diff options
author | Dana Zlotnik <dana.zlotnik@arm.com> | 2021-10-07 15:31:54 +0300 |
---|---|---|
committer | Dana Zlotnik <dana.zlotnik@arm.com> | 2021-10-18 09:37:41 +0000 |
commit | 4cdd6b80754b3abbf54650d9359cf940a4aaf772 (patch) | |
tree | 6f0096f6d2abbc8410d4aec59f96263a05af5c2e /src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp | |
parent | 6d9c982a5aec543d1f7f198f0fee10a7a3a78ddb (diff) | |
download | ComputeLibrary-4cdd6b80754b3abbf54650d9359cf940a4aaf772.tar.gz |
Implement Minimum Workload Size (MWS) in all CPPKernels used by small networks
* create get_mws method in ICPPKernel class that retuns default value for all kernels
* overwrite the default value for all the kernels used by small networks (according to banchmark case)
Resolves COMPMID-4648
Change-Id: I46d7cae61217213279d2ee740edc73f600b6d576
Signed-off-by: Dana Zlotnik <dana.zlotnik@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6412
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp')
-rw-r--r-- | src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp b/src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp index eed4bb9dd5..a71864c10c 100644 --- a/src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp +++ b/src/cpu/kernels/internal/CpuDepthwiseConv2dAssemblyWrapperKernel.cpp @@ -354,6 +354,13 @@ const char *CpuDepthwiseConv2dAssemblyWrapperKernel::name() const { return "CpuDepthwiseConv2dAssemblyWrapperKernel"; } + +size_t CpuDepthwiseConv2dAssemblyWrapperKernel::get_mws(const CPUInfo &platform, size_t thread_count) const +{ + ARM_COMPUTE_UNUSED(platform, thread_count); + + return ICPPKernel::small_network_mws; +} } // namespace kernels } // namespace cpu } // namespace arm_compute |