-
-
Notifications
You must be signed in to change notification settings - Fork 7
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
Rebuild for CUDA 12 #39
Rebuild for CUDA 12 #39
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 ( |
Seeing the following error on CI:
As the new CUDA 12 packages split all these components up and standard conda-forge Docker images are used ( conda-forge/staged-recipes#21382 ), think we need to add host:
- cuda-nvtx # [(cuda_compiler_version or "").startswith("12")] Edit: For more info about the CUDA 12 bringup, please see issue ( conda-forge/conda-forge.github.io#1963 ) |
Trying to push the fixes to the bot branch is giving me an access denied error. Are there any other extra permissions that I need in order to do it? From @h-vetinari's comments on https://matrix.to/#/!SOyumkgPRWoXfQYIFH:matrix.org/$168168775831YVYfK:gitter.im?via=matrix.org&via=gitter.im&via=cadair.com, it seems like having accepted the team invites should let me push to this branch, but is there something else I'm missing? |
@conda-forge-admin, please update team |
Actually just tried that in an issue ( #40 ), which worked Edit: Maybe give it another try? |
Still getting a permission denied error from local machine, tried editing it via codespaces and it gave the following warning:
|
Could you please share how you are trying to push? Also do you use SSH with git / GitHub? |
Yes, it's using SSH with git, both through GitKraken with my github account and manually adding the remote in command line. I tried using HTTPS, but it gave an error "Support for password authentication was removed on August 13, 2021". |
Ok, could you please share the command lines run? |
So even:
gives
|
This is almost certainly a misconfiguration issue on your end. Fetching from a remote has nothing to do with your maintainer rights, it should always be possible. |
@jakirkham Ok got the git issues sorted, but it looks like it's still having the same issue even with cuda-nvtx in host. Do you know if there's any interaction with cmake's find_package for |
Hmm, I added:
to host as well, following conda-forge/conda-forge.github.io#1963, but it's still throwing the same error. |
Think we need to add this to the -DNvToolExt_INCLUDE_DIR:PATH="${PREFIX}/include" |
Ok I added
But now it's giving the error of:
It looks like these issues are fixed via something like |
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.
Sorry looks like it needs to be NvToolExt_ROOT_DIR
Hmm, using just
|
Ok so it looks like the include files from nvtx-c are installed to |
Think we want to only use We may need to tinker with the search path for CUDA 12. Will try to include some suggestions for this a bit later |
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.
Had tried setting the path to the NVTX targets directory (as explained below). However that didn't seem to work. Providing this context in case you know of a better way to try configuring Kaldi with this information
@jakirkham I can play around a bit more with it, but I guess I'm still confused why cuda-nvtx is installing files outside of the normal |
This is better explained in issue ( conda-forge/conda-forge.github.io#1963 ). Please read over that and ask any general questions on that info in that issue |
Hi! This is the friendly automated conda-forge-linting service. I was trying to look for recipes to lint for you, but it appears we have a merge conflict. Please ping the 'conda-forge/core' team (using the @ notation in a comment) if you believe this is a bug. |
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 ( |
@jakirkham ok I think I have everything building correctly, had to add a few more dependencies and ensure that the lib/stubs folder was on the search path for cmake. Can you take a look over for sanity? |
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 pushing through on this @mmcauliffe! Just one nit about the build-script. I'd also like you to squash everything into two commits - your changes in one, and the rerender in another. If you want I can help with that.
It might be slightly more than two after all. But less than 29. ;-) |
3c50d7c
to
8d11b04
Compare
Hi! This is the friendly automated conda-forge-linting service. I was trying to look for recipes to lint for you, but it appears we have a merge conflict. Please ping the 'conda-forge/core' team (using the @ notation in a comment) if you believe this is a bug. |
8d11b04
to
9f7fe4d
Compare
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 ( |
Ok, I think I squashed them right, but let me know if I should have done it a different way! I also rebased onto the latest main commit, and I think everything should pass still. |
The transition to CUDA 12 SDK includes new packages for all CUDA libraries and build tools. Notably, the cudatoolkit package no longer exists, and packages should depend directly on the specific CUDA libraries (libblas, libcusolver, etc) as needed. For an in-depth overview of the changes and to report problems [see this issue]( conda-forge/conda-forge.github.io#1963 ). Please feel free to raise any issues encountered there. Thank you! 🙏
…nda-forge-pinning 2023.06.06.09.57.57
9f7fe4d
to
608c6d1
Compare
It was pretty good already, thanks! The only thing I though would be cleaner to leave in the commits of the migrator. I reinstated those1, and I also separated out the commit aimed at #41, and rerendered again at the end. Otherwise I didn't change anything. Footnotes
|
608c6d1
to
4c0322c
Compare
…nda-forge-pinning 2023.07.28.20.44.13
4c0322c
to
9e42107
Compare
Sorry for the repeated force-pushes, something weird was happening with the rerender, and it seems the new conda build that just got released does a different (=better) job. The last rerender now does a much more focused job too. :) |
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.
This now LGTM! Thanks a lot for your persistence on this! :)
Generally this looks reasonable. Thank you both for working on it! 🙏 |
This PR has been triggered in an effort to update cuda120.
Notes and instructions for merging this PR:
Please note that if you close this PR we presume that the feedstock has been rebuilt, so if you are going to perform the rebuild yourself don't close this PR until the your rebuild has been merged.
Here are some more details about this specific migrator:
If this PR was opened in error or needs to be updated please add the
bot-rerun
label to this PR. The bot will close this PR and schedule another one. If you do not have permissions to add this label, you can use the phrase@conda-forge-admin, please rerun bot
in a PR comment to have theconda-forge-admin
add it for you.This PR was created by the regro-cf-autotick-bot. The regro-cf-autotick-bot is a service to automatically track the dependency graph, migrate packages, and propose package version updates for conda-forge. Feel free to drop us a line if there are any issues! This PR was generated by https://github.com/regro/cf-scripts/actions/runs/5192514335, please use this URL for debugging.