404 Not Found

PATH: lib64 / llvm17 / lib / clang / 17 / include

/*===------------- avx512vlvnniintrin.h - VNNI intrinsics ------------------===
 *
 *
 * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 * See https://llvm.org/LICENSE.txt for license information.
 * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 *
 *===-----------------------------------------------------------------------===
 */
#ifndef __IMMINTRIN_H
#error "Never use <avx512vlvnniintrin.h> directly; include <immintrin.h> instead."
#endif

#ifndef __AVX512VLVNNIINTRIN_H
#define __AVX512VLVNNIINTRIN_H

/* Define the default attributes for the functions in this file. */
#define __DEFAULT_FN_ATTRS128 __attribute__((__always_inline__, __nodebug__, __target__("avx512vl,avx512vnni"), __min_vector_width__(128)))
#define __DEFAULT_FN_ATTRS256 __attribute__((__always_inline__, __nodebug__, __target__("avx512vl,avx512vnni"), __min_vector_width__(256)))

/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with
/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed
/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
/// in \a S, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 7
///      tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j]))
///      tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1]))
///      tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2]))
///      tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3]))
///      DST.dword[j] := S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4
///    ENDFOR
///    DST[MAX:256] := 0
/// \endcode
#define _mm256_dpbusd_epi32(S, A, B) \
  ((__m256i)__builtin_ia32_vpdpbusd256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with
/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed
/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
/// in \a S using signed saturation, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 7
///      tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j]))
///      tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1]))
///      tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2]))
///      tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3]))
///      DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4)
///    ENDFOR
///    DST[MAX:256] := 0
/// \endcode
#define _mm256_dpbusds_epi32(S, A, B) \
  ((__m256i)__builtin_ia32_vpdpbusds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with
/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit
/// results. Sum these 2 results with the corresponding 32-bit integer in \a S,
///  and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 7
///      tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j])
///      tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1])
///      DST.dword[j] := S.dword[j] + tmp1 + tmp2
///    ENDFOR
///    DST[MAX:256] := 0
/// \endcode
#define _mm256_dpwssd_epi32(S, A, B) \
  ((__m256i)__builtin_ia32_vpdpwssd256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with
/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit
/// results. Sum these 2 results with the corresponding 32-bit integer in \a S
/// using signed saturation, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 7
///      tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j])
///      tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1])
///      DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2)
///    ENDFOR
///    DST[MAX:256] := 0
/// \endcode
#define _mm256_dpwssds_epi32(S, A, B) \
  ((__m256i)__builtin_ia32_vpdpwssds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with
/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed
/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
/// in \a S, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPBUSD </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 3
///      tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j]))
///      tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1]))
///      tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2]))
///      tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3]))
///      DST.dword[j] := S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4
///    ENDFOR
///    DST[MAX:128] := 0
/// \endcode
#define _mm_dpbusd_epi32(S, A, B) \
  ((__m128i)__builtin_ia32_vpdpbusd128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

/// Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in \a A with
/// corresponding signed 8-bit integers in \a B, producing 4 intermediate signed
/// 16-bit results. Sum these 4 results with the corresponding 32-bit integer
/// in \a S using signed saturation, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPBUSDS </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 3
///      tmp1.word := Signed(ZeroExtend16(A.byte[4*j]) * SignExtend16(B.byte[4*j]))
///      tmp2.word := Signed(ZeroExtend16(A.byte[4*j+1]) * SignExtend16(B.byte[4*j+1]))
///      tmp3.word := Signed(ZeroExtend16(A.byte[4*j+2]) * SignExtend16(B.byte[4*j+2]))
///      tmp4.word := Signed(ZeroExtend16(A.byte[4*j+3]) * SignExtend16(B.byte[4*j+3]))
///      DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2 + tmp3 + tmp4)
///    ENDFOR
///    DST[MAX:128] := 0
/// \endcode
#define _mm_dpbusds_epi32(S, A, B) \
  ((__m128i)__builtin_ia32_vpdpbusds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with
/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit
/// results. Sum these 2 results with the corresponding 32-bit integer in \a S,
/// and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPWSSD </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 3
///      tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j])
///      tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1])
///      DST.dword[j] := S.dword[j] + tmp1 + tmp2
///    ENDFOR
///    DST[MAX:128] := 0
/// \endcode
#define _mm_dpwssd_epi32(S, A, B) \
  ((__m128i)__builtin_ia32_vpdpwssd128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

