The `cub.bench.segmented_radix_sort.keys.base` benchmark covers too many combinations
Open
- Dominant language
- C++
- Stars
- 2.5k
- Forks
- 486
- Avg merge
- 2d 6h
- Merged PRs (30d)
- 295
Description
When running the ordinary `cub.bench.segmented_radix_sort.keys.base` benchmark, 798 variations are covered. #6731 noted, that this takes 136min of time, which is more than inconvenient for benchmarking a code change.
We should reduce the number of benchmarked combinations to get the benchmark runtime down to a few minutes.
Contributor guide
Assessment
This issue has not been assessed yet.