Changed code to disable k_truss
on CUDA 11.4 differently
#1811
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Outdated suggestions cannot be applied.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.
Changed code to disable k_truss on CUDA 11.4 to not use
numba.cuda.runtime.get_version()
at import time since this creates a CUDA context which breaks daskLocalCUDACluster
init (causes a nccl init invaid usage exception).This code change results in almost the same UX as before, except explicit imports of the k_truss module no longer raise the exception immediately, and instead raise only when a k_truss algo is called.
Tested by ensuring unit tests for the conditional enabling of k_truss still work and also checked that a
LocalCUDACluster
could properly be created.Also changed
dask_client
fixture tomodule
sincesession
was not compatible with tests that tried to create aLocalCUDACluster
independently within the same session (rmat tests).cc @cjnolet @pentschev