From f1a7403e86d001a9df70f03668cfce37f5b70a0f Mon Sep 17 00:00:00 2001 From: Chandler Carruth Date: Mon, 28 Apr 2014 09:34:03 +0000 Subject: [PATCH] Fix very poor compile-time in PR19499 due to excessive tree walks in domtree. When finding a nearest common dominator, if neither A dominates B nor B dominates A, we immediately resorted to a tree walk. The tree walk here is *particularly* expensive because we have to build a (potentially very large) set for one side's dominators and compare it with the other side's. If at any point we have DFS info, we don't need to do any of this. We can just walk up one side's immediate dominators and return the first one which dominates the other side. Because of the DFS info, the dominates queries are trivially constant time. This reduces the optimizers time in the test case on PR19499 by 70%. It now optimizes in about 30 seconds for me. And there is still more to be done for this case. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207406 91177308-0d34-0410-b5e6-96231b3b80d8 --- include/llvm/Support/GenericDomTree.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/include/llvm/Support/GenericDomTree.h b/include/llvm/Support/GenericDomTree.h index 5a44077a2c0..e344220d953 100644 --- a/include/llvm/Support/GenericDomTree.h +++ b/include/llvm/Support/GenericDomTree.h @@ -453,6 +453,21 @@ public: DomTreeNodeBase *NodeA = getNode(A); DomTreeNodeBase *NodeB = getNode(B); + // If we have DFS info, then we can avoid all allocations by just querying + // it from each IDom. Note that because we call 'dominates' twice above, we + // expect to call through this code at most 16 times in a row without + // building valid DFS information. This is important as below is a *very* + // slow tree walk. + if (DFSInfoValid) { + DomTreeNodeBase *IDomA = NodeA->getIDom(); + while (IDomA) { + if (NodeB->DominatedBy(IDomA)) + return IDomA->getBlock(); + IDomA = IDomA->getIDom(); + } + return nullptr; + } + // Collect NodeA dominators set. SmallPtrSet*, 16> NodeADoms; NodeADoms.insert(NodeA); -- 2.34.1