NVIDIA / NVIDIA/cccl

Reduce duplication in reduce kernels if extracting common code to device function does not hurt performance

Open
#9,789 0 comments 0 reactions 1 assignee Claimed by @NaderAlAwar View on GitHub
Dominant language
C++
Stars
2.5k
Forks
486
Avg merge
2d 6h
Merged PRs (30d)
295

Description

#9722 and #9740 introduced new kernels to handle device resident problem sizes. Since they share code with the existing kernels, it would make sense to extract that code to a common helper, however that resulted in SASS differences to the existing kernels.

We should measure if these sass differences have a negative impact on performance. If the impact is negligible, we should refactor the code to avoid duplication.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.