ara
ara copied to clipboard
[hardware] Floating-Point Reductions
This PR depends on https://github.com/pulp-platform/ara/pull/133
This PR introduces FP-reduction support to Ara.
Thanks to @xiaorui-yin for the great work and effort!
Changelog
Added
- Support for vector single-width floating-point reduction instructions:
vfredusum
,vfredosum
,vfredmin
,vfredmax
- Support for Vector widening floating-point reductions:
vfwredusum
,vfwredosum
-
fdotproduct
benchmark to evaluate dot products with reductions -
fredsum
benchmark to evaluate fp-reductions -
riscv-tests
forvfredusum
,vfredosum
,vfredmin
,vfredmax
,vfwredusum
,vfwredosum
Changed
- Updated target
-march
torv64gcv_zfh_zvfh0p1
to enable half-floats support
Checklist
- [x] Automated tests pass
- [x] Changelog updated
- [x] Code style guideline is observed
- [x] No frequency degradation