diff options
| author | Dan Gohman <gohman@apple.com> | 2009-10-22 00:03:58 +0000 |
|---|---|---|
| committer | Dan Gohman <gohman@apple.com> | 2009-10-22 00:03:58 +0000 |
| commit | ff97acd8f1c9a0b067ad97e4987e0e48d0d8f39e (patch) | |
| tree | 502895a183e183e4e3043a4bcb84dab0f42155ac /llvm/lib/Bitcode/Writer/BitcodeWriter.cpp | |
| parent | 649bf5c04b6010f8b1a4cf8fadab941cf0baf420 (diff) | |
| download | llvm-ff97acd8f1c9a0b067ad97e4987e0e48d0d8f39e.zip llvm-ff97acd8f1c9a0b067ad97e4987e0e48d0d8f39e.tar.gz llvm-ff97acd8f1c9a0b067ad97e4987e0e48d0d8f39e.tar.bz2 | |
Revert the main portion of r31856. It was causing BranchFolding
to break up CFG diamonds by banishing one of the blocks to the end of
the function, which is bad for code density and branch size.
This does pessimize MultiSource/Benchmarks/Ptrdist/yacr2, the
benchmark cited as the reason for the change, however I've examined
the code and it looks more like a case of gaming a particular
branch than of being generally applicable.
llvm-svn: 84803
Diffstat (limited to 'llvm/lib/Bitcode/Writer/BitcodeWriter.cpp')
0 files changed, 0 insertions, 0 deletions