clang 19.0.0git
Macros | Functions
avx512vlvnniintrin.h File Reference

Go to the source code of this file.

Macros

#define __DEFAULT_FN_ATTRS128
 
#define __DEFAULT_FN_ATTRS256
 
#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 with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.
 
#define _mm256_dpbusds_epi32(S, A, B)    ((__m256i)__builtin_ia32_vpdpbusds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))
 Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.
 
#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 with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.
 
#define _mm256_dpwssds_epi32(S, A, B)    ((__m256i)__builtin_ia32_vpdpwssds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))
 Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.
 
#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 with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.
 
#define _mm_dpbusds_epi32(S, A, B)    ((__m128i)__builtin_ia32_vpdpbusds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))
 Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.
 
#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 with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.
 
#define _mm_dpwssds_epi32(S, A, B)    ((__m128i)__builtin_ia32_vpdpwssds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))
 Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.
 

Functions

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusd_epi32 (__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpbusd_epi32 (__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusds_epi32 (__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpbusds_epi32 (__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssd_epi32 (__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpwssd_epi32 (__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssds_epi32 (__m256i __S, __mmask8 __U, __m256i __A, __m256i __B)
 
static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpwssds_epi32 (__mmask8 __U, __m256i __S, __m256i __A, __m256i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusd_epi32 (__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpbusd_epi32 (__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusds_epi32 (__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpbusds_epi32 (__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssd_epi32 (__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpwssd_epi32 (__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssds_epi32 (__m128i __S, __mmask8 __U, __m128i __A, __m128i __B)
 
static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpwssds_epi32 (__mmask8 __U, __m128i __S, __m128i __A, __m128i __B)
 

Macro Definition Documentation

◆ __DEFAULT_FN_ATTRS128

#define __DEFAULT_FN_ATTRS128
Value:
__attribute__((__always_inline__, __nodebug__, \
__target__("avx512vl,avx512vnni,no-evex512"), \
__min_vector_width__(128)))
_Float16 __2f16 __attribute__((ext_vector_type(2)))
Zeroes the upper 128 bits (bits 255:128) of all YMM registers.

Definition at line 18 of file avx512vlvnniintrin.h.

◆ __DEFAULT_FN_ATTRS256

#define __DEFAULT_FN_ATTRS256
Value:
__attribute__((__always_inline__, __nodebug__, \
__target__("avx512vl,avx512vnni,no-evex512"), \
__min_vector_width__(256)))

Definition at line 22 of file avx512vlvnniintrin.h.

◆ _mm256_dpbusd_epi32

#define _mm256_dpbusd_epi32 (   S,
  A,
 
)     ((__m256i)__builtin_ia32_vpdpbusd256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.

Sum these 4 results with the corresponding 32-bit integer in S, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPBUSD instructions.

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

Definition at line 44 of file avx512vlvnniintrin.h.

◆ _mm256_dpbusds_epi32

#define _mm256_dpbusds_epi32 (   S,
  A,
 
)     ((__m256i)__builtin_ia32_vpdpbusds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.

Sum these 4 results with the corresponding 32-bit integer in S using signed saturation, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPBUSDS instructions.

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

Definition at line 64 of file avx512vlvnniintrin.h.

◆ _mm256_dpwssd_epi32

#define _mm256_dpwssd_epi32 (   S,
  A,
 
)     ((__m256i)__builtin_ia32_vpdpwssd256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.

Sum these 2 results with the corresponding 32-bit integer in S, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPWSSD instructions.

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

Definition at line 82 of file avx512vlvnniintrin.h.

◆ _mm256_dpwssds_epi32

#define _mm256_dpwssds_epi32 (   S,
  A,
 
)     ((__m256i)__builtin_ia32_vpdpwssds256((__v8si)(S), (__v8si)(A), (__v8si)(B)))

Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.

Sum these 2 results with the corresponding 32-bit integer in S using signed saturation, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPWSSDS instructions.

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

Definition at line 100 of file avx512vlvnniintrin.h.

◆ _mm_dpbusd_epi32

#define _mm_dpbusd_epi32 (   S,
  A,
 
)     ((__m128i)__builtin_ia32_vpdpbusd128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.

Sum these 4 results with the corresponding 32-bit integer in S, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPBUSD instructions.

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

Definition at line 120 of file avx512vlvnniintrin.h.

◆ _mm_dpbusds_epi32

#define _mm_dpbusds_epi32 (   S,
  A,
 
)     ((__m128i)__builtin_ia32_vpdpbusds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

Multiply groups of 4 adjacent pairs of unsigned 8-bit integers in A with corresponding signed 8-bit integers in B, producing 4 intermediate signed 16-bit results.

Sum these 4 results with the corresponding 32-bit integer in S using signed saturation, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPBUSDS instructions.

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

Definition at line 140 of file avx512vlvnniintrin.h.

