diff options
| author | Jean-Didier PAILLEUX <jean-didier.pailleux@sipearl.com> | 2025-04-01 08:07:26 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-04-01 08:07:26 +0200 |
| commit | bae3577002b6bda92837723a06a4ca5c498d300f (patch) | |
| tree | 79b42a1b01f0458a3b6c4c14cdcee584e8b9b897 /llvm/lib | |
| parent | 091dcb8fc2b6ccb88c2975076e94f3cb6530db46 (diff) | |
| download | llvm-bae3577002b6bda92837723a06a4ca5c498d300f.zip llvm-bae3577002b6bda92837723a06a4ca5c498d300f.tar.gz llvm-bae3577002b6bda92837723a06a4ca5c498d300f.tar.bz2 | |
[flang] Define ERF, ERFC and ERFC_SCALED intrinsics with Q and D prefix (#125217)
`ERF`, `ERFC` and `ERFC_SCALED` intrinsics prefixed by `Q` and `D` are
missing. Codes such as `CP2K`(https://github.com/cp2k/cp2k) and
`TurboRVB`(https://github.com/sissaschool/turborvb) use these intrinsics
just like defined in the GNU standard and here:
https://www.ibm.com/docs/fr/xl-fortran-aix/16.1.0?topic=reference-intrinsic-procedures
These intrinsics are based on the existing intrinsics but apply a
restriction on the type kind.
- `DERF`, `DERFC` and `DERFC_SCALED` are for double précision only.
- `QERF`, `QERFC` and `QERFC_SCALED` are for quad précision only.
Diffstat (limited to 'llvm/lib')
0 files changed, 0 insertions, 0 deletions