diff options
author | SiCong Li <sicong.li@arm.com> | 2023-06-28 09:49:45 +0100 |
---|---|---|
committer | SiCong Li <sicong.li@arm.com> | 2023-07-25 15:48:50 +0000 |
commit | 23882a9014eb3972bca958206866c8e0d0b829cc (patch) | |
tree | 9139b91699099160e26a64abd8cf182bd7447278 /src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp | |
parent | 0a59e69fd922b02d9e3b5b043ee7f891061df7be (diff) | |
download | ComputeLibrary-23882a9014eb3972bca958206866c8e0d0b829cc.tar.gz |
Add GpuKernelArgumentBinding for runtime argument setting
* Add flexible runtime argument setting that accept argument bindings
exported from ckw.
* Introduce internal build flag ACL_INTERNAL_TEST_CKW_IN_DF. If set to
true, ckw will be tested in dynamic fusion validation tests. Otherwise
it will not be tested and the dynamic fusion will keep using
ClTemplateWriter instead.
* Fix CKW sampler for elementwise binary to deal with tile sizes > 1
in both dimensions
Resolves: COMPMID-6282
Partially resolves: COMPMID-6260
Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: I0ab225a4484eb2119643d900a4e72806558626ee
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9917
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Anitha Raj <Anitha.Raj@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp')
-rw-r--r-- | src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp | 38 |
1 files changed, 22 insertions, 16 deletions
diff --git a/src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp b/src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp index b273c2a20c..84fb279237 100644 --- a/src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp +++ b/src/dynamic_fusion/runtime/gpu/cl/ckw_driver/GpuCkwKernelArgumentsHelpers.cpp @@ -26,7 +26,11 @@ namespace arm_compute { -void cl_add_tensor_component_argument(cl::Kernel &kernel, unsigned int &idx, ICLTensor *tensor, ckw::TensorComponentType component) +namespace experimental +{ +namespace dynamic_fusion +{ +void cl_add_tensor_component_argument(cl::Kernel &kernel, unsigned int &idx, const ICLTensor *tensor, TensorComponentType component) { ARM_COMPUTE_ERROR_ON(tensor == nullptr); @@ -35,49 +39,49 @@ void cl_add_tensor_component_argument(cl::Kernel &kernel, unsigned int &idx, ICL switch(component) { - case ckw::TensorComponentType::OffsetFirstElement: + case TensorComponentType::OffsetFirstElement: kernel.setArg<cl_uint>(idx++, info->offset_first_element_in_bytes()); break; - case ckw::TensorComponentType::Stride0: + case TensorComponentType::Stride0: kernel.setArg<cl_uint>(idx++, strides[0]); break; - case ckw::TensorComponentType::Stride1: + case TensorComponentType::Stride1: kernel.setArg<cl_uint>(idx++, strides[1]); break; - case ckw::TensorComponentType::Stride2: + case TensorComponentType::Stride2: kernel.setArg<cl_uint>(idx++, strides[2]); break; - case ckw::TensorComponentType::Stride3: + case TensorComponentType::Stride3: kernel.setArg<cl_uint>(idx++, strides[3]); break; - case ckw::TensorComponentType::Stride4: + case TensorComponentType::Stride4: kernel.setArg<cl_uint>(idx++, strides[4]); break; - case ckw::TensorComponentType::Dim0: + case TensorComponentType::Dim0: kernel.setArg<cl_uint>(idx++, info->dimension(0)); break; - case ckw::TensorComponentType::Dim1: + case TensorComponentType::Dim1: kernel.setArg<cl_uint>(idx++, info->dimension(1)); break; - case ckw::TensorComponentType::Dim2: + case TensorComponentType::Dim2: kernel.setArg<cl_uint>(idx++, info->dimension(2)); break; - case ckw::TensorComponentType::Dim3: + case TensorComponentType::Dim3: kernel.setArg<cl_uint>(idx++, info->dimension(3)); break; - case ckw::TensorComponentType::Dim4: + case TensorComponentType::Dim4: kernel.setArg<cl_uint>(idx++, info->dimension(4)); break; - case ckw::TensorComponentType::Dim1xDim2: + case TensorComponentType::Dim1xDim2: kernel.setArg<cl_uint>(idx++, info->dimension(1) * info->dimension(2)); break; - case ckw::TensorComponentType::Dim2xDim3: + case TensorComponentType::Dim2xDim3: kernel.setArg<cl_uint>(idx++, info->dimension(2) * info->dimension(3)); break; - case ckw::TensorComponentType::Dim1xDim2xDim3: + case TensorComponentType::Dim1xDim2xDim3: kernel.setArg<cl_uint>(idx++, info->dimension(1) * info->dimension(2) * info->dimension(3)); break; - case ckw::TensorComponentType::Unknown: + case TensorComponentType::Unknown: default: ARM_COMPUTE_ERROR("Unknown tensor component"); } @@ -93,4 +97,6 @@ void cl_add_texture_argument(cl::Kernel &kernel, unsigned int &idx, const cl::Im kernel.setArg(idx++, image); } +} // namespace dynamic_fusion +} // namespace experimental } // namespace arm_compute |