[PATCH] D119427: [SVE][CodeGen] Add support for scalable vectors in AArch64TargetLowering::ReconstructShuffle

[David Sherwood via Phabricator via llvm-commits]

david-arm created this revision.
david-arm added reviewers: sdesmalen, peterwaller-arm, CarolineConcatto, dmgreen.
Herald added subscribers: ctetreau, psnobl, hiraditya, kristof.beyls, tschuett.
Herald added a reviewer: efriedma.
david-arm requested review of this revision.
Herald added subscribers: llvm-commits, alextsao1999.
Herald added a project: LLVM.

At the moment if AArch64TargetLowering::ReconstructShuffle encounters
a scalable vector it just bails out of the optimisation entirely.
However, there are cases when we can make the transformation succeed
and produce far better code than if we just immediately bail out.

You can see the codegen improvement in the following tests:

  CodeGen/AArch64/sve-fixed-length-reshuffle.ll

I have tried to highlight in sve-fixed-length-reshuffle.ll those tests
that exercise the AArch64TargetLowering::ReconstructShuffle code path
when the transformation succeeds or fails.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D119427

Files:
  llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  llvm/test/CodeGen/AArch64/sve-fixed-length-reshuffle.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D119427.407470.patch
Type: text/x-patch
Size: 11392 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220210/095d9937/attachment.bin>