ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-21 21:39:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK][CK Tile] Fix kbatch check in grouped conv and gemm kernels (#5555)

Browse Source 
## Motivation

Fix kbatch check in grouped conv and gemm kernels, allow tails for
kbatch.

## Technical Details

Round up K / Kperxdl and divide it by Kbatch to allow tail for K.

## Test Plan

test_grouped_convnd_bwd_weight_tile

## Test Result

passed locally

## Submission Checklist

- [x] Look over the contributing guidelines at
https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.
This commit is contained in:
Bartłomiej Kocot
2026-03-21 23:55:24 +01:00
committed by GitHub
parent 59e6fc67d8
commit ce4525e82b
4 changed files with 10 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
profiler/include/profiler/grouped_convolution_backward_weight_tile_algs.hpp
View File

@@ -178,11 +178,11 @@ run_grouped_conv_backward_weight_tile_algs(const ckt::Args<SIGNATURE>& args,
});
const bool valid = report.get_errors().empty();
best_avg_time = std::min(best_avg_time, avg_time);
best_op_name = best_avg_time < avg_time ? best_op_name : op_name;
best_split_k = best_avg_time < avg_time ? best_split_k : k_batch;
if(valid)
{
best_avg_time = std::min(best_avg_time, avg_time);
best_op_name = best_avg_time < avg_time ? best_op_name : op_name;
best_split_k = best_avg_time < avg_time ? best_split_k : k_batch;
std::cout << "[Valid] Perf: " << std::setw(10) << avg_time << " ms," << " "
<< op_name << ", SplitK " << k_batch << std::endl;
}