[all-commits] [llvm/llvm-project] 9d1bea: [SVE][LoopVectorize] Optimise code generated by wi...

[RosieSumpter via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 9d1bea9c88b36aaa450354c4876f025d2c4a14f7
      https://github.com/llvm/llvm-project/commit/9d1bea9c88b36aaa450354c4876f025d2c4a14f7
  Author: Rosie Sumpter <rosie.sumpter at arm.com>
  Date:   2021-09-10 (Fri, 10 Sep 2021)

  Changed paths:
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    M llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll
    M llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-phi.ll

  Log Message:
  -----------
  [SVE][LoopVectorize] Optimise code generated by widenPHIInstruction

For SVE, when scalarising the PHI instruction the whole vector part is
generated as opposed to creating instructions for each lane for fixed-
width vectors. However, in some cases the lane values may be needed
later (e.g for a load instruction) so we still need to calculate
these values to avoid extractelement being called on the vector part.

Differential Revision: https://reviews.llvm.org/D109445