Add optimization remarks to the loop unroller and vectorizer.

Summary:
This calls emitOptimizationRemark from the loop unroller and vectorizer
at the point where they make a positive transformation. For the
vectorizer, it reports vectorization and interleave factors. For the
loop unroller, it reports all the different supported types of
unrolling.

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D3456

llvm-svn: 207528

3 files changed, 45 insertions, 0 deletions

diff --git a/llvm/include/llvm/Analysis/LoopInfo.h b/llvm/include/llvm/Analysis/LoopInfo.h
index 045d4dc565e..bef03e91bbb 100644
--- a/llvm/include/llvm/Analysis/LoopInfo.h
+++ b/llvm/include/llvm/Analysis/LoopInfo.h
@@ -453,6 +453,31 @@ public:
 
   void dump() const;
 
+  /// \brief Return the debug location of the start of this loop.
+  /// This looks for a BB terminating instruction with a known debug
+  /// location by looking at the preheader and header blocks. If it
+  /// cannot find a terminating instruction with location information,
+  /// it returns an unknown location.
+  DebugLoc getStartLoc() const {
+    DebugLoc StartLoc;
+    BasicBlock *HeadBB;
+
+    // Try the pre-header first.
+    if ((HeadBB = getLoopPreheader()) != nullptr) {
+      StartLoc = HeadBB->getTerminator()->getDebugLoc();
+      if (!StartLoc.isUnknown())
+        return StartLoc;
+    }
+
+    // If we have no pre-header or there are no instructions with debug
+    // info in it, try the header.
+    HeadBB = getHeader();
+    if (HeadBB)
+      StartLoc = HeadBB->getTerminator()->getDebugLoc();
+
+    return StartLoc;
+  }
+
 private:
   friend class LoopInfoBase<BasicBlock, Loop>;
   explicit Loop(BasicBlock *BB) : LoopBase<BasicBlock, Loop>(BB) {}
diff --git a/llvm/lib/Transforms/Utils/LoopUnroll.cpp b/llvm/lib/Transforms/Utils/LoopUnroll.cpp
index 668324881e3..faaab5c12e4 100644
--- a/llvm/lib/Transforms/Utils/LoopUnroll.cpp
+++ b/llvm/lib/Transforms/Utils/LoopUnroll.cpp
@@ -24,6 +24,7 @@
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/IR/BasicBlock.h"
 #include "llvm/IR/Dominators.h"
+#include "llvm/IR/LLVMContext.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/raw_ostream.h"
 #include "llvm/Transforms/Utils/BasicBlockUtils.h"
@@ -228,20 +229,33 @@ bool llvm::UnrollLoop(Loop *L, unsigned Count, unsigned TripCount,
       (unsigned)GreatestCommonDivisor64(Count, TripMultiple);
   }
 
+  // Report the unrolling decision.
+  DebugLoc LoopLoc = L->getStartLoc();
+  Function *F = Header->getParent();
+  LLVMContext &Ctx = F->getContext();
+
   if (CompletelyUnroll) {
     DEBUG(dbgs() << "COMPLETELY UNROLLING loop %" << Header->getName()
           << " with trip count " << TripCount << "!\n");
+    Ctx.emitOptimizationRemark(DEBUG_TYPE, *F, LoopLoc,
+                               Twine("completely unrolled loop with ") +
+                                   Twine(TripCount) + " iterations");
   } else {
     DEBUG(dbgs() << "UNROLLING loop %" << Header->getName()
           << " by " << Count);
+    Twine DiagMsg("unrolled loop by a factor of " + Twine(Count));
     if (TripMultiple == 0 || BreakoutTrip != TripMultiple) {
       DEBUG(dbgs() << " with a breakout at trip " << BreakoutTrip);
+      DiagMsg.concat(" with a breakout at trip " + Twine(BreakoutTrip));
     } else if (TripMultiple != 1) {
       DEBUG(dbgs() << " with " << TripMultiple << " trips per branch");
+      DiagMsg.concat(" with " + Twine(TripMultiple) + " trips per branch");
     } else if (RuntimeTripCount) {
       DEBUG(dbgs() << " with run-time trip count");
+      DiagMsg.concat(" with run-time trip count");
     }
     DEBUG(dbgs() << "!\n");
+    Ctx.emitOptimizationRemark(DEBUG_TYPE, *F, LoopLoc, DiagMsg);
   }
 
   bool ContinueOnTrue = L->contains(BI->getSuccessor(0));
diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index cf210d679d8..586c220162f 100644
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -1222,6 +1222,12 @@ struct LoopVectorize : public FunctionPass {
     // Mark the loop as already vectorized to avoid vectorizing again.
     Hints.setAlreadyVectorized(L);
 
+    // Report the vectorization decision.
+    F->getContext().emitOptimizationRemark(
+        DEBUG_TYPE, *F, L->getStartLoc(),
+        Twine("vectorized loop (vectorization factor: ") + Twine(VF.Width) +
+            ", unroll factor: " + Twine(UF) + ")");
+
     DEBUG(verifyFunction(*L->getHeader()->getParent()));
     return true;
   }