Skip to content

Ensure GemmTuner doesn't generate invalid input combinations for SplitK kernels #2721

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
apicciau wants to merge 1 commit into
base: main
Choose a base branch
from
+7 −0
Draft

Ensure GemmTuner doesn't generate invalid input combinations for SplitK kernels #2721

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: 7 additions & 0 deletions gradlib/gradlib/GemmTuner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -396,6 +396,13 @@ def asm_gemm_all_solutions(self):
)
if self.k / splitK < subK:
break
# splitK kernels use a semaphore array of size gdx*gdy; skip
# candidates where the grid exceeds the 1024-entry limit.
if splitK > 1:
gdx = (self.n + tile_n - 1) // tile_n
gdy = (self.m + tile_m - 1) // tile_m
if gdx * gdy > 1024:
continue
task_asm.append(
(
info,
Expand Down
Loading