ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-19 04:19:36 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix split N for large images in groupd conv fwd (#2004)

Browse Source 
* Fix split N for large images in groupd conv fwd

* Fix comments

[ROCm/composable_kernel commit: 5b0873c31a]
This commit is contained in:
Bartłomiej Kocot
2025-03-22 23:19:49 +01:00
committed by GitHub
parent e1122c5c27
commit ceb078163f
3 changed files with 12 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
include/ck/tensor_operation/operator_transform/transform_conv_bwd_data_to_gemm_v1.hpp
View File

@@ -1,5 +1,5 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
@@ -106,9 +106,10 @@ struct TransformConvBwdDataToGemm_v1
}
else
{
// Not possible to support even after split N.
// Too large tensor.
return N;
// Split Convolution's N dimension into N workgroups. However
// this still might not result in sufficiently small tensor,
// but at least later on we could divide the image as well.
return 1;
}
}
else

7
include/ck/tensor_operation/operator_transform/transform_conv_fwd_to_gemm.hpp
View File

@@ -83,9 +83,10 @@ struct TransformConvFwdToGemm
}
else
{
// Not possible to support even after split N.
// Too large tensor.
return N;
// Split Convolution's N dimension into N workgroups. However
// this still might not result in sufficiently small tensor,
// but at least later on we could divide the image as well.
return 1;
}
}
else