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

Jump to bottom

feat: update torch and modify cuda graphs #220

Merged
merged 7 commits into from
Dec 19, 2022
Merged

feat: update torch and modify cuda graphs #220

merged 7 commits into from
Dec 19, 2022

Conversation

pommedeterresautee
Copy link
Member

@pommedeterresautee pommedeterresautee commented Dec 14, 2022 •
edited
Loading

fix #197

  • update PyTorch nightly to last version
  • rely on CUDA graph PyTorch wrapper (instead of our own)

@pommedeterresautee pommedeterresautee marked this pull request as draft December 14, 2022 21:19
@github-actions github-actions bot added feature and removed feature labels Dec 14, 2022
gaetansnl
gaetansnl reviewed Dec 15, 2022
View reviewed changes
requirements.txt
@@ -1,5 +1,5 @@
triton==2.0.0.dev20221202
torch==1.14.0.dev20221029+cu117
torch== 2.0.0.dev20221214+cu117
Copy link
Contributor

Choose a reason for hiding this comment

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

note to remember: version in docker needs to be changed

@github-actions github-actions bot added feature and removed feature labels Dec 15, 2022
@pommedeterresautee pommedeterresautee self-assigned this Dec 15, 2022
@pommedeterresautee pommedeterresautee added the dependencies Pull requests that update a dependency file label Dec 15, 2022
@pommedeterresautee pommedeterresautee marked this pull request as ready for review December 15, 2022 12:38
@pommedeterresautee
Copy link
Member Author

simple test pass

pytest test/test_torchdynamo.py

============================================================================================== 268 passed in 5296.85s (1:28:16) ==============================================================================================

@pommedeterresautee
Copy link
Member Author

all tests pass

====================== 2849 passed in 9063.33s (2:31:03) =======================

gaetansnl
gaetansnl requested changes Dec 16, 2022
View reviewed changes
Copy link
Contributor

@gaetansnl gaetansnl left a comment

Choose a reason for hiding this comment

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

needs docker change

src/kernl/model_optimization.py
@@ -22,15 +22,11 @@
from kernl.optimizer.dynamo_backend import dynamo_backend_ofi


# single shared pool by default
_pool: (int, int) = torch.cuda.graph_pool_handle()
Copy link
Contributor

Choose a reason for hiding this comment

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

should we open an issue to add pool back ?

Copy link
Member Author

Choose a reason for hiding this comment

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

it will be covered here:

@pommedeterresautee
Copy link
Member Author

@gaetansnl docker updated

@pommedeterresautee pommedeterresautee merged commit e604c11 into main Dec 19, 2022
@pommedeterresautee pommedeterresautee deleted the feat/update_pytorch branch December 19, 2022 19:24
@pommedeterresautee pommedeterresautee mentioned this pull request Dec 21, 2022
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@gaetansnl gaetansnl gaetansnl approved these changes

Assignees

@pommedeterresautee pommedeterresautee

Labels
dependencies Pull requests that update a dependency file feature
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Comparison with pytorch 2.0
2 participants
@pommedeterresautee @gaetansnl