-
Notifications
You must be signed in to change notification settings - Fork 101
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
Add classifiers for CUDA 11 / 12 / 12.0 #124
Conversation
LGTM. Thanks Leo! 🙏 |
cc @di |
I feel like having both Would you be opposed to:
instead? Or is there another way to refer to everything compatible with a major version? |
Thanks, Dustin. I don't have strong opinion about Also, we're not changing any existing But in any case I agree that we are creating a new convention (for major-ver only classifiers). As I said I don't have strong opinion 🙂 Perhaps @jakirkham or @m3vaz could chime in? |
Would this be more or less confusing in your opinion?
I don't feel too strongly about this, but whatever we decide will be set in stone, more or less. |
That's interesting, I am thinking
Indeed. I wouldn't mind if you want to keep this PR open for a bit longer, Dustin. I can also ask NV folks internally and see if any of them has strong opinion on the 3 proposals (this, this, this) on the table. Thanks! |
Hi @di we discussed internally briefly about this, with CUDA product management involved. The 2nd proposal is slightly preferred (though none of us has strong opinion), so I just updated the PR. PTAL 🙂 thx! |
Friendly nudge @di 🙂 |
Another friendly nudge @di 🙂 |
Thanks Leo and Dustin! 🙏 |
Request to add a new Trove classifier.
The name of the classifier(s) you would like to add:
Environment :: GPU :: NVIDIA CUDA :: 11
Environment :: GPU :: NVIDIA CUDA :: 12
UPDATE:Environment :: GPU :: NVIDIA CUDA :: 12.0
Environment :: GPU :: NVIDIA CUDA 12 :: 12.0
(see Add classifiers for CUDA 11 / 12 / 12.0 #124 (comment))Why do you want to add this classifier?
CUDA 12.0 just came out today, so it'd be nice to add it now.
As for the major-version classifiers (CUDA 11/12), I consider them useful because starting with CUDA 11 there is guaranteed minor version compatibility at the programming model level (with caveats that users/developers would understand), so wheels can tag
CUDA :: 11
orCUDA :: 12
to declare such compatibility. For example, NVIDIA and some GPU frameworks have been uploading wheels with suffix-cuXX
(XX=11 currently) on PyPI.org.cc: @jakirkham @m3vaz