From: Owen Anderson <resistor@mac.com>
Date: Wed, 29 Sep 2010 22:57:02 +0000 (+0000)
Subject: Early CFG simplification can fold conditionals down to selects, which is often a... 
X-Git-Url: http://plrg.eecs.uci.edu/git/?a=commitdiff_plain;h=235859d66cf3faf7453f59b92cc4056ae1855d0d;p=oota-llvm.git

Early CFG simplification can fold conditionals down to selects, which is often a good thing, but it can also
hide jump threading opportunities by turning control flow into data flow.  Run an early JumpThreading pass
(adds approximately an additional 1% to optimization time on SPEC), allowing it to get a shot at these cases
first.  Fixes <rdar://problem/8447345>.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115099 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/include/llvm/Support/StandardPasses.h b/include/llvm/Support/StandardPasses.h
index 1b40ede14b8..6e76426b205 100644
--- a/include/llvm/Support/StandardPasses.h
+++ b/include/llvm/Support/StandardPasses.h
@@ -104,6 +104,7 @@ namespace llvm {
       PM->add(createDeadArgEliminationPass());  // Dead argument elimination
     }
     PM->add(createInstructionCombiningPass());  // Clean up after IPCP & DAE
+    PM->add(createJumpThreadingPass());         // Clean up after IPCP & DAE
     PM->add(createCFGSimplificationPass());     // Clean up after IPCP & DAE
     
     // Start of CallGraph SCC passes.