[PATCH] D96978: [WIP] Bias statepoint operands towards spilling

[Philip Reames via Phabricator via llvm-commits]

reames created this revision.
Herald added subscribers: dantrushin, pengfei, hiraditya, mcrosier, qcolombet.
Herald added a reviewer: bollu.
reames requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

Not for review, simply saving an idea for later.

The idea behind this is the a statepoint operand which can be spilled should contribute to the spill cost of a live interval as it's equally happy to use register and stack.

This made a difference on a larger test I'd looked at, but I haven't managed to reduce a small test case.  All of the small test cases I write have obviously bigger problems (quality wise).

Setting this aside until other issues resolved.


https://reviews.llvm.org/D96978

Files:
  llvm/lib/CodeGen/CalcSpillWeights.cpp
  llvm/test/CodeGen/X86/statepoint-allocas.ll
  llvm/test/CodeGen/X86/statepoint-call-lowering.ll
  llvm/test/CodeGen/X86/statepoint-duplicates-export.ll
  llvm/test/CodeGen/X86/statepoint-gc-live.ll
  llvm/test/CodeGen/X86/statepoint-invoke.ll
  llvm/test/CodeGen/X86/statepoint-live-in-remat.ll
  llvm/test/CodeGen/X86/statepoint-live-in.ll
  llvm/test/CodeGen/X86/statepoint-no-realign-stack.ll
  llvm/test/CodeGen/X86/statepoint-regs.ll
  llvm/test/CodeGen/X86/statepoint-uniqueing.ll
  llvm/test/CodeGen/X86/statepoint-vector.ll
  llvm/test/CodeGen/X86/statepoint-vreg.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D96978.324700.patch
Type: text/x-patch
Size: 31488 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210218/93cb5ef1/attachment.bin>