diff options
| author | Sam Laynton <sam.laynton@arm.com> | 2018-04-05 13:26:08 +0100 |
|---|---|---|
| committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:49:37 +0000 |
| commit | 56e8e8627e5c6912a482a68ec4d65051a5ff4ee8 (patch) | |
| tree | 0af347b19d97fb55f8a231c5e7ab373cefc33ea6 /src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp | |
| parent | 9fb2b667b8cced30c4585e2aa164b92d8d406676 (diff) | |
| download | ComputeLibrary-56e8e8627e5c6912a482a68ec4d65051a5ff4ee8.tar.gz | |
COMPMID-1031: Use LWS hints for G51, G51BIG, G51LIT, and TNOX
Change-Id: Ie07d9225faaef778bdcfdcb56ae42ec95962e48d
Signed-off-by: Sam Laynton <sam.laynton@arm.com>
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/126735
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Tested-by: Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp')
| -rw-r--r-- | src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp b/src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp index 3309775c36..d409fdbc87 100644 --- a/src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp +++ b/src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp @@ -52,7 +52,7 @@ std::pair<Status, Window> validate_and_configure_window(ITensorInfo *accum, ITen unsigned int &num_elems_processed_per_iteration) { // Select the vector size to use (8 for Bifrost; 16 for Midgard). - num_elems_processed_per_iteration = gpu_target_is_in(gpu_target, GPUTarget::G71, GPUTarget::G72) ? 8 : 16; + num_elems_processed_per_iteration = gpu_target_is_in(gpu_target, GPUTarget::G71, GPUTarget::G72, GPUTarget::G51, GPUTarget::G51BIG, GPUTarget::G51LIT, GPUTarget::TNOX) ? 8 : 16; // Configure kernel window Window win = calculate_max_window(*accum, Steps(num_elems_processed_per_iteration)); |