Skip to content

Pin cmake version to workaround pytorch build issues #75045

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

Closed
Parent: Add extra debug info for the super common "index out of bounds" error.
wants to merge 1 commit into from
Closed

Pin cmake version to workaround pytorch build issues #75045

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 3 additions & 1 deletion .github/workflows/run_torchbench.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,9 @@ jobs:
# shellcheck disable=SC1091
. "${HOME}"/anaconda3/etc/profile.d/conda.sh
conda activate pr-ci
conda install -y numpy requests ninja pyyaml mkl mkl-include setuptools cmake cffi typing_extensions \
# pin cmake version to 3.22 since 3.23 breaks pytorch build
# see details at: https://github.com/pytorch/pytorch/issues/74985
conda install -y numpy requests ninja pyyaml mkl mkl-include setuptools cmake=3.22 cffi typing_extensions \
Copy link
Member

Choose a reason for hiding this comment

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

Maybe a good follow up to this would be to just pin all of these particular dependencies

Copy link
Contributor Author

Choose a reason for hiding this comment

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

One problem of doing that is we need to keep consistent with pytorch/benchmark workflows such as in https://github.com/pytorch/benchmark/blob/main/.github/workflows/v2-bisection.yml#L31. Perhaps it is better to create a script in pytorch/benchmark to manage these dependencies in a unified manner.

future six dataclasses pillow pytest tabulate gitpython git-lfs tqdm psutil
# install magma
conda install -y -c pytorch "${MAGMA_VERSION}"
Expand Down