Extend HPX benchmark performance investigations to latest performant CPUs
Open
- Dominant language
- C++
- Stars
- 2.5k
- Forks
- 486
- Avg merge
- 2d 6h
- Merged PRs (30d)
- 295
Description
The PR https://github.com/NVIDIA/cccl/pull/7680 introduces HPX to be used as CPU backed for parallel algorithms in CCCL. The initial benchmark's on the description shows promising results, but we would like to further extend and see how the the performance is on the newer NVIDIA Grace CPUs (ARM) or multi-socket EPYC CPU systems.
Contributor guide
Assessment
This issue has not been assessed yet.