[TTI] Add a hook for specifying per-target defaults for Interleaved Accesses

Summary:
This adds a hook to TTI which enables us to selectively turn on by default
interleaved access vectorization for targets on which we have have performed
the required benchmarking.

Reviewers: rengolin

Subscribers: rengolin, llvm-commits

Differential Revision: http://reviews.llvm.org/D11901

llvm-svn: 244449

llvm/include/llvm/Analysis/TargetTransformInfo.h

@@ -357,6 +357,9 @@ public:
   /// \brief Don't restrict interleaved unrolling to small loops.
   bool enableAggressiveInterleaving(bool LoopHasReductions) const;
 
+  /// \brief Enable matching of interleaved access groups.
+  bool enableInterleavedAccessVectorization() const;
+
   /// \brief Return hardware support for population count.
   PopcntSupportKind getPopcntSupport(unsigned IntTyWidthInBit) const;
 
@@ -578,6 +581,7 @@ public:
   virtual unsigned getJumpBufSize() = 0;
   virtual bool shouldBuildLookupTables() = 0;
   virtual bool enableAggressiveInterleaving(bool LoopHasReductions) = 0;
+  virtual bool enableInterleavedAccessVectorization() = 0;
   virtual PopcntSupportKind getPopcntSupport(unsigned IntTyWidthInBit) = 0;
   virtual bool haveFastSqrt(Type *Ty) = 0;
   virtual int getFPOpCost(Type *Ty) = 0;
@@ -719,6 +723,9 @@ public:
   bool enableAggressiveInterleaving(bool LoopHasReductions) override {
     return Impl.enableAggressiveInterleaving(LoopHasReductions);
   }
+  bool enableInterleavedAccessVectorization() override {
+    return Impl.enableInterleavedAccessVectorization();
+  }
   PopcntSupportKind getPopcntSupport(unsigned IntTyWidthInBit) override {
     return Impl.getPopcntSupport(IntTyWidthInBit);
   }

llvm/include/llvm/Analysis/TargetTransformInfoImpl.h

@@ -230,6 +230,8 @@ public:
 
   bool enableAggressiveInterleaving(bool LoopHasReductions) { return false; }
 
+  bool enableInterleavedAccessVectorization() { return false; }
+
   TTI::PopcntSupportKind getPopcntSupport(unsigned IntTyWidthInBit) {
     return TTI::PSK_Software;
   }

llvm/lib/Analysis/TargetTransformInfo.cpp

@@ -166,6 +166,10 @@ bool TargetTransformInfo::enableAggressiveInterleaving(bool LoopHasReductions) c
   return TTIImpl->enableAggressiveInterleaving(LoopHasReductions);
 }
 
+bool TargetTransformInfo::enableInterleavedAccessVectorization() const {
+  return TTIImpl->enableInterleavedAccessVectorization();
+}
+
 TargetTransformInfo::PopcntSupportKind
 TargetTransformInfo::getPopcntSupport(unsigned IntTyWidthInBit) const {
   return TTIImpl->getPopcntSupport(IntTyWidthInBit);

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

@@ -3880,9 +3880,15 @@ bool LoopVectorizationLegality::canVectorize() {
                        : "")
                << "!\n");
 
+  bool UseInterleaved = TTI->enableInterleavedAccessVectorization();
+
+  // If an override option has been passed in for interleaved accesses, use it.
+  if (EnableInterleavedMemAccesses.getNumOccurrences() > 0)
+    UseInterleaved = EnableInterleavedMemAccesses;
+
   // Analyze interleaved memory accesses.
-  if (EnableInterleavedMemAccesses)
+  if (UseInterleaved)
-    InterleaveInfo.analyzeInterleaving(Strides);
+     InterleaveInfo.analyzeInterleaving(Strides);
 
   // Okay! We can vectorize. At this point we don't have any other mem analysis
   // which may limit our maximum vectorization factor, so just return true with