From e041a30bb9abf8080841991dce5d00e3d663a80a Mon Sep 17 00:00:00 2001 From: Chandler Carruth Date: Thu, 5 Jan 2012 11:05:55 +0000 Subject: Prevent a DAGCombine from firing where there are two uses of a combined-away node and the result of the combine isn't substantially smaller than the input, it's just canonicalized. This is the first part of a significant (7%) performance gain for Snappy's hot decompression loop. llvm-svn: 147604 --- llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp') diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp index 8a8f0cf2bc7..5b2d86f511d 100644 --- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp @@ -3331,7 +3331,9 @@ SDValue DAGCombiner::visitSHL(SDNode *N) { // fold (shl (srl x, c1), c2) -> (and (shl x, (sub c2, c1), MASK) or // (and (srl x, (sub c1, c2), MASK) - if (N1C && N0.getOpcode() == ISD::SRL && + // Only fold this if the inner shift has no other uses -- if it does, folding + // this will increase the total number of instructions. + if (N1C && N0.getOpcode() == ISD::SRL && N0.hasOneUse() && N0.getOperand(1).getOpcode() == ISD::Constant) { uint64_t c1 = cast(N0.getOperand(1))->getZExtValue(); if (c1 < VT.getSizeInBits()) { -- cgit v1.2.3