Skip to content

revert max split heuristics #121

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

Draft
wants to merge 1 commit into
base: lwilkinson/fix-cutlass-mla-hang
Choose a base branch
from
Draft

revert max split heuristics #121

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
7 changes: 0 additions & 7 deletions csrc/attention/mla/cutlass_sm100_mla/device/sm100_mla.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,13 +134,6 @@ class MLA {
int max_splits = ceil_div(K, 128);
max_splits = min(16, max_splits);

// TODO: This avoids a hang when the batch size larger than 1 and
// there is more than 1 kv_splits.
// Discuss with NVIDIA how this can be fixed.
if (B > 1) {
max_splits = min(1, max_splits);
}

// printf(" max_splits = %d\n", max_splits);
int sms_per_batch = max(1, sm_count / B);
// printf(" sms_per_batch = %d\n", sms_per_batch);
Expand Down