[WIP] Include GPU in wheels #479
Draft
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Bazel page is returning 404s, need to wait for it to come back and retry. |
95-martin-orion
force-pushed
the
qsim-gha-gpu
branch
from
96f86eb
to
74942aa
Compare
95-martin-orion
force-pushed
the
qsim-gha-gpu
branch
from
2c6d87d
to
29d4993
Compare
NOUIY
approved these changes
Dec 29, 2021
|
An inevitable side effect of building GPU support on machines that lack GPU support: we have no way to test GPU behavior in continuous integration. However, since the wheels generated by GHA are accessible from the test artifacts, it will be possible for us to spin up a GPU-enabled machine, copy over the wheels, and test before release. If we move forward with this, the process outlined above will become a mandatory part of all GPU changes. |
|
@95-martin-orion @sergeisakov FYI, we prepared a demo to show how to build a project upon CUDA/cuQuantum wheels: https://github.com/NVIDIA/cuQuantum/tree/main/extra/demo_build_with_wheels |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #430.
Since cross-platform support requires building on the manylinux image, installing NVCC on the runner is the best we can do unless we decide to push a manylinux + CUDA image to Docker Hub.