Skip to content

[core] Turn off GPU communication overlap for Ray executor #12051

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
merged 2 commits into from
Jan 15, 2025
Merged

[core] Turn off GPU communication overlap for Ray executor #12051

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b4f4f14
[core] Turn off GPU communication overlap for Ray executor
ruisearch42 Jan 14, 2025
535ea41
up
ruisearch42 Jan 14, 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
8 changes: 4 additions & 4 deletions vllm/envs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,7 +45,7 @@
VLLM_USE_RAY_SPMD_WORKER: bool = False
VLLM_USE_RAY_COMPILED_DAG: bool = False
VLLM_USE_RAY_COMPILED_DAG_NCCL_CHANNEL: bool = True
VLLM_USE_RAY_COMPILED_DAG_OVERLAP_COMM: bool = True
VLLM_USE_RAY_COMPILED_DAG_OVERLAP_COMM: bool = False
VLLM_WORKER_MULTIPROC_METHOD: str = "fork"
VLLM_ASSETS_CACHE: str = os.path.join(VLLM_CACHE_ROOT, "assets")
VLLM_IMAGE_FETCH_TIMEOUT: int = 5
Expand Down Expand Up @@ -340,11 +340,11 @@ def get_default_config_root():
lambda: bool(int(os.getenv("VLLM_USE_RAY_COMPILED_DAG_NCCL_CHANNEL", "1"))
),

# If the env var is set, it enables GPU communication overlap in
# Ray's compiled DAG. This flag is ignored if
# If the env var is set, it enables GPU communication overlap
# (experimental feature) in Ray's compiled DAG. This flag is ignored if
# VLLM_USE_RAY_COMPILED_DAG is not set.
"VLLM_USE_RAY_COMPILED_DAG_OVERLAP_COMM":
lambda: bool(int(os.getenv("VLLM_USE_RAY_COMPILED_DAG_OVERLAP_COMM", "1"))
lambda: bool(int(os.getenv("VLLM_USE_RAY_COMPILED_DAG_OVERLAP_COMM", "0"))
),

# Use dedicated multiprocess context for workers.
Expand Down
Loading