Skip to content

[Fix] Prevent symbolic over-unification in multi-modality torch.split and add comprehensive tests #26

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
jiahy0825 merged 1 commit into from
Apr 26, 2026
+663 −1
Merged

[Fix] Prevent symbolic over-unification in multi-modality torch.split and add comprehensive tests #26

Show file tree
Hide file tree
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
9 changes: 8 additions & 1 deletion magi_compiler/magi_backend/piecewise_compiler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,7 +173,14 @@ def compile(
) -> tuple[Callable | None, CacheHandle | None]:
# Step1: Update compile settings
compilation_counter.num_inductor_compiles += 1
current_config = {}
current_config = {
# standalone_compile hardcodes autotune_at_compile_time=True, but
# Triton autotune benchmarks with unbacked SymInt dimensions cause
# CUDA illegal-memory-access errors. Disable compile-time autotune
# so that tuning happens at first runtime invocation instead (same
# kernel quality, avoids the crash, and tunes on real shapes).
"triton.autotune_at_compile_time": False
}
if inductor_compile_config is not None:
current_config.update(inductor_compile_config)
if isinstance(runtime_shape, int):
Expand Down
Loading