[PATCH] D95795: [AMDGPU] Add new CostPerUse values for VGPRs

[Matt Arsenault via Phabricator via llvm-commits]

arsenm requested changes to this revision.
arsenm added inline comments.
This revision now requires changes to proceed.


================
Comment at: llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp:159
+
+  // For kernels with no calling convention.
+  if (CSRegs && *CSRegs == AMDGPU::NoRegister)
----------------
Comment is wrong, kernel *is* a calling convention


================
Comment at: llvm/lib/Target/AMDGPU/SIRegisterInfo.td:284
 foreach Index = 0...255 in {
-  defm VGPR#Index :
-    SIRegLoHi16 <"v"#Index, Index, 0, 1>,
-    DwarfRegNum<[!add(Index, 2560), !add(Index, 1536)]>;
+  let CostPerUse=[0, !if(!gt(Index, 31), !srl(Index, 2), 0)] in {
+    defm VGPR#Index :
----------------
Needs a comment explaining the cost. Also space around =


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95795/new/

https://reviews.llvm.org/D95795