JACC v0.0.2
Major Changes
- Added CI for NVIDIA RTX-A4000 and AMD M100 GPUs on ORNL systems, e.g. ExCL
- Added single and threaded CI for CPUs on Linux and macOS runners with macro
maybe_threaded
- Initial work on oneAPI back end
What's Changed
- Added oneAPI implementation and test by @pedrovalerolara in #14
- Fixed bugs, optimized implementation, added test codes in test/test-p… by @pedrovalerolara in #20
- New optimization for parallel reduce on CUDA, AMDGPU and oneAPI using… by @pedrovalerolara in #22
- Parallel reduce MN optimized for CUDA, AMDGPU, and oneAPI using multi… by @pedrovalerolara in #24
- Add AMD GPU CI on ExCL cousteau by @williamfgc in #25
- Add commands to verify CI environment. by @Geekdude in #28
- CI: Switched back to guibranco/github-status-action-v2 since the proxy fix was merged upstream. by @Geekdude in #38
- Revert "CI: Switched back to guibranco/github-status-action-v2 since … by @Geekdude in #39
- CI: Update github-status-action to v1.1.10 by @Geekdude in #41
- Bugs fixed in AMDGPU back end, indexes, group size in second kernel i… by @pedrovalerolara in #42
- Added performance tests for AMDGPU by @pedrovalerolara in #43
- Refactor CI and address AMD sync issue by @williamfgc in #47
- Upgrade GA CPU CI by @williamfgc in #48
- CI: Update Instantiate to load rocm, just like Test. by @Geekdude in #54
- Fix precompilation by @michel2323 in #52
- Introduce maybe_threaded by @williamfgc in #60
New Contributors
Full Changelog: v0.0.1...v0.0.2