diff options
| author | Konstantin Zhuravlyov <kzhuravl_dev@outlook.com> | 2016-03-29 15:15:44 +0000 |
|---|---|---|
| committer | Konstantin Zhuravlyov <kzhuravl_dev@outlook.com> | 2016-03-29 15:15:44 +0000 |
| commit | ecc7cbf611f77dada2b4124c4f021364c1c5d288 (patch) | |
| tree | cd467ab59914de405cd71316395d7182bbc94ff7 /llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp | |
| parent | 773c3b0d7c1e58ecc431402117e47a94e3b252fc (diff) | |
| download | llvm-ecc7cbf611f77dada2b4124c4f021364c1c5d288.zip llvm-ecc7cbf611f77dada2b4124c4f021364c1c5d288.tar.gz llvm-ecc7cbf611f77dada2b4124c4f021364c1c5d288.tar.bz2 | |
Test commit access
llvm-svn: 264736
Diffstat (limited to 'llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp')
| -rw-r--r-- | llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp b/llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp index 1452d81..d98bd70 100644 --- a/llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp +++ b/llvm/lib/Target/AMDGPU/SILoadStoreOptimizer.cpp @@ -26,7 +26,7 @@ // // - Live interval recomputing seems inefficient. This currently only matches // one pair, and recomputes live intervals and moves on to the next pair. It -// would be better to compute a list of all merges that need to occur +// would be better to compute a list of all merges that need to occur. // // - With a list of instructions to process, we can also merge more. If a // cluster of loads have offsets that are too large to fit in the 8-bit |