commit a998735def5f04459fa6cbeaaffbb42f1a1dec5c
author Kevin Qin <Kevin.Qin@arm.com> Mon Mar 09 06:14:07 2015 +0000
committer Kevin Qin <Kevin.Qin@arm.com> Mon Mar 09 06:14:07 2015 +0000
tree 82a5ff86e24b5e27caffbb29b94df869abb4974a
parent 7b68b90c159924f76d25fadf43ff72469b7ba573

Run LICM pass after loop unrolling pass.

Runtime unrollng will introduce a runtime check in loop prologue.
If the unrolled loop is a inner loop, then the proglogue will be inside
the outer loop. LICM pass can help to promote the runtime check out if
the checked value is loop invariant.

llvm-svn: 231630

llvm/lib/Transforms/IPO/PassManagerBuilder.cpp

diff --git a/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp b/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
index 21fa34d..7ff55a7 100644
--- a/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
+++ b/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp
@@ -364,8 +364,14 @@
   MPM.add(createCFGSimplificationPass());
   MPM.add(createInstructionCombiningPass());
 
-  if (!DisableUnrollLoops)
+  if (!DisableUnrollLoops) {
     MPM.add(createLoopUnrollPass());    // Unroll small loops
+    // Runtime unrollng will introduce runtime check in loop prologue. If the
+    // unrolled loop is a inner loop, then the prologue will be inside the
+    // outer loop. LICM pass can help to promote the runtime check out if the
+    // checked value is loop invariant.
+    MPM.add(createLICMPass());
+  }
 
   // After vectorization and unrolling, assume intrinsics may tell us more
   // about pointer alignments.

llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp

diff --git a/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp b/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp
index 91b688c..9e2d8c9 100644
--- a/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp
+++ b/llvm/lib/Transforms/Utils/LoopUnrollRuntime.cpp
@@ -142,7 +142,7 @@
                             BasicBlock *InsertTop, BasicBlock *InsertBot,
                             std::vector<BasicBlock *> &NewBlocks,
                             LoopBlocksDFS &LoopBlocks, ValueToValueMapTy &VMap,
-                            LoopInfo *LI) {
+                            LoopInfo *LI, LPPassManager *LPM) {
   BasicBlock *Preheader = L->getLoopPreheader();
   BasicBlock *Header = L->getHeader();
   BasicBlock *Latch = L->getLoopLatch();
@@ -153,10 +153,7 @@
   Loop *ParentLoop = L->getParentLoop();
   if (!UnrollProlog) {
     NewLoop = new Loop();
-    if (ParentLoop)
-      ParentLoop->addChildLoop(NewLoop);
-    else
-      LI->addTopLevelLoop(NewLoop);
+    LPM->insertLoop(NewLoop, ParentLoop);
   }
 
   // For each block in the original loop, create a new copy,
@@ -390,7 +387,7 @@
   // the loop, otherwise we create a cloned loop to execute the extra
   // iterations. This function adds the appropriate CFG connections.
   CloneLoopBlocks(L, ModVal, UnrollPrologue, PH, PEnd, NewBlocks, LoopBlocks,
-                  VMap, LI);
+                  VMap, LI, LPM);
 
   // Insert the cloned blocks into function just before the original loop
   F->getBasicBlockList().splice(PEnd, F->getBasicBlockList(), NewBlocks[0],