Based on the testcase for pr3120, running on my MacPro with Xeon processors,

it is definitely profitable to tail duplicate indirect branches for x86.
This is likely to be true to various degrees for all modern x86 processors.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@89865 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86InstrInfo.h b/lib/Target/X86/X86InstrInfo.h
index c6daa25..3d4c2f6 100644 (file)
--- a/lib/Target/X86/X86InstrInfo.h
+++ b/lib/Target/X86/X86InstrInfo.h
@@ -632,6 +632,8 @@ public:
   ///
   unsigned getGlobalBaseReg(MachineFunction *MF) const;
 
+  virtual bool isProfitableToDuplicateIndirectBranch() const { return true; }
+
 private:
   MachineInstr* foldMemoryOperandImpl(MachineFunction &MF,
                                      MachineInstr* MI,