From 1d08a310b7316f2b731e60ac36dc68989d15b546 Mon Sep 17 00:00:00 2001 From: Georgios Pinitas Date: Wed, 3 Jan 2018 12:29:22 +0000 Subject: COMPMID-765: Collapse execution window in CL kernels. Updated following kernels to collapse their execution window and reduce number of kernel enqueues: -CLArithmeticAddition -CLArithmeticSubtraction -CLPixelWiseMultiplication Change-Id: I13d503515a20fa9be1401ead1e27e9bbc6627975 Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/114878 Reviewed-by: Anthony Barbier Tested-by: Anthony Barbier --- src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp') diff --git a/src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp b/src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp index fd5e5d5862..6dba9c0f95 100644 --- a/src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp +++ b/src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp @@ -227,7 +227,8 @@ void CLPixelWiseMultiplicationKernel::run(const Window &window, cl::CommandQueue ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window); - Window slice = window.first_slice_window_3D(); + Window collapsed = window.collapse_if_possible(ICLKernel::window(), Window::DimZ); + Window slice = collapsed.first_slice_window_3D(); do { @@ -237,5 +238,5 @@ void CLPixelWiseMultiplicationKernel::run(const Window &window, cl::CommandQueue add_3D_tensor_argument(idx, _output, slice); enqueue(queue, *this, slice); } - while(window.slide_window_slice_3D(slice)); + while(collapsed.slide_window_slice_3D(slice)); } -- cgit v1.2.1