Skip to content

[BugFix] Fix DeepSeek-R1 hang with DP and MTP #30119

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

Open
LucasWilkinson wants to merge 3 commits into
base: main
Choose a base branch
from
+9 −3
Open

[BugFix] Fix DeepSeek-R1 hang with DP and MTP #30119

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c6f3f31
fix hang
LucasWilkinson Dec 5, 2025
71019e3
cleanup
LucasWilkinson Dec 5, 2025
3f56bb0
Update vllm/v1/worker/gpu_model_runner.py
LucasWilkinson Dec 5, 2025
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
12 changes: 9 additions & 3 deletions vllm/v1/worker/gpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4104,10 +4104,16 @@ def _dummy_run(

if self.speculative_config and self.speculative_config.use_eagle():
assert isinstance(self.drafter, EagleProposer)
# Eagle currently only supports PIECEWISE cudagraphs.
# Therefore only use cudagraphs if the main model uses PIECEWISE
# NOTE(lucas): this is a hack, need to clean up.
use_cudagraphs = (
cudagraph_runtime_mode.has_mode(CUDAGraphMode.PIECEWISE)
and not self.speculative_config.enforce_eager
)
(
is_graph_capturing
and cudagraph_runtime_mode == CUDAGraphMode.PIECEWISE
)
or (cudagraph_runtime_mode != CUDAGraphMode.NONE)
Comment on lines +4111 to +4115
Copy link
Member

@tlrmchlsmth tlrmchlsmth Dec 5, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@LucasWilkinson could you double check this logic?

Copy link
Collaborator Author

@LucasWilkinson LucasWilkinson Dec 5, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the basic idea here is during a DP dummy run we may end up using FULL CGs for the target model; this was set to only using drafter PIECEWISE CGs when he target model was using PIECEWISE CGs (mostly for CG capture) meaning if the target model used FULL CGs this would revert the drafter to eager. This didn't match the behavior of non-dummy runs where if the main model used FULL CGs the drafter would still use PIECEWISE CGs.

) and not self.speculative_config.enforce_eager

# Note(gnovack) - We need to disable cudagraphs for one of the two
# lora cases when cudagraph_specialize_lora is enabled. This is a
Expand Down