Add getUnrollingPreferences to TTI

Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190542 91177308-0d34-0410-b5e6-96231b3b80d8

include/llvm/Analysis/TargetTransformInfo.h

@@ -29,6 +29,7 @@
 namespace llvm {
 
 class GlobalValue;
+class Loop;
 class Type;
 class User;
 class Value;
@@ -191,6 +192,36 @@ public:
   /// incurs significant execution cost.
   virtual bool isLoweredToCall(const Function *F) const;
 
+  /// Parameters that control the generic loop unrolling transformation.
+  struct UnrollingPreferences {
+    /// The cost threshold for the unrolled loop, compared to
+    /// CodeMetrics.NumInsts aggregated over all basic blocks in the loop body.
+    /// The unrolling factor is set such that the unrolled loop body does not
+    /// exceed this cost. Set this to UINT_MAX to disable the loop body cost
+    /// restriction.
+    unsigned Threshold;
+    /// The cost threshold for the unrolled loop when optimizing for size (set
+    /// to UINT_MAX to disable).
+    unsigned OptSizeThreshold;
+    /// A forced unrolling factor (the number of concatenated bodies of the
+    /// original loop in the unrolled loop body). When set to 0, the unrolling
+    /// transformation will select an unrolling factor based on the current cost
+    /// threshold and other factors.
+    unsigned Count;
+    /// Allow partial unrolling (unrolling of loops to expand the size of the
+    /// loop body, not only to eliminate small constant-trip-count loops).
+    bool     Partial;
+    /// Allow runtime unrolling (unrolling of loops to expand the size of the
+    /// loop body even when the number of loop iterations is not known at compile
+    /// time).
+    bool     Runtime;
+  };
+
+  /// \brief Get target-customized preferences for the generic loop unrolling
+  /// transformation. The caller will initialize UP with the current
+  /// target-independent defaults.
+  virtual void getUnrollingPreferences(Loop *L, UnrollingPreferences &UP) const;
+
   /// @}
 
   /// \name Scalar Target Information

lib/Analysis/TargetTransformInfo.cpp

@@ -96,6 +96,11 @@ bool TargetTransformInfo::isLoweredToCall(const Function *F) const {
   return PrevTTI->isLoweredToCall(F);
 }
 
+void TargetTransformInfo::getUnrollingPreferences(Loop *L,
+                            UnrollingPreferences &UP) const {
+  PrevTTI->getUnrollingPreferences(L, UP);
+}
+
 bool TargetTransformInfo::isLegalAddImmediate(int64_t Imm) const {
   return PrevTTI->isLegalAddImmediate(Imm);
 }
@@ -469,6 +474,8 @@ struct NoTTI : ImmutablePass, TargetTransformInfo {
     return true;
   }
 
+  void getUnrollingPreferences(Loop *, UnrollingPreferences &) const { }
+
   bool isLegalAddImmediate(int64_t Imm) const {
     return false;
   }

lib/CodeGen/BasicTargetTransformInfo.cpp

@@ -84,6 +84,7 @@ public:
   virtual unsigned getJumpBufSize() const;
   virtual bool shouldBuildLookupTables() const;
   virtual bool haveFastSqrt(Type *Ty) const;
+  virtual void getUnrollingPreferences(Loop *L, UnrollingPreferences &UP) const;
 
   /// @}
 
@@ -189,6 +190,8 @@ bool BasicTTI::haveFastSqrt(Type *Ty) const {
   return TLI->isTypeLegal(VT) && TLI->isOperationLegalOrCustom(ISD::FSQRT, VT);
 }
 
+void BasicTTI::getUnrollingPreferences(Loop *, UnrollingPreferences &) const { }
+
 //===----------------------------------------------------------------------===//
 //
 // Calls used by the vectorizers.

lib/Transforms/Scalar/LoopUnrollPass.cpp

@@ -55,6 +55,9 @@ namespace {
       CurrentAllowPartial = (P == -1) ? UnrollAllowPartial : (bool)P;
 
