Skip to content

CUDA: Deprecate support for CC < 5.3 and CTK < 10.2 - #7755

Merged
sklam merged 2 commits into
numba:masterfrom
gmarkall:deprecate-cc-52-ctk101
Jan 26, 2022
Merged

sklam merged 2 commits into
numba:masterfrom
gmarkall:deprecate-cc-52-ctk101

Conversation

@gmarkall

Copy link
Copy Markdown
Member

This deprecates support for:

  • Compute capabilities less than 5.3
  • CUDA toolkits less than 10.2

in preparation for removing support for them in Numba 0.56. This is in accordance with supporting 3 year's worth of toolkits (9.2 is almost 4 years old now), and all hardware released in the last 6 years is still supported.

I've also updated the Numba sysinfo example so it looks similar to what a user might see today (rather than referencing outdated versions of things) and added a small note to cuda.detect() to show supported but deprecated devices, and to correctly identify devices with CC < 3.0 as unsupported.

@gmarkall gmarkall added CUDA CUDA related issue/PR doc Effort - short Short size effort needed 3 - Ready for Review labels Jan 18, 2022
@gmarkall gmarkall added this to the Numba 0.55.1 milestone Jan 18, 2022
@gmarkall

Copy link
Copy Markdown
Member Author

gpuci run tests

@stuartarchibald stuartarchibald left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Thanks for the patch, looks good, just a couple of minor things to resolve. Thanks again!

Comment thread docs/source/cuda/overview.rst
Comment thread docs/source/cuda/overview.rst Outdated
Comment thread docs/source/user/installing.rst
Comment thread numba/cuda/api.py Outdated
@stuartarchibald stuartarchibald added 4 - Waiting on author Waiting for author to respond to review and removed 3 - Ready for Review labels Jan 21, 2022
Co-authored-by: stuartarchibald <stuartarchibald@users.noreply.github.com>
@gmarkall

Copy link
Copy Markdown
Member Author

gpuci run tests

@gmarkall gmarkall added 4 - Waiting on reviewer Waiting for reviewer to respond to author and removed 4 - Waiting on author Waiting for author to respond to review labels Jan 21, 2022
@stuartarchibald stuartarchibald mentioned this pull request Jan 21, 2022
4 tasks done

@stuartarchibald stuartarchibald left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Thanks for the patch and fixes.

@stuartarchibald stuartarchibald added 4 - Waiting on CI Review etc done, waiting for CI to finish and removed 4 - Waiting on reviewer Waiting for reviewer to respond to author labels Jan 21, 2022
@gmarkall

Copy link
Copy Markdown
Member Author

This failed gpuCI only because #7740 is not yet merged.

@stuartarchibald

Copy link
Copy Markdown
Contributor

Looks like CTK=11.4 with CUDA 11.2 is failing due to an unexpected success on gpuci:

11:35:09 test_profiling (numba.cuda.tests.cudadrv.test_profiler.TestProfiler) ... unexpected success

@stuartarchibald

Copy link
Copy Markdown
Contributor

Looks like CTK=11.4 with CUDA 11.2 is failing due to an unexpected success on gpuci:

11:35:09 test_profiling (numba.cuda.tests.cudadrv.test_profiler.TestProfiler) ... unexpected success

ah #7755 (comment) notes why, thanks @gmarkall .

@gmarkall

Copy link
Copy Markdown
Member Author

gpuci run tests

@stuartarchibald stuartarchibald added 5 - Ready to merge Review and testing done, is ready to merge and removed 4 - Waiting on CI Review etc done, waiting for CI to finish labels Jan 26, 2022
@sklam
sklam merged commit 8c0feea into numba:master Jan 26, 2022
esc pushed a commit to esc/numba that referenced this pull request Jan 27, 2022
CUDA: Deprecate support for CC < 5.3 and CTK < 10.2
esc pushed a commit to esc/numba that referenced this pull request Jan 27, 2022
CUDA: Deprecate support for CC < 5.3 and CTK < 10.2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

5 - Ready to merge Review and testing done, is ready to merge CUDA CUDA related issue/PR doc Effort - short Short size effort needed

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants