);
523 /// \param V
526 #define _mm256_floor_ps(V) _mm256_round_ps((V), _MM_FROUND_FLOOR)  argument
858 ///      10: Bits [95:64] of the source are copied to bits [31:0] of the
867 ///      10: Bits [95:64] of the source are copied to bits [63:32] of the
872 ///      00: Bits [31:0] of the source are copied to bits [95:64] of the
874 ///      01: Bits [63:32] of the source are copied to bits [95:64] of the
876 ///      10: Bits [95:64] of the source are copied to bits [95:64] of the
878 ///      11: Bits [127:96] of the source are copied to bits [95:64] of the
885 ///      10: Bits [95:64] of the source are copied to bits [127:96] of the
913 ///      10: Bits [95:64] of the source are copied to bits [31:0] of the
922 ///      10: Bits [95:64] of the source are copied to bits [63:32] of the
927 ///      00: Bits [31:0] of the source are copied to bits [95:64] of the
929 ///      01: Bits [63:32] of the source are copied to bits [95:64] of the
931 ///      10: Bits [95:64] of the source are copied to bits [95:64] of the
933 ///      11: Bits [127:96] of the source are copied to bits [95:64] of the
940 ///      10: Bits [95:64] of the source are copied to bits [127:96] of the
1078 ///      10: Bits [95:64] of the source are copied to bits [31:0] of the
1087 ///      10: Bits [95:64] of the source are copied to bits [63:32] of the
1092 ///      00: Bits [31:0] of the source are copied to bits [95:64] of the
1094 ///      01: Bits [63:32] of the source are copied to bits [95:64] of the
1096 ///      10: Bits [95:64] of the source are copied to bits [95:64] of the
1098 ///      11: Bits [127:96] of the source are copied to bits [95:64] of the
1105 ///      10: Bits [95:64] of the source are copied to bits [127:96] of the
1134 ///      10: Bits [95:64] of the source are copied to bits [31:0] of the
1143 ///      10: Bits [95:64] of the source are copied to bits [63:32] of the
1148 ///      00: Bits [31:0] of the source are copied to bits [95:64] of the
1150 ///      01: Bits [63:32] of the source are copied to bits [95:64] of the
1152 ///      10: Bits [95:64] of the source are copied to bits [95:64] of the
1154 ///      11: Bits [127:96] of the source are copied to bits [95:64] of the
1161 ///      10: Bits [95:64] of the source are copied to bits [127:96] of the
1425 ///    A 256-bit vector operand, with mask bits 255, 223, 191, 159, 127, 95, 63,
1522 ///    Bits [5:4] are used to assign values to bits [95:64] and [223:192] in the
1529 ///    10: Bits [95:64] and [223:192] are copied from the selected operand. \n
2372 ///    Bits [127:96] of \a __a are written to bits [127:96] and [95:64] of the
2397 ///    Bits [95:64] of \a __a are written to bits [127:96] and [95:64] of the
2486 ///    Bits [95:64] are written to bits [31:0] of the return value. \n
2487 ///    Bits [127:96] are written to bits [95:64] of the return value. \n
2492 ///    Bits [95:64] are written to bits [63:32] of the return value. \n
2514 ///    Bits [63:32] are written to bits [95:64] of the return value. \n
3722 ///    A single-precision floating-point value used to initialize bits [95:64]
3757 ///    A 32-bit integral value used to initialize bits [95:64] of the result.
3799 ///    A 16-bit integral value used to initialize bits [95:80] of the result.
3870 ///    An 8-bit integral value used to initialize bits [95:88] of the result.
3980 ///    A single-precision floating-point value used to initialize bits [95:64]
4018 ///    A 32-bit integral value used to initialize bits [95:64] of the result.
4056 ///    A 16-bit integral value used to initialize bits [95:80] of the result.
4121 ///    An 8-bit integral value used to initialize bits [95:88] of the result.
4762 /// __m128 _mm256_extractf128_ps(__m256 V, const int M);
4767 /// \param V
4772 ///    If bit [0] of \a M is 0, bits [127:0] of \a V are copied to the
4774 ///    If bit [0] of \a M is 1, bits [255:128] of \a V are copied to the result.
4776 #define _mm256_extractf128_ps(V, M) \  argument
4777   ((__m128)__builtin_ia32_vextractf128_ps256((__v8sf)(__m256)(V), (int)(M)))
4786 /// __m128d _mm256_extractf128_pd(__m256d V, const int M);
4791 /// \param V
4796 ///    If bit [0] of \a M is 0, bits [127:0] of \a V are copied to the
4798 ///    If bit [0] of \a M is 1, bits [255:128] of \a V are copied to the result.
4800 #define _mm256_extractf128_pd(V, M) \  argument
4801   ((__m128d)__builtin_ia32_vextractf128_pd256((__v4df)(__m256d)(V), (int)(M)))
4810 /// __m128i _mm256_extractf128_si256(__m256i V, const int M);
4815 /// \param V
4820 ///    If bit [0] of \a M is 0, bits [127:0] of \a V are copied to the
4822 ///    If bit [0] of \a M is 1, bits [255:128] of \a V are copied to the result.
4824 #define _mm256_extractf128_si256(V, M) \  argument
4825   ((__m128i)__builtin_ia32_vextractf128_si256((__v8si)(__m256i)(V), (int)(M)))