ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-03 05:01:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Grouped Conv Bwd Data out index calculation optimizations (#2917)

Browse Source 
* Grouped Conv Bwd Data index calculation optimizations

* fixes

* refactor instances

* gfx12 fixes

* temporary disable splitK for gfx12
This commit is contained in:
Bartłomiej Kocot
2025-09-29 15:59:11 +02:00
committed by GitHub
parent 0f10e6d921
commit 5477811670
17 changed files with 895 additions and 75 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_xdl_cshuffle_v1.hpp
View File

@@ -1485,7 +1485,8 @@ struct DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
static bool IsSupportedArgument(const Argument& arg)
{
// gfx11 doesn't support float atomic
if(ck::is_gfx11_supported() && arg.k_batch_ > 1)
// Todo: Enable splitK for gfx12
if((ck::is_gfx12_supported() || ck::is_gfx11_supported()) && arg.k_batch_ > 1)
{
return false;
}