projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: d0458e5) | patch
When we do the single-store optimization, delete both the store
authorChris Lattner <sabre@nondot.org>
Sat, 4 Aug 2007 02:38:38 +0000 (02:38 +0000)
committerChris Lattner <sabre@nondot.org>
Sat, 4 Aug 2007 02:38:38 +0000 (02:38 +0000)
commit4f63e76cda86b676e4f0e31fd35b812e2f8dd57f
tree790e1e26645ebfae8c5f71787fb2df2eb738de34tree | snapshot
parentd0458e5d4f29ce1e9ba1def22d06b362ae2c1999commit | diff
When we do the single-store optimization, delete both the store
and the alloca so they don't get reprocessed.

This speeds up PR1432 from 2.20s to 2.17s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40812 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Transforms/Utils/PromoteMemoryToRegister.cpp diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors