| commit | 544b9b426f0dfa1beef6dc3640607dee3ca1160e | [log] [tgz] |
|---|---|---|
| author | Chad Rosier <mcrosier@apple.com> | Thu Mar 22 00:21:17 2012 +0000 |
| committer | Chad Rosier <mcrosier@apple.com> | Thu Mar 22 00:21:17 2012 +0000 |
| tree | 72e4f0b64ca11a3a78633db46c6d1f94c61445b9 | |
| parent | 55f43d6b7e3ad1f08f5392205ab0feabf4c9933d [diff] |
[fast-isel] Fold "urem x, pow2" -> "and x, pow2-1". This should fix the 271% execution-time regression for nsieve-bits on the ARMv7 -O0 -g nightly tester. This may also improve compile-time on architectures that would otherwise generate a libcall for urem (e.g., ARM) or fall back to the DAG selector. rdar://10810716 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153230 91177308-0d34-0410-b5e6-96231b3b80d8