/// Multiply groups of 2 adjacent pairs of signed 16-bit integers in \a A with
/// corresponding 16-bit integers in \a B, producing 2 intermediate signed 32-bit
/// results. Sum these 2 results with the corresponding 32-bit integer in \a S
/// using signed saturation, and store the packed 32-bit results in DST.
///
/// This intrinsic corresponds to the <c> VPDPWSSDS </c> instructions.
///
/// \code{.operation}
///    FOR j := 0 to 3
///      tmp1.dword := SignExtend32(A.word[2*j]) * SignExtend32(B.word[2*j])
///      tmp2.dword := SignExtend32(A.word[2*j+1]) * SignExtend32(B.word[2*j+1])
///      DST.dword[j] := Saturate32(S.dword[j] + tmp1 + tmp2)
///    ENDFOR
///    DST[MAX:128] := 0
/// \endcode
#define _mm_dpwssds_epi32(S, A, B) \
  ((__m128i)__builtin_ia32_vpdpwssds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_mask_dpbusd_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                     (__v8si)_mm256_dpbusd_epi32(__S, __A, __B),
                                     (__v8si)__S);
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_maskz_dpbusd_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                     (__v8si)_mm256_dpbusd_epi32(__S, __A, __B),
                                     (__v8si)_mm256_setzero_si256());
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_mask_dpbusds_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                    (__v8si)_mm256_dpbusds_epi32(__S, __A, __B),
                                    (__v8si)__S);
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_maskz_dpbusds_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                     (__v8si)_mm256_dpbusds_epi32(__S, __A, __B),
                                     (__v8si)_mm256_setzero_si256());
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_mask_dpwssd_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                     (__v8si)_mm256_dpwssd_epi32(__S, __A, __B),
                                     (__v8si)__S);
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_maskz_dpwssd_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                     (__v8si)_mm256_dpwssd_epi32(__S, __A, __B),
                                     (__v8si)_mm256_setzero_si256());
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_mask_dpwssds_epi32(__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                    (__v8si)_mm256_dpwssds_epi32(__S, __A, __B),
                                    (__v8si)__S);
}

static __inline__ __m256i __DEFAULT_FN_ATTRS256
_mm256_maskz_dpwssds_epi32(__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
{
  return (__m256i)__builtin_ia32_selectd_256(__U,
                                    (__v8si)_mm256_dpwssds_epi32(__S, __A, __B),
                                    (__v8si)_mm256_setzero_si256());
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_mask_dpbusd_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                        (__v4si)_mm_dpbusd_epi32(__S, __A, __B),
                                        (__v4si)__S);
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_maskz_dpbusd_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                        (__v4si)_mm_dpbusd_epi32(__S, __A, __B),
                                        (__v4si)_mm_setzero_si128());
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_mask_dpbusds_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                       (__v4si)_mm_dpbusds_epi32(__S, __A, __B),
                                       (__v4si)__S);
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_maskz_dpbusds_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                       (__v4si)_mm_dpbusds_epi32(__S, __A, __B),
                                       (__v4si)_mm_setzero_si128());
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_mask_dpwssd_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                        (__v4si)_mm_dpwssd_epi32(__S, __A, __B),
                                        (__v4si)__S);
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_maskz_dpwssd_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                        (__v4si)_mm_dpwssd_epi32(__S, __A, __B),
                                        (__v4si)_mm_setzero_si128());
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_mask_dpwssds_epi32(__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                       (__v4si)_mm_dpwssds_epi32(__S, __A, __B),
                                       (__v4si)__S);
}

static __inline__ __m128i __DEFAULT_FN_ATTRS128
_mm_maskz_dpwssds_epi32(__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
{
  return (__m128i)__builtin_ia32_selectd_128(__U,
                                       (__v4si)_mm_dpwssds_epi32(__S, __A, __B),
                                       (__v4si)_mm_setzero_si128());
}

