[PATCH] D72796: AMDGPU/GlobalISel: Legalize unpacked d16 image operations

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: nhaehnle, kerbowa.
Herald added subscribers: Petar.Avramovic, jfb, hiraditya, t-tye, tpr, dstuttard, rovka, yaxunl, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

On targets that don't have the normal packed f16 layout, handle these
during legalization. Directly modify the register types. We can infer
this was a d16 load based on the mem operand size during selection.

A16 operands should possibly be handled here as well, but don't worry
about that yet.


https://reviews.llvm.org/D72796

Files:
  llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp
  llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.h
  llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-llvm.amdgcn.image.load.2d.d16.ll
  llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-llvm.amdgcn.image.store.2d.d16.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D72796.238326.patch
Type: text/x-patch
Size: 33891 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200115/cbc94745/attachment.bin>