Skip to content

[AUTOGENERATED] [release/2.8] [ROCm] Improve reduction sum performance #2505

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Aug 15, 2025
Merged

[AUTOGENERATED] [release/2.8] [ROCm] Improve reduction sum performance #2505

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion aten/src/ATen/native/cuda/Reduce.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1062,7 +1062,7 @@ ReduceConfig setReduceConfig(const TensorIterator& iter){
// In such case, values in each loaded vector always correspond to different outputs.
if (fastest_moving_stride == sizeof(scalar_t)) {
#ifdef USE_ROCM
if (reduction_on_fastest_striding_dimension && dim0 > 128 && iter.num_reduce_dims() == 1) {
if (reduction_on_fastest_striding_dimension && dim0 >= 128 && iter.num_reduce_dims() == 1) {
#else
if (reduction_on_fastest_striding_dimension && dim0 > 128 && iter.num_reduce_dims() == 1 && vt0 >= input_vec_size) {
#endif
Expand Down