Skip to content
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

[misc][distributed] add VLLM_SKIP_P2P_CHECK flag #8911

Merged
merged 2 commits into from
Sep 27, 2024
Merged

[misc][distributed] add VLLM_SKIP_P2P_CHECK flag #8911

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
901c3f2
enable skip
youkaichao Sep 27, 2024
a854d5d
add logging
youkaichao Sep 27, 2024
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
4 changes: 4 additions & 0 deletions vllm/distributed/device_communicators/custom_all_reduce.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,10 @@ def _can_p2p(rank: int, world_size: int) -> bool:
for i in range(world_size):
if i == rank:
continue
if envs.VLLM_SKIP_P2P_CHECK:
logger.info(
"Skipping P2P check and trusting the driver's P2P report.")
return torch.cuda.can_device_access_peer(rank, i)
Copy link
Contributor

Choose a reason for hiding this comment

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

how long will this take? torch.cuda.can_device_access_peer(rank, i)

Copy link
Member Author

Choose a reason for hiding this comment

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

this is just one api call, probably less than ms .

if not gpu_p2p_access_check(rank, i):
return False
return True
Expand Down
8 changes: 8 additions & 0 deletions vllm/envs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,7 @@
VLLM_USE_TRITON_AWQ: bool = False
VLLM_ALLOW_RUNTIME_LORA_UPDATING: bool = False
VLLM_ALLOW_DEPRECATED_BEAM_SEARCH: bool = False
VLLM_SKIP_P2P_CHECK: bool = False


def get_default_cache_root():
Expand Down Expand Up @@ -423,6 +424,13 @@ def get_default_config_root():
lambda:
(os.environ.get("VLLM_ALLOW_RUNTIME_LORA_UPDATING", "0").strip().lower() in
("1", "true")),

# By default, vLLM will check the peer-to-peer capability itself,
# in case of broken drivers. See https://github.com/vllm-project/vllm/blob/a9b15c606fea67a072416ea0ea115261a2756058/vllm/distributed/device_communicators/custom_all_reduce_utils.py#L101-L108 for details. # noqa
# If this env var is set to 1, vLLM will skip the peer-to-peer check,
# and trust the driver's peer-to-peer capability report.
"VLLM_SKIP_P2P_CHECK":
lambda: os.getenv("VLLM_SKIP_P2P_CHECK", "0") == "1",
}

# end-env-vars-definition
Expand Down
Loading