NVIDIA / NVIDIA/cccl

Fix `BlockRadixRankMatchEarlyCounts` or constrain it

Open
#901 2 comments 0 reactions 1 assignee Claimed by @wmaxey View on GitHub
cub
Dominant language
C++
Stars
2.5k
Forks
487
Avg merge
2d 7h
Merged PRs (30d)
296

Description

Currently, `BlockRadixRankMatchEarlyCounts` doesn't work in some specific cases `(1 << RADIX_BITS) % WARP_THREADS != 0`. This use case should be addressed or the structure has to be complemented with a static assert that validates template arguments. To make the structure conforming, we might as well provide `BlockDimY` and `BlockDimZ` template parameters.

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.