diff options
Diffstat (limited to 'src/PackDepthwiseConvMatrixAvx2.cc')
| -rw-r--r-- | src/PackDepthwiseConvMatrixAvx2.cc | 28 |
1 files changed, 18 insertions, 10 deletions
diff --git a/src/PackDepthwiseConvMatrixAvx2.cc b/src/PackDepthwiseConvMatrixAvx2.cc index 0e17bcd..ab2e1f2 100644 --- a/src/PackDepthwiseConvMatrixAvx2.cc +++ b/src/PackDepthwiseConvMatrixAvx2.cc @@ -5,6 +5,8 @@ * LICENSE file in the root directory of this source tree. */ #include "fbgemm/FbgemmI8DepthwiseAvx2.h" +#include "fbgemm/Utils.h" +#include "fbgemm/Fbgemm.h" #include <immintrin.h> @@ -33,7 +35,9 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( const int8_t* smat) : K_(K), kernel_prod_(kernel_prod) { // Transpose the input matrix to make packing faster. - alignas(64) int8_t smat_transposed[K * kernel_prod]; + int8_t* smat_transposed + = static_cast<int8_t*>(ALIGNED_MALLOC(K * kernel_prod * sizeof(int8_t), 64)); + for (int i = 0; i < kernel_prod; ++i) { for (int j = 0; j < K; ++j) { smat_transposed[i * K + j] = smat[i + j * kernel_prod]; @@ -42,12 +46,11 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( // Allocate packed arrays int kernel_prod_aligned = (kernel_prod + 1) / 2 * 2; - // pmat_ = static_cast<int8_t *>(fbgemmAlignedAlloc( - // 64, ((K + 31) / 32) * KERNEL_PROD_ALIGNED * 32 * sizeof(int8_t))); - posix_memalign( - (void**)&pmat_, - 64, - ((K + 31) / 32) * kernel_prod_aligned * 32 * sizeof(int8_t)); + pmat_ = static_cast<int8_t *>(fbgemmAlignedAlloc(64, ((K + 31) / 32) * kernel_prod_aligned * 32 * sizeof(int8_t))); + //posix_memalign( + // (void**)&pmat_, + // 64, + // ((K + 31) / 32) * kernel_prod_aligned * 32 * sizeof(int8_t)); // Pack input matrix // The layout is optimized to use vpmaddubsw efficiently (see @@ -102,7 +105,7 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( // (12, 8), (12, 9), (12, 10), zero, ..., (15, 8), (15, 9), (15, 10), zero // (28, 8), (28, 9), (28, 10), zero, ..., (31, 8), (31, 9), (31, 10), zero for (int k1 = 0; k1 < K; k1 += 32) { - __m256i b_v[kernel_prod]; + __m256i* b_v = new __m256i[kernel_prod]; int remainder = K - k1; if (remainder < 32) { __m256i mask_v = _mm256_loadu_si256( @@ -119,7 +122,7 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( } // Interleave 2 SIMD registers - __m256i b_interleaved_epi16[kernel_prod_aligned]; + __m256i* b_interleaved_epi16 = new __m256i[kernel_prod_aligned]; __m256i zero_v = _mm256_setzero_si256(); for (int i = 0; i < kernel_prod_aligned / 2; ++i) { if (2 * i + 1 >= kernel_prod) { @@ -135,7 +138,7 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( } // Interleave 4 SIMD registers - __m256i b_interleaved_epi32[kernel_prod_aligned]; + __m256i* b_interleaved_epi32 = new __m256i[kernel_prod_aligned]; for (int i = 0; i < kernel_prod_aligned / 4; ++i) { b_interleaved_epi32[4 * i] = _mm256_unpacklo_epi16( b_interleaved_epi16[4 * i], b_interleaved_epi16[4 * i + 2]); @@ -156,7 +159,12 @@ PackedDepthWiseConvMatrix::PackedDepthWiseConvMatrix( &pmat_[((k1 / 32) * kernel_prod_aligned + i) * 32]), b_interleaved_epi32[i]); } + + delete[] b_v; + delete[] b_interleaved_epi16; + delete[] b_interleaved_epi32; } + FREE(smat_transposed); } int PackedDepthWiseConvMatrix::addr(int r, int c) { |