#undef __DEFAULT_FN_ATTRS128
#undef __DEFAULT_FN_ATTRS256

#endif

[+] ..
[-] riscv_ntlh.h [edit]
[-] avx512vlbitalgintrin.h [edit]
[-] avxifmaintrin.h [edit]
[-] mm_malloc.h [edit]
[-] avx512vlbf16intrin.h [edit]
[-] mwaitxintrin.h [edit]
[-] adxintrin.h [edit]
[-] arm_cmse.h [edit]
[-] __clang_cuda_intrinsics.h [edit]
[-] module.modulemap [edit]
[-] crc32intrin.h [edit]
[+] cuda_wrappers
[-] msa.h [edit]
[-] __clang_hip_stdlib.h [edit]
[-] arm_acle.h [edit]
[-] arm_neon_sve_bridge.h [edit]
[-] xsaveintrin.h [edit]
[-] velintrin_gen.h [edit]
[-] stdatomic.h [edit]
[-] hvx_hexagon_protos.h [edit]
[-] uintrintrin.h [edit]
[-] fxsrintrin.h [edit]
[-] sha512intrin.h [edit]
[-] __clang_cuda_texture_intrinsics.h [edit]
[-] iso646.h [edit]
[-] unwind.h [edit]
[-] avx512vlbwintrin.h [edit]
[-] avxvnniint8intrin.h [edit]
[-] avx512erintrin.h [edit]
[-] avxvnniintrin.h [edit]
[-] ia32intrin.h [edit]
[-] rdseedintrin.h [edit]
[-] prfchiintrin.h [edit]
[-] amxcomplexintrin.h [edit]
[-] clflushoptintrin.h [edit]
[-] htmxlintrin.h [edit]
[-] avx512fintrin.h [edit]
[-] gfniintrin.h [edit]
[-] arm_cde.h [edit]
[+] ppc_wrappers
[-] amxfp16intrin.h [edit]
[-] velintrin_approx.h [edit]
[-] avx512pfintrin.h [edit]
[-] stdarg.h [edit]
[-] cmpccxaddintrin.h [edit]
[-] avx512vlintrin.h [edit]
[-] __clang_hip_math.h [edit]
[-] vecintrin.h [edit]
[-] xtestintrin.h [edit]
[-] __wmmintrin_aes.h [edit]
[-] arm_neon.h [edit]
[-] immintrin.h [edit]
[-] ammintrin.h [edit]
[-] waitpkgintrin.h [edit]
[-] vpclmulqdqintrin.h [edit]
[-] fmaintrin.h [edit]
[-] tsxldtrkintrin.h [edit]
[-] prfchwintrin.h [edit]
[-] avx512bitalgintrin.h [edit]
[-] bmiintrin.h [edit]
[-] __wmmintrin_pclmul.h [edit]
[-] htmintrin.h [edit]
[-] mm3dnow.h [edit]
[-] __clang_cuda_builtin_vars.h [edit]
[-] __clang_hip_runtime_wrapper.h [edit]
[-] stdbool.h [edit]
[-] altivec.h [edit]
[-] wbnoinvdintrin.h [edit]
[-] keylockerintrin.h [edit]
[-] tgmath.h [edit]
[-] hexagon_circ_brev_intrinsics.h [edit]
[-] x86intrin.h [edit]
[-] pkuintrin.h [edit]
[-] avx512vbmivlintrin.h [edit]
[-] avxneconvertintrin.h [edit]
[-] __clang_hip_cmath.h [edit]
[-] sgxintrin.h [edit]
[-] f16cintrin.h [edit]
[-] opencl-c-base.h [edit]
[-] cpuid.h [edit]
[-] raointintrin.h [edit]
[-] builtins.h [edit]
[-] emmintrin.h [edit]
[-] smmintrin.h [edit]
[-] vaesintrin.h [edit]
[-] larchintrin.h [edit]
[-] avx512ifmaintrin.h [edit]
[-] intrin.h [edit]
[-] avx512vlvp2intersectintrin.h [edit]
