projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 03fce70) | patch
This patch aims to improve compile time performance by increasing
authorPreston Gurd <preston.gurd@intel.com>
Fri, 1 Feb 2013 20:41:27 +0000 (20:41 +0000)
committerPreston Gurd <preston.gurd@intel.com>
Fri, 1 Feb 2013 20:41:27 +0000 (20:41 +0000)
commit83474ee594b5b14e9071564814a90571805cc433
tree3e14a0c453560222f29831d51dfb668724d30276tree | snapshot
parent03fce7074ebf875485a74e026e3e3a165a602166commit | diff
This patch aims to improve compile time performance by increasing
the SCEV vector size in LoopStrengthReduce. It is observed that
the BaseRegs vector size is 4 in most cases,
and elements are frequently copied when it is initialized as
SmallVector<const SCEV *, 2> BaseRegs.
Our benchmark results show that the compilation time performance
improved by ~0.5%.

Patch by Wan Xiaofei.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174219 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Scalar/LoopStrengthReduce.cpp diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors