diff options
| author | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-04-22 21:13:21 +0100 |
|---|---|---|
| committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-05-18 14:48:39 +0000 |
| commit | 856f66e6c61b77d03f754cd0fa8439891f0e4aca (patch) | |
| tree | f9379cd0853ac407109e54c3d53b385ceee066c2 /src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp | |
| parent | 37f4b2ef1ea225a90ccb563fcb2c08f8fb0fb5d5 (diff) | |
| download | ComputeLibrary-856f66e6c61b77d03f754cd0fa8439891f0e4aca.tar.gz | |
Port CLGEMM to memory injecting interface
Moves the following kernels:
- CLGEMMMatrixMultiplyKernel
- CLGEMMMatrixMultiplyNativeKernel
- CLGEMMMatrixMultipluReshapedKernel
- CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
Moves the following functions
- CLGEMM
Introduces facilities to easy handling of auxiliary temporary buffers
under then new run interface. Such are:
- CLAuxTensorHandler: That allows wrapping of workspace buffers memory
to CLBuffer objects
- Ability to inject TensorInfo to allocator without transferring
ownership. This reduce the copy overhead if needed.
Resolves: COMPMID-4188
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I7055435d831b05b749b26302082e4ac45f26dfb0
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5498
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp')
| -rw-r--r-- | src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp b/src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp index 18d648d2f2..0a5101f564 100644 --- a/src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp +++ b/src/core/gpu/cl/kernels/ClDirectConvolutionKernel.cpp @@ -35,7 +35,7 @@ #include "src/core/AccessWindowStatic.h" #include "src/core/CL/CLUtils.h" #include "src/core/CL/CLValidate.h" -#include "src/core/CL/gemm/CLGEMMHelpers.h" +#include "src/core/gpu/cl/kernels/gemm/ClGemmHelpers.h" #include "src/core/helpers/AutoConfiguration.h" #include "src/core/helpers/WindowHelpers.h" #include "support/Cast.h" @@ -416,7 +416,7 @@ void ClDirectConvolutionKernel::configure(const CLCompileContext &compile_contex const unsigned int n0 = win_config.second.x().step(); const unsigned int m0 = win_config.second.y().step(); - const unsigned int k0 = adjust_vec_size(is_data_type_quantized(data_type)? 16u : 8u, src->dimension(channel_idx)); + const unsigned int k0 = adjust_vec_size(is_data_type_quantized(data_type) ? 16u : 8u, src->dimension(channel_idx)); const unsigned int partial_store_n0 = dst->dimension(channel_idx) % n0; const unsigned int pad_left = conv_info.pad_left(); const unsigned int pad_top = conv_info.pad_top(); @@ -425,7 +425,7 @@ void ClDirectConvolutionKernel::configure(const CLCompileContext &compile_contex // Update the padding for the weights tensor if we can export to cl_image if(export_to_cl_image) { - arm_compute::cl_gemm::update_padding_for_cl_image(weights); + gemm::update_padding_for_cl_image(weights); } if(biases != nullptr) |