diff --git a/include/llvm/CodeGen/SelectionDAG.h b/include/llvm/CodeGen/SelectionDAG.h
index 28535d12737..5434fb632b4 100644
--- a/include/llvm/CodeGen/SelectionDAG.h
+++ b/include/llvm/CodeGen/SelectionDAG.h
@@ -477,6 +477,11 @@ public:
                         const SDOperand *Ops, unsigned NumOps);
   SDNode *getTargetNode(unsigned Opcode, std::vector<MVT::ValueType> &ResultTys,
                         const SDOperand *Ops, unsigned NumOps);
+
+  /// getNodeIfExists - Get the specified node if it's already available, or
+  /// else return NULL.
+  SDNode *getNodeIfExists(unsigned Opcode, SDVTList VTs,
+                          const SDOperand *Ops, unsigned NumOps);
   
   /// DAGUpdateListener - Clients of various APIs that cause global effects on
   /// the DAG can optionally implement this interface.  This allows the clients
diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 579870e6d07..9feefb9e303 100644
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -758,6 +758,23 @@ SDOperand DAGCombiner::combine(SDNode *N) {
     }
   }
 
+  // If N is a commutative binary node, try commuting it to enable more 
+  // sdisel CSE.
+  if (RV.Val == 0 && 
+      SelectionDAG::isCommutativeBinOp(N->getOpcode()) &&
+      N->getNumValues() == 1) {
+    SDOperand N0 = N->getOperand(0);
+    SDOperand N1 = N->getOperand(1);
+    // Constant operands are canonicalized to RHS.
+    if (isa<ConstantSDNode>(N0) || !isa<ConstantSDNode>(N1)) {
+      SDOperand Ops[] = { N1, N0 };
+      SDNode *CSENode = DAG.getNodeIfExists(N->getOpcode(), N->getVTList(),
+                                            Ops, 2);
+      if (CSENode && CSENode->use_size() <= N->use_size())
+        return SDOperand(CSENode, 0);
+    }
+  }
+
   return RV;
 } 
 
diff --git a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 9331e8b7ee9..94eba83e05d 100644
--- a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -3291,6 +3291,20 @@ SDNode *SelectionDAG::getTargetNode(unsigned Opcode,
                  Ops, NumOps).Val;
 }
 
+/// getNodeIfExists - Get the specified node if it's already available, or
+/// else return NULL.
+SDNode *SelectionDAG::getNodeIfExists(unsigned Opcode, SDVTList VTList,
+                                      const SDOperand *Ops, unsigned NumOps) {
+  if (VTList.VTs[VTList.NumVTs-1] != MVT::Flag) {
+    FoldingSetNodeID ID;
+    AddNodeIDNode(ID, Opcode, VTList, Ops, NumOps);
+    void *IP = 0;
+    if (SDNode *E = CSEMap.FindNodeOrInsertPos(ID, IP))
+      return E;
+  }
+  return NULL;
+}
+
 
 /// ReplaceAllUsesWith - Modify anything using 'From' to use 'To' instead.
 /// This can cause recursive merging of nodes in the DAG.
diff --git a/test/CodeGen/X86/dagcombine-cse.ll b/test/CodeGen/X86/dagcombine-cse.ll
new file mode 100644
index 00000000000..7427678db00
--- /dev/null
+++ b/test/CodeGen/X86/dagcombine-cse.ll
@@ -0,0 +1,27 @@
+; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 -stats |& grep asm-printer | grep 14
+
+define i32 @t(i8* %ref_frame_ptr, i32 %ref_frame_stride, i32 %idxX, i32 %idxY) nounwind  {
+entry:
+	%tmp7 = mul i32 %idxY, %ref_frame_stride		; <i32> [#uses=2]
+	%tmp9 = add i32 %tmp7, %idxX		; <i32> [#uses=1]
+	%tmp11 = getelementptr i8* %ref_frame_ptr, i32 %tmp9		; <i8*> [#uses=1]
+	%tmp1112 = bitcast i8* %tmp11 to i32*		; <i32*> [#uses=1]
+	%tmp13 = load i32* %tmp1112, align 4		; <i32> [#uses=1]
+	%tmp18 = add i32 %idxX, 4		; <i32> [#uses=1]
+	%tmp20.sum = add i32 %tmp18, %tmp7		; <i32> [#uses=1]
+	%tmp21 = getelementptr i8* %ref_frame_ptr, i32 %tmp20.sum		; <i8*> [#uses=1]
+	%tmp2122 = bitcast i8* %tmp21 to i16*		; <i16*> [#uses=1]
+	%tmp23 = load i16* %tmp2122, align 2		; <i16> [#uses=1]
+	%tmp2425 = zext i16 %tmp23 to i64		; <i64> [#uses=1]
+	%tmp26 = shl i64 %tmp2425, 32		; <i64> [#uses=1]
+	%tmp2728 = zext i32 %tmp13 to i64		; <i64> [#uses=1]
+	%tmp29 = or i64 %tmp26, %tmp2728		; <i64> [#uses=1]
+	%tmp3454 = bitcast i64 %tmp29 to double		; <double> [#uses=1]
+	%tmp35 = insertelement <2 x double> undef, double %tmp3454, i32 0		; <<2 x double>> [#uses=1]
+	%tmp36 = insertelement <2 x double> %tmp35, double 0.000000e+00, i32 1		; <<2 x double>> [#uses=1]
+	%tmp42 = bitcast <2 x double> %tmp36 to <8 x i16>		; <<8 x i16>> [#uses=1]
+	%tmp43 = shufflevector <8 x i16> %tmp42, <8 x i16> undef, <8 x i32> < i32 0, i32 1, i32 1, i32 2, i32 4, i32 5, i32 6, i32 7 >		; <<8 x i16>> [#uses=1]
+	%tmp47 = bitcast <8 x i16> %tmp43 to <4 x i32>		; <<4 x i32>> [#uses=1]
+	%tmp48 = extractelement <4 x i32> %tmp47, i32 0		; <i32> [#uses=1]
+	ret i32 %tmp48
+}