-
-
Notifications
You must be signed in to change notification settings - Fork 16
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
Switch to cross compiling for aarch64/ppc64le #88
Switch to cross compiling for aarch64/ppc64le #88
Conversation
Hi! This is the friendly automated conda-forge-linting service. I just wanted to let you know that I linted all conda-recipes in your PR ( |
…nda-forge-pinning 2023.04.15.08.59.11
try to use nccl's default gencode for all platforms
@conda-forge-admin, please rerender |
…nda-forge-pinning 2023.04.15.08.59.11
Seems to be compiling fine. Would be nice if someone can help test the uploaded artifacts 🙂 @conda-forge/nccl @pentschev for vis |
Note: This would drop CUDA 11.0/11.1 for aarch64 and 10.2/11.0/11.1 for ppc64le. But the download count for these versions are very low anyway, so it shouldn't be a big deal. |
@conda-forge-admin, please rerender |
…nda-forge-pinning 2023.04.17.13.38.21
Chuck from the cuGraph team tested the artifact on an aarch64 machine with 4 GPUs. His tests passed. I think this is good to go. However, given that we're still doing CUDA 12 bring-up in parallel (#85), I'd like to discuss with @jakirkham @pentschev first how to proceed and merge, next week. |
@conda-forge-admin, please rerender |
…nda-forge-pinning 2023.04.21.18.24.51
recipe/conda_build_config.yaml
Outdated
# The added keys below are needed to enable ppc64le/aarch64 CUDA builds; the linux64 | ||
# and win64 builds are already handled in the conda-forge-pinning feedstock. Once we | ||
# migrate to enable CUDA builds on all supported platforms, these can be removed too. | ||
# conda-forge only supports cross-compiling with CUDA 11.2 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Wonder if we can drop this file entirely after making this change
IIUC this file was used for old CUDA versions on aarch64
and ppc64le
, which are no longer relevant after this change
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yup. Even for CUDA 11 it's irrelevant after conda-forge/conda-forge-pinning-feedstock#3624
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok dropped in the commit below and re-rendering
Let's see how this goes
Appears this is now redundant. So try dropping this file.
@conda-forge-admin, please re-render |
…nda-forge-pinning 2023.05.31.19.00.54
@conda-forge-admin, please re-render |
Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do. This message was generated by GitHub actions workflow run https://github.com/conda-forge/nccl-feedstock/actions/runs/5136819837. |
@conda-forge-admin, please re-render |
…nda-forge-pinning 2023.05.31.19.00.54
@leofang, could you please review? |
Borrow simplifications from Peter's CUDA 12 PR to call `make` once.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for finishing it @jakirkham @h-vetinari! Let's merge to unblock other PRs.
Agreed. Thanks Leo! 🙏 |
Hi! This is the friendly automated conda-forge-webservice.
I've rerendered the recipe as instructed in #87.
Here's a checklist to do before merging.
Fixes #87