Cold-Air-Networks/mirror-llvm-project
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 7

[flang][ReleaseNotes] Add do concurrent parallelization note

Browse source
This commit is contained in:
ergawy 2026-02-23 03:11:37 -06:00 committed by Cullen Rhodes
parent 1a802d58de
commit 9356fe4d1f
1 changed files with 12 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

12
flang/docs/ReleaseNotes.md
View file

@ -33,6 +33,18 @@ page](https://llvm.org/releases/).
* Experimental support for multi-image program launch, collective subroutines,
synchronization, teams, and image enumeration.
### `do concurrent` mapping to OpenMP
Flang now has better support for `do concurrent` parallelization on the CPU and
GPU. On the CPU, we validated the feature using
[FIATS'](https://github.com/BerkeleyLab/fiats) inference and training codes. On
the GPU, we have basic support that is still in progress; we have offload tests
for 1D and 2D saxpy. We also validated using codes that do not use extensive
user-defined types and/or allocatables.
To use the feature, use `-fdo-concurrent-to-openmp=[none|host|device]`. OpenMP
must be enabled as well, for example: `-fopenmp [--offload-arch=<target_arch>]`.
## Bug Fixes
## Non-comprehensive list of changes in this release