[-] fma4intrin.h [edit]
[-] pmmintrin.h [edit]
[-] __clang_hip_libdevice_declares.h [edit]
[-] limits.h [edit]
[-] clwbintrin.h [edit]
[-] rtmintrin.h [edit]
[-] mmintrin.h [edit]
[-] stddef.h [edit]
[-] invpcidintrin.h [edit]
[-] avx512vp2intersectintrin.h [edit]
[-] cet.h [edit]
[-] xopintrin.h [edit]
[-] avx512vlvnniintrin.h [edit]
[-] avx512vlfp16intrin.h [edit]
[-] stdint.h [edit]
[-] arm64intr.h [edit]
[-] sm4intrin.h [edit]
[-] avx512vnniintrin.h [edit]
[-] avx2intrin.h [edit]
[-] movdirintrin.h [edit]
[-] tbmintrin.h [edit]
[-] arm_mve.h [edit]
[-] avx512ifmavlintrin.h [edit]
[-] amxintrin.h [edit]
[-] opencl-c.h [edit]
[-] stdalign.h [edit]
[-] __clang_cuda_device_functions.h [edit]
[-] pconfigintrin.h [edit]
[-] avx512fp16intrin.h [edit]
[-] inttypes.h [edit]
[-] arm_bf16.h [edit]
[-] __clang_cuda_math_forward_declares.h [edit]
[-] vadefs.h [edit]
[-] shaintrin.h [edit]
[-] hexagon_protos.h [edit]
[-] ptwriteintrin.h [edit]
[-] xsaveoptintrin.h [edit]
[-] enqcmdintrin.h [edit]
[-] x86gprintrin.h [edit]
[-] tmmintrin.h [edit]
[-] stdnoreturn.h [edit]
[-] avx512bf16intrin.h [edit]
[-] varargs.h [edit]
[-] s390intrin.h [edit]
[-] avx512vbmiintrin.h [edit]
[+] openmp_wrappers
[-] wmmintrin.h [edit]
[-] __clang_cuda_cmath.h [edit]
[-] clzerointrin.h [edit]
[-] xsavesintrin.h [edit]
[-] __clang_cuda_libdevice_declares.h [edit]
[-] nmmintrin.h [edit]
[-] wasm_simd128.h [edit]
[-] xsavecintrin.h [edit]
[-] avx512dqintrin.h [edit]
[-] lwpintrin.h [edit]
[-] serializeintrin.h [edit]
[-] arm_sme_draft_spec_subject_to_change.h [edit]
[-] avxintrin.h [edit]
[-] __stddef_max_align_t.h [edit]
[-] sm3intrin.h [edit]
[-] velintrin.h [edit]
[-] __clang_cuda_complex_builtins.h [edit]
[-] armintr.h [edit]
[-] avx512cdintrin.h [edit]
[-] float.h [edit]
[-] avx512vbmi2intrin.h [edit]
[-] lzcntintrin.h [edit]
[-] sifive_vector.h [edit]
[-] rdpruintrin.h [edit]
[-] arm_sve.h [edit]
[-] avx512vpopcntdqintrin.h [edit]
[-] xmmintrin.h [edit]
[-] hresetintrin.h [edit]
[-] bmi2intrin.h [edit]
[-] hexagon_types.h [edit]
[-] avx512bwintrin.h [edit]
[-] cetintrin.h [edit]
[-] __clang_cuda_math.h [edit]
[-] avx512vlvbmi2intrin.h [edit]
[-] arm_fp16.h [edit]
[-] avx512vpopcntdqvlintrin.h [edit]
[-] avxvnniint16intrin.h [edit]
[+] llvm_libc_wrappers
[-] __clang_cuda_runtime_wrapper.h [edit]
[-] cldemoteintrin.h [edit]
[-] avx512vldqintrin.h [edit]
[-] avx512vlcdintrin.h [edit]
[-] popcntintrin.h [edit]