commit d02d5edfa15ba6c04a9986a8a362a945cb38ac31
author Michele Di Giorgio <michele.digiorgio@arm.com> Fri Jan 22 09:47:04 2021 +0000
committer Michele Di Giorgio <michele.digiorgio@arm.com> Fri Jun 18 10:33:48 2021 +0000
tree ced4f49691d6c7038e347a8709b315bff59c64cf
parent b014c27ba6db9840e4a72519760d51a87a2af7e7

Integrate improved CPU depthwise convolution kernels

* Replace assembly kernels for depthwise convolution with more optimized
  ones.
* Add int8 assembly kernels.
* Fix implicit padding on optimized kernels

Resolves: COMPMID-3867, COMPMID-4361

Change-Id: I0b0867e05f61be4f368f62190d55e14d0ab3ebf2
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5622
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>

tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h

diff --git a/tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h b/tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h
index 2c94373..19ec6b2 100644
--- a/tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h
+++ b/tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h
@@ -117,8 +117,7 @@
 
     void allocate_and_run_target()
     {
-        // TODO: uncomment after COMPMID-4361
-        // add_padding_x({ &_src, &_weights, &_biases, &_target }, _data_layout);
+        add_padding_x({ &_src, &_weights, &_biases, &_target }, _data_layout);
 
         // Allocate tensors
         _src.allocator()->allocate();