◆ _mm_dpwssd_epi32

#define _mm_dpwssd_epi32 (   S,
  A,
 
)     ((__m128i)__builtin_ia32_vpdpwssd128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.

Sum these 2 results with the corresponding 32-bit integer in S, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPWSSD instructions.

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

Definition at line 158 of file avx512vlvnniintrin.h.

◆ _mm_dpwssds_epi32

#define _mm_dpwssds_epi32 (   S,
  A,
 
)     ((__m128i)__builtin_ia32_vpdpwssds128((__v4si)(S), (__v4si)(A), (__v4si)(B)))

Multiply groups of 2 adjacent pairs of signed 16-bit integers in A with corresponding 16-bit integers in B, producing 2 intermediate signed 32-bit results.

Sum these 2 results with the corresponding 32-bit integer in S using signed saturation, and store the packed 32-bit results in DST.

This intrinsic corresponds to the VPDPWSSDS instructions.

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

Definition at line 176 of file avx512vlvnniintrin.h.

Function Documentation

◆ _mm256_mask_dpbusd_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusd_epi32 ( __m256i  __S,
__mmask8  __U,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 180 of file avx512vlvnniintrin.h.

References _mm256_dpbusd_epi32.

◆ _mm256_mask_dpbusds_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpbusds_epi32 ( __m256i  __S,
__mmask8  __U,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 196 of file avx512vlvnniintrin.h.

References _mm256_dpbusds_epi32.

◆ _mm256_mask_dpwssd_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssd_epi32 ( __m256i  __S,
__mmask8  __U,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 212 of file avx512vlvnniintrin.h.

References _mm256_dpwssd_epi32.

◆ _mm256_mask_dpwssds_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_mask_dpwssds_epi32 ( __m256i  __S,
__mmask8  __U,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 228 of file avx512vlvnniintrin.h.

References _mm256_dpwssds_epi32.

◆ _mm256_maskz_dpbusd_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpbusd_epi32 ( __mmask8  __U,
__m256i  __S,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 188 of file avx512vlvnniintrin.h.

References _mm256_dpbusd_epi32, and _mm256_setzero_si256().

◆ _mm256_maskz_dpbusds_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpbusds_epi32 ( __mmask8  __U,
__m256i  __S,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 204 of file avx512vlvnniintrin.h.

References _mm256_dpbusds_epi32, and _mm256_setzero_si256().

◆ _mm256_maskz_dpwssd_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpwssd_epi32 ( __mmask8  __U,
__m256i  __S,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 220 of file avx512vlvnniintrin.h.

References _mm256_dpwssd_epi32, and _mm256_setzero_si256().

◆ _mm256_maskz_dpwssds_epi32()

static __inline__ __m256i __DEFAULT_FN_ATTRS256 _mm256_maskz_dpwssds_epi32 ( __mmask8  __U,
__m256i  __S,
__m256i  __A,
__m256i  __B 
)
static

Definition at line 236 of file avx512vlvnniintrin.h.

References _mm256_dpwssds_epi32, and _mm256_setzero_si256().

◆ _mm_mask_dpbusd_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusd_epi32 ( __m128i  __S,
__mmask8  __U,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 244 of file avx512vlvnniintrin.h.

References _mm_dpbusd_epi32.

◆ _mm_mask_dpbusds_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpbusds_epi32 ( __m128i  __S,
__mmask8  __U,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 260 of file avx512vlvnniintrin.h.

References _mm_dpbusds_epi32.

◆ _mm_mask_dpwssd_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssd_epi32 ( __m128i  __S,
__mmask8  __U,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 276 of file avx512vlvnniintrin.h.

References _mm_dpwssd_epi32.

◆ _mm_mask_dpwssds_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_mask_dpwssds_epi32 ( __m128i  __S,
__mmask8  __U,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 292 of file avx512vlvnniintrin.h.

References _mm_dpwssds_epi32.

◆ _mm_maskz_dpbusd_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpbusd_epi32 ( __mmask8  __U,
__m128i  __S,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 252 of file avx512vlvnniintrin.h.

References _mm_dpbusd_epi32, and _mm_setzero_si128().

◆ _mm_maskz_dpbusds_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpbusds_epi32 ( __mmask8  __U,
__m128i  __S,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 268 of file avx512vlvnniintrin.h.

References _mm_dpbusds_epi32, and _mm_setzero_si128().

◆ _mm_maskz_dpwssd_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpwssd_epi32 ( __mmask8  __U,
__m128i  __S,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 284 of file avx512vlvnniintrin.h.

References _mm_dpwssd_epi32, and _mm_setzero_si128().

◆ _mm_maskz_dpwssds_epi32()

static __inline__ __m128i __DEFAULT_FN_ATTRS128 _mm_maskz_dpwssds_epi32 ( __mmask8  __U,
__m128i  __S,
__m128i  __A,
__m128i  __B 
)
static

Definition at line 300 of file avx512vlvnniintrin.h.

References _mm_dpwssds_epi32, and _mm_setzero_si128().