[PATCH] D156398: [AMDGPU] Remove post-PromoteAlloca SROA run

Thu Jul 27 11:22:17 PDT 2023

arsenm accepted this revision.
arsenm added inline comments.
This revision is now accepted and ready to land.

================
Comment at: llvm/test/CodeGen/AMDGPU/resource-optimization-remarks.ll:182
 ; STDERR-NEXT: remark: foo.cl:74:0:     AGPRs: 10
-; STDERR-NEXT: remark: foo.cl:74:0:     ScratchSize [bytes/lane]: 64
+; STDERR-NEXT: remark: foo.cl:74:0:     ScratchSize [bytes/lane]: 144
 ; STDERR-NEXT: remark: foo.cl:74:0:     Dynamic Stack: True
----------------
Joe_Nash wrote:
> This is a question for my curiosity. Presumably if the 80 bytes/lane for %alloca is now on the stack, shouldn't we expect some other value like VGPRs to go down by 80 bytes (-20 VGPRs)?
Mechanically that's not really how it works. In this case the stack isn't actually used for anything other than filler content (it's kind of a bug this was optimized out to begin with, this memset probably should have been volatile)

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D156398/new/

https://reviews.llvm.org/D156398