Skip to content

hotfix: revert torch.library register #709

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
Dec 30, 2024
Merged

hotfix: revert torch.library register #709

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fb1c1a6
upd
yzh119 Dec 30, 2024
3a5b9a0
upd
yzh119 Dec 30, 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
21 changes: 13 additions & 8 deletions flashinfer/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -236,19 +236,24 @@ def register_custom_op(
device_types: Optional[Union[str, Sequence[str]]] = None,
schema: Optional[str] = None,
) -> Callable:
return torch.library.custom_op(
name,
fn,
mutates_args=mutates_args,
device_types=device_types,
schema=schema,
)
# NOTE(Zihao): torch.library.custom_op has significant overhead as mentioned in the following link
# https://github.com/vllm-project/vllm/blob/36e76700453924c8d421db99af70a88a1df835cd/vllm/utils.py#L1660-L1674

# return torch.library.custom_op(
# name,
# fn,
# mutates_args=mutates_args,
# device_types=device_types,
# schema=schema,
# )
return lambda x: x

def register_fake_op(
name: str,
fn: Optional[Callable] = None,
) -> Callable:
return torch.library.register_fake(name, fn)
# return torch.library.register_fake(name, fn)
return lambda x: x


def get_cuda_stream(device: torch.device) -> int:
Expand Down