diff options
| author | Marek Olsak <marek.olsak@amd.com> | 2018-01-31 20:18:11 +0000 |
|---|---|---|
| committer | Marek Olsak <marek.olsak@amd.com> | 2018-01-31 20:18:11 +0000 |
| commit | d4bb329d0ea29bf6882b8f3bee9b944c161980a3 (patch) | |
| tree | fd66d564231a8a63a11ac159574fcb730a499eb7 /llvm/unittests/CodeGen/MachineInstrTest.cpp | |
| parent | 13e4741275a1169c2ff50645cd6f4e964a94169f (diff) | |
| download | llvm-d4bb329d0ea29bf6882b8f3bee9b944c161980a3.zip llvm-d4bb329d0ea29bf6882b8f3bee9b944c161980a3.tar.gz llvm-d4bb329d0ea29bf6882b8f3bee9b944c161980a3.tar.bz2 | |
AMDGPU: Fold inline offset for loads properly in moveToVALU on GFX9
Summary:
This enables load merging into x2, x4, which is driven by inline offsets.
6500 shaders are affected:
Code Size in affected shaders: -15.14 %
Reviewers: arsenm, nhaehnle
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D42078
llvm-svn: 323909
Diffstat (limited to 'llvm/unittests/CodeGen/MachineInstrTest.cpp')
0 files changed, 0 insertions, 0 deletions