Skip to content

[Bug Fix] Reject incompatible combination of --disable-cuda-graph-padding and --enable-torch-compile #23903

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
ShangmingCai merged 1 commit into from
May 12, 2026
Merged

[Bug Fix] Reject incompatible combination of --disable-cuda-graph-padding and --enable-torch-compile #23903

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
8 changes: 8 additions & 0 deletions python/sglang/srt/server_args.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6626,6 +6626,14 @@ def check_server_args(self):
f"Got: {self.pp_max_micro_batch_size}"
)

assert not (self.disable_cuda_graph_padding and self.enable_torch_compile), (
"--disable-cuda-graph-padding is incompatible with --enable-torch-compile. "
"With padding disabled, every distinct batch size gets its own torch.compile + "
"Triton autotune cycle (O(max_batch_size) compilations) instead of the small fixed "
"set of padded bucket sizes, causing engine initialisation to stall for many minutes. "
"Remove --disable-cuda-graph-padding or --enable-torch-compile."
)

if self.pp_size > 1:
assert (
self.disable_overlap_schedule and self.speculative_algorithm is None
Expand Down
Loading