       UserThreshold = (T != -1) || (UnrollThreshold.getNumOccurrences() > 0);
+      UserAllowPartial = (P != -1) ||
+                         (UnrollAllowPartial.getNumOccurrences() > 0);
+      UserCount = (C != -1) || (UnrollCount.getNumOccurrences() > 0);
 
       initializeLoopUnrollPass(*PassRegistry::getPassRegistry());
     }
@@ -75,7 +78,9 @@ namespace {
     unsigned CurrentCount;
     unsigned CurrentThreshold;
     bool     CurrentAllowPartial;
+    bool     UserCount;            // CurrentCount is user-specified.
     bool     UserThreshold;        // CurrentThreshold is user-specified.
+    bool     UserAllowPartial;     // CurrentAllowPartial is user-specified.
 
     bool runOnLoop(Loop *L, LPPassManager &LPM);
 
@@ -145,16 +150,24 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
         << "] Loop %" << Header->getName() << "\n");
   (void)Header;
 
+  TargetTransformInfo::UnrollingPreferences UP;
+  UP.Threshold = CurrentThreshold;
+  UP.OptSizeThreshold = OptSizeUnrollThreshold;
+  UP.Count = CurrentCount;
+  UP.Partial = CurrentAllowPartial;
+  UP.Runtime = UnrollRuntime;
+  TTI.getUnrollingPreferences(L, UP);
+
   // Determine the current unrolling threshold.  While this is normally set
   // from UnrollThreshold, it is overridden to a smaller value if the current
   // function is marked as optimize-for-size, and the unroll threshold was
   // not user specified.
-  unsigned Threshold = CurrentThreshold;
+  unsigned Threshold = UserThreshold ? CurrentThreshold : UP.Threshold;
   if (!UserThreshold &&
       Header->getParent()->getAttributes().
         hasAttribute(AttributeSet::FunctionIndex,
                      Attribute::OptimizeForSize))
-    Threshold = OptSizeUnrollThreshold;
+    Threshold = UP.OptSizeThreshold;
 
   // Find trip count and trip multiple if count is not available
   unsigned TripCount = 0;
@@ -167,11 +180,15 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
     TripCount = SE->getSmallConstantTripCount(L, LatchBlock);
     TripMultiple = SE->getSmallConstantTripMultiple(L, LatchBlock);
   }
+
+  bool Runtime = UnrollRuntime.getNumOccurrences() == 0 ?
+                 UP.Runtime : UnrollRuntime;
+
   // Use a default unroll-count if the user doesn't specify a value
   // and the trip count is a run-time value.  The default is different
   // for run-time or compile-time trip count loops.
-  unsigned Count = CurrentCount;
-  if (UnrollRuntime && CurrentCount == 0 && TripCount == 0)
+  unsigned Count = UserCount ? CurrentCount : UP.Count;
+  if (Runtime && Count == 0 && TripCount == 0)
     Count = UnrollRuntimeCount;
 
   if (Count == 0) {
@@ -204,7 +221,8 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
     if (TripCount != 1 && Size > Threshold) {
       DEBUG(dbgs() << "  Too large to fully unroll with count: " << Count
             << " because size: " << Size << ">" << Threshold << "\n");
-      if (!CurrentAllowPartial && !(UnrollRuntime && TripCount == 0)) {
+      bool AllowPartial = UserAllowPartial ? CurrentAllowPartial : UP.Partial;
+      if (!AllowPartial && !(Runtime && TripCount == 0)) {
         DEBUG(dbgs() << "  will not try to unroll partially because "
               << "-unroll-allow-partial not given\n");
         return false;
@@ -215,7 +233,7 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
         while (Count != 0 && TripCount%Count != 0)
           Count--;
       }
-      else if (UnrollRuntime) {
+      else if (Runtime) {
         // Reduce unroll count to be a lower power-of-two value
         while (Count != 0 && Size > Threshold) {
           Count >>= 1;
@@ -231,7 +249,7 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
   }
 
   // Unroll the loop.
-  if (!UnrollLoop(L, Count, TripCount, UnrollRuntime, TripMultiple, LI, &LPM))
+  if (!UnrollLoop(L, Count, TripCount, Runtime, TripMultiple, LI, &LPM))
     return false;
 
   return true;