Simplify (X<<1)-X into X. According to my auto-simplier this is the most common missed

simplification in fully optimized code.  It occurs sporadically in the testsuite, and
many times in 403.gcc: the final bitcode has 131 fewer subtractions after this change.
The reason that the multiplies are not eliminated is the same reason that instcombine
did not catch this: they are used by other instructions (instcombine catches this with
a more general transform which in general is only profitable if the operands have only
one use).

llvm-svn: 123754

1 files changed, 6 insertions, 0 deletions

diff --git a/llvm/lib/Analysis/InstructionSimplify.cpp b/llvm/lib/Analysis/InstructionSimplify.cpp
index 1d4ff0fc9e0..f737354b36a 100644
--- a/llvm/lib/Analysis/InstructionSimplify.cpp
+++ b/llvm/lib/Analysis/InstructionSimplify.cpp
@@ -593,6 +593,12 @@ static Value *SimplifySubInst(Value *Op0, Value *Op1, bool isNSW, bool isNUW,
       match(Op0, m_Add(m_Specific(Op1), m_Value(X))))
     return X;
 
+  // (X*2) - X -> X
+  // (X<<1) - X -> X
+  if (match(Op0, m_Mul(m_Specific(Op1), m_ConstantInt<2>())) ||
+      match(Op0, m_Shl(m_Specific(Op1), m_One())))
+    return Op1;
+
   // i1 sub -> xor.
   if (MaxRecurse && Op0->getType()->isIntegerTy(1))
     if (Value *V = SimplifyXorInst(Op0, Op1, TD, DT, MaxRecurse-1))