Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

[v1][torch.compile] support managing cudagraph buffer #10203

Merged
merged 6 commits into from
Nov 11, 2024

Conversation

youkaichao
Copy link
Member

advantage: caller does not need to explicitly manage the input buffer

disadvantage: the cost of one more input tensor, and one more copy kernel.

given the code simplification from user side, i think it should be worthwhile.

Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Copy link

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

  • Add ready label to the PR
  • Enable auto-merge.

🚀

@@ -94,13 +94,14 @@ def test_simple_piecewise_compile():
with set_compile_context([1, 2]):
model(input_buffer)
Copy link
Collaborator

Choose a reason for hiding this comment

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

To double check: Here, input_buffer is not directly cached and reused by the backend, right?

Copy link
Member Author

Choose a reason for hiding this comment

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

right. let me rename the variable. it is not a persistent buffer anymore. it can be a newly created tensor.

youkaichao and others added 2 commits November 10, 2024 17:56
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: youkaichao <youkaichao@gmail.com>
Signed-off-by: youkaichao <youkaichao@gmail.com>
@youkaichao youkaichao changed the title [v1][torch.compile] manage cudagraph buffer in compiler [v1][torch.compile] support managing cudagraph buffer Nov 11, 2024
@youkaichao youkaichao merged commit 330e82d into vllm-project:main Nov 11, 2024
19 of 23 checks passed
@youkaichao youkaichao deleted the input_buffer branch November 11, 2024 19:10
rickyyx pushed a commit to rickyyx/vllm that referenced this pull request Nov 13, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
sumitd2 pushed a commit to sumitd2/vllm that referenced this pull request Nov 14, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Sumit Dubey <sumit.dubey2@ibm.com>
KuntaiDu pushed a commit to KuntaiDu/vllm that referenced this pull request Nov 20, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
mfournioux pushed a commit to mfournioux/vllm that referenced this pull request Nov 20, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Maxime Fournioux <55544262+mfournioux@users.noreply.github.com>
tlrmchlsmth pushed a commit to neuralmagic/vllm that referenced this pull request Nov 23, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
sleepwalker2017 pushed a commit to sleepwalker2017/vllm that referenced this pull request Dec 13, 2024
…0203)

Signed-off-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants