[PATCH] D157133: [AMDGPU] Extend CalculateByteProvider to capture vectors and signed

[Jeffrey Byrnes via Phabricator via llvm-commits]

jrbyrnes created this revision.
jrbyrnes added reviewers: arsenm, gandhi21299.
Herald added subscribers: foad, wenlei, kerbowa, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
jrbyrnes requested review of this revision.
Herald added subscribers: llvm-commits, wangpc, wdng.
Herald added a project: LLVM.

Trees often involve extract i8 / i16s, and ultimate sources may come from extract i32s. Capturing trees that involve e.g. SIGN_EXTEND has limited impact on combining into perms, but significantly helps for combining trees for arithmetic ops (e.g. v_dot4)


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D157133

Files:
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/test/CodeGen/AMDGPU/image-load-d16-tfe.ll
  llvm/test/CodeGen/AMDGPU/insert_vector_elt.v2i16.ll
  llvm/test/CodeGen/AMDGPU/llvm.amdgcn.image.sample.d16.dim.ll
  llvm/test/CodeGen/AMDGPU/permute_i8.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D157133.547324.patch
Type: text/x-patch
Size: 16031 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230804/32b3eafc/attachment.bin>