[PATCH] D78255: [AMDGPU] copyPhysReg() for 16 bit SGPR subregs

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec updated this revision to Diff 258132.
rampitec added a comment.

Improved handling for GFX8. In a nutshell GFX8 cannot use hi16 in any situation because SDWA does not support SGPR sources on GFX8. At least there hi16 shall not be used, but it does not have packed f16 for legitimate use of it anyway.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D78255/new/

https://reviews.llvm.org/D78255

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/test/CodeGen/AMDGPU/lo16-hi16-physreg-copy.mir
  llvm/test/CodeGen/AMDGPU/lo16-lo16-physreg-copy-sgpr.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D78255.258132.patch
Type: text/x-patch
Size: 4295 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200416/fcd72f64/attachment-0001.bin>