[all-commits] [llvm/llvm-project] a36d26: [VPlan] Avoid collecting scalars for SVE

[malJaj via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a36d269658dfd8b16dd8bdcb1726b9206b1d97ff
      https://github.com/llvm/llvm-project/commit/a36d269658dfd8b16dd8bdcb1726b9206b1d97ff
  Author: Malhar Jajoo <malhar.jajoo at arm.com>
  Date:   2022-03-16 (Wed, 16 Mar 2022)

  Changed paths:
    M llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
    A llvm/test/Transforms/LoopVectorize/AArch64/scalable-avoid-scalarization.ll

  Log Message:
  -----------
  [VPlan] Avoid collecting scalars for SVE

This patch ensures scalars (except for uniforms) are no
longer collected (prior to LVP planning phase) for
scalable vectorization.

This is to avoid the chances of generating scalarized
instructions later (during LVP execute phase) as they
are not supported for scalable vectorization.

Relevant test has also been added.

Differential Revision: https://reviews.llvm.org/D121452