[all-commits] [llvm/llvm-project] 4eecf1: [AMDGPU] Always expand ext/insertelement with dive...

[Stanislav Mekhanoshin via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 4eecf171645e064ff251ea3f706d3252b54da2f8
      https://github.com/llvm/llvm-project/commit/4eecf171645e064ff251ea3f706d3252b54da2f8
  Author: Stanislav Mekhanoshin <Stanislav.Mekhanoshin at amd.com>
  Date:   2020-05-20 (Wed, 20 May 2020)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/extract_vector_dynelt.ll
    M llvm/test/CodeGen/AMDGPU/indirect-addressing-si-gfx9.ll
    M llvm/test/CodeGen/AMDGPU/indirect-addressing-si-pregfx9.ll
    M llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll
    M llvm/test/CodeGen/AMDGPU/insert_vector_dynelt.ll
    M llvm/test/CodeGen/AMDGPU/scratch-simple.ll

  Log Message:
  -----------
  [AMDGPU] Always expand ext/insertelement with divergent idx

Even though series of cmd/cndmask can produce quite a lot of
code that is still better than a loop. In case of doubles we
would even produce two loops.

Differential Revision: https://reviews.llvm.org/D80032