Improve Primitives a Little Bit #2001
Closed
Conversation
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
CUDA: Replace thrust::reduce with CUB DeviceReduce::Sum in primitives::sum
Summary
Inspired by this conversation here NVIDIA/cccl#520
This PR replaces the CUDA implementation of primitivesDevice::CUDA::sum from thrust::reduce to cub::DeviceReduce::Sum with cudaMallocAsync/cudaFreeAsync for temporary storage. No interfaces change. No other code paths are modified.
The underlying GPU reduction kernel is the same family of CUB kernels Thrust would eventually dispatch to, but this change removes framework overhead.
This PR intentionally changes only "sum." Other Thrust-based reductions can be migrated if desired, such as for "max," "max_element," "logsumexp" or what not to CUB.
Hope this helps! I can provide some benchmarks if you want.