From ada6cbc057ff725e57d301a99a1816ce602485b9 Mon Sep 17 00:00:00 2001 From: Giorgio Arena Date: Fri, 16 Apr 2021 17:03:39 +0100 Subject: Remove OpenCL padding: CLPixelWiseMultiplicationKernel - Change kernel's vec_size to 16 / sizeof(output) - Change ICLKernel.cpp to handle broadcast without padding Resolve COMPMID-3913 Signed-off-by: Giorgio Arena Change-Id: I03e884b250ef5784dc109bff8cf2c96b345d119f Signed-off-by: Giorgio Arena Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5450 Tested-by: Arm Jenkins Comments-Addressed: Arm Jenkins Reviewed-by: Michele Di Giorgio Reviewed-by: Gian Marco Iodice --- src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) (limited to 'src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h') diff --git a/src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h b/src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h index 5889b84938..5b827262a1 100644 --- a/src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h +++ b/src/core/gpu/cl/kernels/ClPixelWiseMultiplicationKernel.h @@ -41,7 +41,7 @@ public: /** Default constructor */ ClPixelWiseMultiplicationKernel() = default; ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClPixelWiseMultiplicationKernel); - /** Initialise the kernel's src, dst and border mode. + /** Initialise the kernel's src and dst. * * Valid configurations (Input1,Input2) -> Output : * @@ -101,10 +101,6 @@ public: // Inherited methods overridden: void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; - BorderSize border_size() const override; - -public: - BorderSize _border_size{}; }; /** Interface for the complex pixelwise multiplication kernel. */ @@ -114,7 +110,7 @@ public: /** Default constructor */ ClComplexPixelWiseMultiplicationKernel() = default; ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClComplexPixelWiseMultiplicationKernel); - /** Initialise the kernel's src, dst and border mode. + /** Initialise the kernel's src and dst. * * @param[in] compile_context The compile context to be used. * @param[in] src1 An src tensor info. Data types supported: F32. Number of channels supported: 2. @@ -136,10 +132,6 @@ public: // Inherited methods overridden: void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override; - BorderSize border_size() const override; - -public: - BorderSize _border_size{}; }; } // namespace kernels } // namespace opencl -- cgit v1.2.1