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.

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

Update libcudf compiler requirements in contributing doc #16103

Merged
merged 5 commits into from
Jun 28, 2024
Merged
Changes from 1 commit
Commits
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
8 changes: 4 additions & 4 deletions CONTRIBUTING.md
Original file line number Diff line number Diff line change
Expand Up @@ -71,13 +71,13 @@ for a minimal build of libcudf without using conda are also listed below.

Compilers:

* `gcc` version 9.3+
* `nvcc` version 11.5+
* `cmake` version 3.26.4+
* `gcc` version 11.4+
* `nvcc` version 11.8+
* `cmake` version 3.29.6+

CUDA/GPU:

* CUDA 11.5+
* CUDA 11.8+
Copy link
Contributor

Choose a reason for hiding this comment

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

We still support older CUDA versions at runtime, we test 11.4 in our nightlies. We only require 11.8 for builds.

Suggested change
* CUDA 11.8+
* CUDA 11.4+

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks, I realized that this was probably runtime after I pushed this up. Glad to get confirmation on that.
Should I change this to say CUDA/GPU Runtime perhaps? I think the driver version is still correct.

Copy link
Contributor

@bdice bdice Jun 26, 2024

Choose a reason for hiding this comment

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

Yes, you can clarify this is runtime information. Driver version is still correct as far as I know, but we have discussed removing driver information because it is implied by the compatible CUDA versions — and it would be one less piece of information to track and update. Remove that driver information (and maybe add a link to a page like https://docs.nvidia.com/deploy/cuda-compatibility/) if you think that is reasonable.

* NVIDIA driver 450.80.02+
* Volta architecture or better (Compute Capability >=7.0)

Expand Down
Loading