[LVI] Use a SmallDenseMap instead of std::map for ValueCacheEntryTy

Historically there seems to be some resistance regarding the change to DenseMap
(r147980). However, I couldn't find cases of iterator invalidation for
ValueCacheEntryTy, but only for ValueCache, which I left untouched.

This reduces 20s on an internal testcase. Follow up from r245309.

Differential Revision: http://reviews.llvm.org/D11651

rdar://problem/21320066

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@245314 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Analysis/LazyValueInfo.cpp b/lib/Analysis/LazyValueInfo.cpp
index 5904257492be66e621c3a7335b2219e76e119d99..a984bc40ac1034d9fea8a7a258ec7eb9acef2969 100644 (file)
--- a/lib/Analysis/LazyValueInfo.cpp
+++ b/lib/Analysis/LazyValueInfo.cpp
@@ -315,7 +315,8 @@ namespace {
     /// This is all of the cached block information for exactly one Value*.
     /// The entries are sorted by the BasicBlock* of the
     /// entries, allowing us to do a lookup with a binary search.
-    typedef std::map<AssertingVH<BasicBlock>, LVILatticeVal> ValueCacheEntryTy;
+    typedef SmallDenseMap<AssertingVH<BasicBlock>, LVILatticeVal, 4>
+        ValueCacheEntryTy;
 
     /// This is all of the cached information for all values,
     /// mapped from Value* to key information.