aboutsummaryrefslogtreecommitdiff
path: root/src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp
diff options
context:
space:
mode:
authorFrank Lei <frank.lei@arm.com>2018-02-01 14:47:14 +0800
committerAnthony Barbier <anthony.barbier@arm.com>2018-11-02 16:47:40 +0000
commit4406fd6cc4abded564d3791324e1f48bdfd34273 (patch)
tree22fe402fe9ac7ca338df49e9eccd6eb1587ae875 /src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp
parent898d399a0f62c15612a52df4bff5018e783214e4 (diff)
downloadComputeLibrary-4406fd6cc4abded564d3791324e1f48bdfd34273.tar.gz
APPBROWSER-391: Fix GLES COMPUTE alignment issues
APPBROWSER-402: Performance optimization for squeezenet/xray model Change-Id: If31b186b99a6d6087164019fe94d3ac9279e3204 Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/119526 Tested-by: Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp')
-rw-r--r--src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp12
1 files changed, 9 insertions, 3 deletions
diff --git a/src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp b/src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp
index bc9c7eb55a..fac29024e3 100644
--- a/src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp
+++ b/src/core/GLES_COMPUTE/kernels/GCNormalizePlanarYUVLayerKernel.cpp
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2017 ARM Limited.
+ * Copyright (c) 2017-2018 ARM Limited.
*
* SPDX-License-Identifier: MIT
*
@@ -89,6 +89,8 @@ void GCNormalizePlanarYUVLayerKernel::run(const Window &window)
_kernel.use();
+ _output->set_needs_shifting(true);
+
Window slice = window.first_slice_window_3D();
Window slice_in;
@@ -100,15 +102,19 @@ void GCNormalizePlanarYUVLayerKernel::run(const Window &window)
add_1D_tensor_argument(idx, _mean, 3, slice_in);
add_1D_tensor_argument(idx, _sd, 4, slice_in);
+ slice_in = window.first_slice_window_3D();
+
+ slice.shift(Window::DimX, -(_output->info()->padding()).left);
+
do
{
idx = 0;
- add_3D_tensor_argument(idx, _input, 1, slice);
+ add_3D_tensor_argument(idx, _input, 1, slice_in);
add_3D_tensor_argument(idx, _output, 2, slice);
_kernel.update_shader_params();
enqueue(*this, slice);
}
- while(window.slide_window_slice_3D(slice));
+ while(window.slide_window_slice_3D(slice) && window.slide_window_slice_3D(slice_in));
}