diff options
| author | Thurston Dang <thurston@google.com> | 2025-02-04 14:41:18 -0800 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-02-04 14:41:18 -0800 |
| commit | c9fccbd90cadfad53d68a199cd11c7afc2d01378 (patch) | |
| tree | 4152c6f46eb54e762da54cfd5ea4f74f25dba092 /flang/lib/Frontend/CompilerInvocation.cpp | |
| parent | 806e35175bc66f002cc600ab23abe221c905ef98 (diff) | |
| download | llvm-c9fccbd90cadfad53d68a199cd11c7afc2d01378.zip llvm-c9fccbd90cadfad53d68a199cd11c7afc2d01378.tar.gz llvm-c9fccbd90cadfad53d68a199cd11c7afc2d01378.tar.bz2 | |
[msan][NFCI] Add tests for Arm NEON floating-point min/max (vector) (#125729)
Currently handled (suboptimally) by handleUnknownInstruction:
- llvm.aarch64.neon.fmaxv (Floating-point Maximum (vector))
- llvm.aarch64.neon.fminv
- llvm.aarch64.neon.fmaxnmv (Floating-point Maximum Number across
Vector)
- llvm.aarch64.neon.fminnmv
(not to be mistaken with llvm.aarch64.neon.f{max,min}, which are
correctly handled by `maybeHandleSimpleNomemIntrinsic`)
Forked from llvm/test/CodeGen/AArch64/arm64-fminv.ll
Diffstat (limited to 'flang/lib/Frontend/CompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions