[PATCH] D88291: [AMDGPU] Insert waterfall loops for divergent calls

[Sebastian Neubauer via Phabricator via llvm-commits]

Flakebi created this revision.
Flakebi added reviewers: arsenm, madhur13490, cfang.
Herald added subscribers: llvm-commits, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, jvesely, kzhuravl.
Herald added a project: LLVM.
Flakebi requested review of this revision.
Herald added a subscriber: wdng.

Extend loadSRsrcFromVGPR to allow moving a range of instructions into
the loop. The call instruction is preceded by copies into physical
registers which should be part of the waterfall loop, as the registers
can be overwritten by the call.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D88291

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/test/CodeGen/AMDGPU/indirect-call.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D88291.294269.patch
Type: text/x-patch
Size: 13156 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200925/9d8b131f/attachment.bin>