Home
last modified time | relevance | path

Searched refs:vxb23 (Results 1 – 9 of 9) sorted by relevance

/external/XNNPACK/src/qs8-gemm/gen/
D3x8c8-minmax-avx2.c98 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2() local
100 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2()
101 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2()
102 vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23)); in xnn_qs8_gemm_minmax_ukernel_3x8c8__avx2()
D3x8c8-xw-minmax-avx2.c96 … const __m256i vxb23 = _mm256_load_si256((const __m256i*) ((uintptr_t) w + 16 * sizeof(int16_t))); in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2() local
98 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2()
99 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2()
100 vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_3x8c8__avx2()
D2x8c8-xw-minmax-avx2.c82 … const __m256i vxb23 = _mm256_load_si256((const __m256i*) ((uintptr_t) w + 16 * sizeof(int16_t))); in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2() local
84 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2()
85 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_2x8c8__avx2()
D2x8c8-minmax-avx2.c84 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2() local
86 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2()
87 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_gemm_minmax_ukernel_2x8c8__avx2()
D1x8c8-xw-minmax-avx2.c68 … const __m256i vxb23 = _mm256_load_si256((const __m256i*) ((uintptr_t) w + 16 * sizeof(int16_t))); in xnn_qs8_gemm_xw_minmax_ukernel_1x8c8__avx2() local
70 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_xw_minmax_ukernel_1x8c8__avx2()
D1x8c8-minmax-avx2.c70 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_gemm_minmax_ukernel_1x8c8__avx2() local
72 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_gemm_minmax_ukernel_1x8c8__avx2()
/external/XNNPACK/src/qs8-igemm/gen/
D3x8c8-minmax-avx2.c113 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2() local
115 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2()
116 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2()
117 vacc2x23 = _mm256_add_epi32(vacc2x23, _mm256_madd_epi16(vxa2, vxb23)); in xnn_qs8_igemm_minmax_ukernel_3x8c8__avx2()
D2x8c8-minmax-avx2.c97 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2() local
99 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2()
100 vacc1x23 = _mm256_add_epi32(vacc1x23, _mm256_madd_epi16(vxa1, vxb23)); in xnn_qs8_igemm_minmax_ukernel_2x8c8__avx2()
D1x8c8-minmax-avx2.c81 const __m256i vxb23 = _mm256_cvtepi8_epi16(vb23); in xnn_qs8_igemm_minmax_ukernel_1x8c8__avx2() local
83 vacc0x23 = _mm256_add_epi32(vacc0x23, _mm256_madd_epi16(vxa0, vxb23)); in xnn_qs8_igemm_minmax_ukernel_1x8c8__avx2()