X86 cost model: Add cost for vectorized gather/scather

radar://14351991

llvm-svn: 186189

1 files changed, 15 insertions, 0 deletions

diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 68e1a672436..3bbddadca88 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -100,6 +100,8 @@ public:
                                    unsigned Alignment,
                                    unsigned AddressSpace) const;
 
+  virtual unsigned getAddressComputationCost(Type *PtrTy, bool IsComplex) const;
+
   /// @}
 };
 
@@ -598,3 +600,16 @@ unsigned X86TTI::getMemoryOpCost(unsigned Opcode, Type *Src, unsigned Alignment,
 
   return Cost;
 }
+
+unsigned X86TTI::getAddressComputationCost(Type *Ty, bool IsComplex) const {
+  // Address computations in vectorized code with non-consecutive addresses will
+  // likely result in more instructions compared to scalar code where the
+  // computation can more often be merged into the index mode. The resulting
+  // extra micro-ops can significantly decrease throughput.
+  unsigned NumVectorInstToHideOverhead = 10;
+
+  if (Ty->isVectorTy() && IsComplex)
+    return NumVectorInstToHideOverhead;
+
+  return TargetTransformInfo::getAddressComputationCost(Ty, IsComplex);
+}