Skip to content

[V1][PP] Fix memory profiling in PP #13315

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
WoosukKwon merged 1 commit into from
Feb 15, 2025
Merged

[V1][PP] Fix memory profiling in PP #13315

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
11 changes: 6 additions & 5 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 @@ -1158,11 +1158,12 @@ def profile_run(self) -> None:
# Trigger compilation for general shape.
hidden_states = self._dummy_run(self.max_num_tokens,
dummy_kv_caches)
if not get_pp_group().is_last_rank:
return hidden_states
hidden_states = hidden_states[logit_indices]
logits = self.model.compute_logits(hidden_states, None)
# TODO(woosuk): Consider the memory usage of the sampler.
if get_pp_group().is_last_rank:
hidden_states = hidden_states[logit_indices]
logits = self.model.compute_logits(hidden_states, None)
# TODO(woosuk): Consider the memory usage of the sampler.
else:
logits = None
torch.cuda.synchronize()
del hidden_states, logits
self.encoder_cache.clear()
Expand Down