diff options
Diffstat (limited to 'src/FbgemmI8Depthwise3DAvx2.cc')
| -rw-r--r-- | src/FbgemmI8Depthwise3DAvx2.cc | 26 |
1 files changed, 17 insertions, 9 deletions
diff --git a/src/FbgemmI8Depthwise3DAvx2.cc b/src/FbgemmI8Depthwise3DAvx2.cc index 925d265..2114b20 100644 --- a/src/FbgemmI8Depthwise3DAvx2.cc +++ b/src/FbgemmI8Depthwise3DAvx2.cc @@ -19,7 +19,7 @@ template < bool SUM_A, bool REMAINDER = false, bool PER_CHANNEL_QUANTIZATION = false> -static inline __attribute__((always_inline)) void inner_prod_3x3x3_packed_( +static inline ALWAYS_INLINE void inner_prod_3x3x3_packed_( int T, int H, int W, @@ -272,7 +272,7 @@ template < bool A_SYMMETRIC, bool B_SYMMETRIC, typename BIAS_TYPE> -static inline __attribute__((always_inline)) void depthwise_3x3x3_kernel_( +static inline ALWAYS_INLINE void depthwise_3x3x3_kernel_( int T, int H, int W, @@ -359,7 +359,7 @@ static inline __attribute__((always_inline)) void depthwise_3x3x3_kernel_( } template <bool FUSE_RELU, bool HAS_BIAS, bool A_SYMMETRIC, typename BIAS_TYPE> -static inline __attribute__((always_inline)) void +static inline ALWAYS_INLINE void depthwise_3x3x3_per_channel_quantization_kernel_( int T, int H, @@ -457,7 +457,7 @@ template < bool A_SYMMETRIC, bool B_SYMMETRIC, typename BIAS_TYPE> -static inline __attribute__((always_inline)) void depthwise_3x3x3_pad_1_( +static inline ALWAYS_INLINE void depthwise_3x3x3_pad_1_( int N, int T, int H, @@ -488,7 +488,9 @@ static inline __attribute__((always_inline)) void depthwise_3x3x3_pad_1_( int W_OUT = (W + PAD_L + PAD_R - K_W) / stride_w + 1; const int8_t* Bp = B.PackedMat(); - int32_t row_offsets[(K + 31) / 32 * 32] __attribute__((aligned(64))); + //int32_t row_offsets[(K + 31) / 32 * 32] __attribute__((aligned(64))); + int32_t* row_offsets + = static_cast<int32_t*>(ALIGNED_MALLOC((K + 31) / 32 * 32 * sizeof(int32_t), 64)); int n_begin, n_end; int t_begin, t_end, h_begin, h_end; @@ -566,10 +568,11 @@ static inline __attribute__((always_inline)) void depthwise_3x3x3_pad_1_( } // h } // t } // for each n + FREE(row_offsets); }; template <bool FUSE_RELU, bool HAS_BIAS, bool A_SYMMETRIC, typename BIAS_TYPE> -static inline __attribute__((always_inline)) void +static inline ALWAYS_INLINE void depthwise_3x3x3_per_channel_quantization_pad_1_( int N, int T, @@ -601,7 +604,9 @@ depthwise_3x3x3_per_channel_quantization_pad_1_( int W_OUT = (W + PAD_L + PAD_R - K_W) / stride_w + 1; const int8_t* Bp = B.PackedMat(); - int32_t row_offsets[(K + 31) / 32 * 32] __attribute__((aligned(64))); + //int32_t row_offsets[(K + 31) / 32 * 32] __attribute__((aligned(64))); + int32_t* row_offsets + = static_cast<int32_t*>(ALIGNED_MALLOC((K + 31) / 32 * 32 * sizeof(int32_t), 64)); int n_begin, n_end; int t_begin, t_end, h_begin, h_end; @@ -679,6 +684,7 @@ depthwise_3x3x3_per_channel_quantization_pad_1_( } // h } // t } // for each n + FREE(row_offsets); }; // Dispatch A_SYMMETRIC and B_SYMMETRIC @@ -704,7 +710,7 @@ static void depthwise_3x3x3_pad_1_( float act_times_w_scale, int thread_id, int num_threads) { - int32_t C_int32_temp[(K + 31) / 32 * 32]; + int32_t* C_int32_temp = new int32_t[(K + 31) / 32 * 32]; if (A_zero_point == 0 || col_offsets == nullptr) { if (B_zero_point == 0) { depthwise_3x3x3_pad_1_< @@ -822,6 +828,7 @@ static void depthwise_3x3x3_pad_1_( num_threads); } } + delete[] C_int32_temp; } // Dispatch HAS_BIAS @@ -1004,7 +1011,7 @@ static void depthwise_3x3x3_per_channel_quantization_pad_1_( const float* act_times_w_scale, int thread_id, int num_threads) { - int32_t C_int32_temp[(K + 31) / 32 * 32]; + int32_t* C_int32_temp = new int32_t[(K + 31) / 32 * 32]; if (A_zero_point == 0 || col_offsets == nullptr) { depthwise_3x3x3_per_channel_quantization_pad_1_< FUSE_RELU, @@ -1060,6 +1067,7 @@ static void depthwise_3x3x3_per_channel_quantization_pad_1_( thread_id, num_threads); } + delete[] C_int32_temp; } // Dispatch HAS_BIAS |