diff options
author | Matthias Braun <matze@braunis.de> | 2024-10-29 17:16:17 -0700 |
---|---|---|
committer | GitHub <noreply@github.com> | 2024-10-29 17:16:17 -0700 |
commit | 255e441613e39a391e9f85d6a605cc9e46dcf273 (patch) | |
tree | b9d7881e4ad4c62ff967fd425730e731335c0dcc /llvm/lib/CodeGen/MachineModuleSlotTracker.cpp | |
parent | 0fa2fb3ed0bc726e5dcf8258bf764aacd1c2e6dc (diff) | |
download | llvm-255e441613e39a391e9f85d6a605cc9e46dcf273.zip llvm-255e441613e39a391e9f85d6a605cc9e46dcf273.tar.gz llvm-255e441613e39a391e9f85d6a605cc9e46dcf273.tar.bz2 |
X86: Do not return invalid cost for fp16 conversion (#114128)
Returning invalid instruction costs when converting from/to fp16 in
`X86TTIImpl::getCastInstrCost` when there is no hardware support
available was triggering asserts. This changes the code to return a
large (arbitrary) number to model the fact that libcalls are used to
implement the conversion.
This also simplifies the code by only reporting costs for the scalar
fp16 conversion; vectorized costs being left to the fallback assuming
scalarization.
This is a follow-up to assertion issues reported for the changes in
#113195
Diffstat (limited to 'llvm/lib/CodeGen/MachineModuleSlotTracker.cpp')
0 files changed, 0 insertions, 0 deletions