[llvm] [AMDGPU][True16][CodeGen] select vgpr16 for asm inline 16bit vreg (PR #140946)
Jay Foad via llvm-commits
llvm-commits at lists.llvm.org
Thu May 22 01:39:04 PDT 2025
================
@@ -16062,7 +16062,8 @@ SITargetLowering::getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI_,
case 'v':
switch (BitWidth) {
----------------
jayfoad wrote:
`switch` seems overkill here. Could just handle it with:
```
if (BitWidth == 16 && !Subtarget->useRealTrue16Insts())
BitWidth = 32;
```
https://github.com/llvm/llvm-project/pull/140946
More information about the llvm-commits
mailing list