| commit | f06f68a796f28576a5f1464753de31493b1b3da9 | [log] [tgz] |
|---|---|---|
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | Wed Mar 08 01:06:58 2017 +0000 |
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | Wed Mar 08 01:06:58 2017 +0000 |
| tree | 9378d724b508f0c25b1675e9449f814b9e768b54 | |
| parent | 9ac562c93382e55ad60545ff086de85b9e3ca084 [diff] |
AMDGPU: Don't wait at end of block with a trivial successor If there is only one successor, and that successor only has one predecessor the wait can obviously be delayed until uses or the end of the next block. This avoids code quality regressions when there are trivial fallthrough blocks inserted for structurization. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297251 91177308-0d34-0410-b5e6-96231b3b80d8