diff options
author | Anthony Barbier <anthony.barbier@arm.com> | 2018-07-16 18:00:11 +0100 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:54:54 +0000 |
commit | 578225e8000c0fc2c4f81a44ae7f06fe44b0fc6a (patch) | |
tree | 828df50795a0cf004b009eaa354893f390bdb29d /arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h | |
parent | d02eb45b7431d6e7fe9ef32e16475c2f9917e62a (diff) | |
download | ComputeLibrary-578225e8000c0fc2c4f81a44ae7f06fe44b0fc6a.tar.gz |
COMPMID-1357: Port NEWinogradConvolution to use NEGEMMAssemblyDispatch
Change-Id: I9a6a1ebf3956b2b5747027f659cb326b86335f50
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/140155
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h')
-rw-r--r-- | arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h | 8 |
1 files changed, 3 insertions, 5 deletions
diff --git a/arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h index c1260977c0..d897ae00e7 100644 --- a/arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h +++ b/arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h @@ -27,11 +27,11 @@ #include "arm_compute/runtime/IFunction.h" #include "arm_compute/core/NEON/INEKernel.h" -#include "arm_compute/core/NEON/kernels/assembly/arm_gemm.hpp" #include "arm_compute/core/Types.h" #include "arm_compute/runtime/CPP/functions/CPPPermute.h" #include "arm_compute/runtime/MemoryGroup.h" #include "arm_compute/runtime/NEON/functions/NEActivationLayer.h" +#include "arm_compute/runtime/NEON/functions/NEGEMMAssemblyDispatch.h" #include "arm_compute/runtime/Tensor.h" #include <memory> @@ -102,9 +102,8 @@ public: NEWinogradConvolutionLayer &operator=(const NEWinogradConvolutionLayer &) = delete; private: - MemoryGroup _memory_group; - std::unique_ptr<arm_gemm::GemmCommon<float, float>> _arm_gemm; - std::unique_ptr<INEKernel> _gemm_kernel; + MemoryGroup _memory_group; + NEGEMMAssemblyDispatchF32 _asm_glue; std::unique_ptr<INEKernel> _transform_input_kernel; std::unique_ptr<INEKernel> _transform_output_kernel; std::unique_ptr<INEKernel> _transform_weights_kernel; @@ -119,7 +118,6 @@ private: Tensor _input_nhwc; Tensor _output_nhwc; Tensor _weights_hwio; - Tensor _workspace; const ITensor *_input; const ITensor *_weights; ITensor *_output; |