Skip to content

[Log] Only Print Profiler Results on Rank 0 #23370

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
yewentao256 merged 7 commits into from
Sep 2, 2025
Merged

[Log] Only Print Profiler Results on Rank 0 #23370

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
07b0b7c
only print profiler results on rank 0
yewentao256 Aug 21, 2025
e7f1143
use local rank
yewentao256 Aug 28, 2025
f9b3f31
Merge branch 'main' into wye-only-print-cuda-profile-time-on-rank-0
yewentao256 Aug 28, 2025
470055b
Merge branch 'main' into wye-only-print-cuda-profile-time-on-rank-0
yewentao256 Aug 28, 2025
8334366
Merge branch 'main' into wye-only-print-cuda-profile-time-on-rank-0
yewentao256 Aug 29, 2025
867918a
Merge branch 'main' into wye-only-print-cuda-profile-time-on-rank-0
yewentao256 Aug 29, 2025
5ab8983
Merge branch 'main' into wye-only-print-cuda-profile-time-on-rank-0
yewentao256 Aug 31, 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
6 changes: 4 additions & 2 deletions vllm/v1/worker/gpu_worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -400,8 +400,10 @@ def profile(self, is_start: bool = True):
self.profiler.start()
else:
self.profiler.stop()
print(self.profiler.key_averages().table(
sort_by="self_cuda_time_total"))
# only print profiler results on rank 0
if self.local_rank == 0:
print(self.profiler.key_averages().table(
sort_by="self_cuda_time_total"))

def execute_dummy_batch(self) -> None:
self.model_runner._dummy_run(1)
Expand Down
6 changes: 4 additions & 2 deletions vllm/worker/worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -128,8 +128,10 @@ def stop_profile(self):
if self.profiler is None:
raise RuntimeError("Profiler is not enabled.")
self.profiler.stop()
print(
self.profiler.key_averages().table(sort_by="self_cuda_time_total"))
# only print profiler results on rank 0
if self.local_rank == 0:
print(self.profiler.key_averages().table(
sort_by="self_cuda_time_total"))

def sleep(self, level: int = 1) -> None:
free_bytes_before_sleep = torch.cuda.mem_get_info()[0]
Expand Down