Index: llvm/trunk/lib/Target/X86/X86InstrSSE.td
===================================================================
--- llvm/trunk/lib/Target/X86/X86InstrSSE.td
+++ llvm/trunk/lib/Target/X86/X86InstrSSE.td
@@ -241,21 +241,20 @@
 /// sse12_fp_scalar - SSE 1 & 2 scalar instructions class
 multiclass sse12_fp_scalar<bits<8> opc, string OpcodeStr, SDNode OpNode,
                            RegisterClass RC, X86MemOperand x86memop,
-                           OpndItins itins,
-                           bit Is2Addr = 1> {
+                           Domain d, OpndItins itins, bit Is2Addr = 1> {
   let isCommutable = 1 in {
     def rr : SI<opc, MRMSrcReg, (outs RC:$dst), (ins RC:$src1, RC:$src2),
        !if(Is2Addr,
            !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
            !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}")),
-       [(set RC:$dst, (OpNode RC:$src1, RC:$src2))], itins.rr>,
+       [(set RC:$dst, (OpNode RC:$src1, RC:$src2))], itins.rr, d>,
        Sched<[itins.Sched]>;
   }
   def rm : SI<opc, MRMSrcMem, (outs RC:$dst), (ins RC:$src1, x86memop:$src2),
        !if(Is2Addr,
            !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
            !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}")),
-       [(set RC:$dst, (OpNode RC:$src1, (load addr:$src2)))], itins.rm>,
+       [(set RC:$dst, (OpNode RC:$src1, (load addr:$src2)))], itins.rm, d>,
        Sched<[itins.Sched.Folded, ReadAfterLd]>;
 }
 
@@ -263,8 +262,7 @@
 multiclass sse12_fp_scalar_int<bits<8> opc, string OpcodeStr, RegisterClass RC,
                              string asm, string SSEVer, string FPSizeStr,
                              Operand memopr, ComplexPattern mem_cpat,
-                             OpndItins itins,
-                             bit Is2Addr = 1> {
+                             Domain d, OpndItins itins, bit Is2Addr = 1> {
 let isCodeGenOnly = 1 in {
   def rr_Int : SI<opc, MRMSrcReg, (outs RC:$dst), (ins RC:$src1, RC:$src2),
        !if(Is2Addr,
@@ -272,7 +270,7 @@
            !strconcat(asm, "\t{$src2, $src1, $dst|$dst, $src1, $src2}")),
        [(set RC:$dst, (!cast<Intrinsic>(
                  !strconcat("int_x86_sse", SSEVer, "_", OpcodeStr, FPSizeStr))
-             RC:$src1, RC:$src2))], itins.rr>,
+             RC:$src1, RC:$src2))], itins.rr, d>,
        Sched<[itins.Sched]>;
   def rm_Int : SI<opc, MRMSrcMem, (outs RC:$dst), (ins RC:$src1, memopr:$src2),
        !if(Is2Addr,
@@ -280,7 +278,7 @@
            !strconcat(asm, "\t{$src2, $src1, $dst|$dst, $src1, $src2}")),
        [(set RC:$dst, (!cast<Intrinsic>(!strconcat("int_x86_sse",
                                           SSEVer, "_", OpcodeStr, FPSizeStr))
-             RC:$src1, mem_cpat:$src2))], itins.rm>,
+             RC:$src1, mem_cpat:$src2))], itins.rm, d>,
        Sched<[itins.Sched.Folded, ReadAfterLd]>;
 }
 }
@@ -3054,15 +3052,19 @@
 multiclass basic_sse12_fp_binop_s<bits<8> opc, string OpcodeStr, SDNode OpNode,
                                   SizeItins itins> {
   defm V#NAME#SS : sse12_fp_scalar<opc, !strconcat(OpcodeStr, "ss"),
-                         OpNode, FR32, f32mem, itins.s, 0>, XS, VEX_4V, VEX_LIG;
+                         OpNode, FR32, f32mem, SSEPackedSingle, itins.s, 0>,
+                         XS, VEX_4V, VEX_LIG;
   defm V#NAME#SD : sse12_fp_scalar<opc, !strconcat(OpcodeStr, "sd"),
-                         OpNode, FR64, f64mem, itins.d, 0>, XD, VEX_4V, VEX_LIG;
+                         OpNode, FR64, f64mem, SSEPackedDouble, itins.d, 0>,
+                         XD, VEX_4V, VEX_LIG;
 
   let Constraints = "$src1 = $dst" in {
     defm SS : sse12_fp_scalar<opc, !strconcat(OpcodeStr, "ss"),
-                              OpNode, FR32, f32mem, itins.s>, XS;
+                              OpNode, FR32, f32mem, SSEPackedSingle,
+                              itins.s>, XS;
     defm SD : sse12_fp_scalar<opc, !strconcat(OpcodeStr, "sd"),
-                              OpNode, FR64, f64mem, itins.d>, XD;
+                              OpNode, FR64, f64mem, SSEPackedDouble,
+                              itins.d>, XD;
   }
 }
 
@@ -3070,18 +3072,18 @@
                                       SizeItins itins> {
   defm V#NAME#SS : sse12_fp_scalar_int<opc, OpcodeStr, VR128,
                    !strconcat(OpcodeStr, "ss"), "", "_ss", ssmem, sse_load_f32,
-                   itins.s, 0>, XS, VEX_4V, VEX_LIG;
+                   SSEPackedSingle, itins.s, 0>, XS, VEX_4V, VEX_LIG;
   defm V#NAME#SD : sse12_fp_scalar_int<opc, OpcodeStr, VR128,
                    !strconcat(OpcodeStr, "sd"), "2", "_sd", sdmem, sse_load_f64,
-                   itins.d, 0>, XD, VEX_4V, VEX_LIG;
+                   SSEPackedDouble, itins.d, 0>, XD, VEX_4V, VEX_LIG;
 
   let Constraints = "$src1 = $dst" in {
     defm SS : sse12_fp_scalar_int<opc, OpcodeStr, VR128,
                    !strconcat(OpcodeStr, "ss"), "", "_ss", ssmem, sse_load_f32,
-                   itins.s>, XS;
+                   SSEPackedSingle, itins.s>, XS;
     defm SD : sse12_fp_scalar_int<opc, OpcodeStr, VR128,
                    !strconcat(OpcodeStr, "sd"), "2", "_sd", sdmem, sse_load_f64,
-                   itins.d>, XD;
+                   SSEPackedDouble, itins.d>, XD;
   }
 }
 
@@ -3170,7 +3172,7 @@
           (Op (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
       (!cast<I>(OpcPrefix#SSrr_Int) v4f32:$dst, v4f32:$src)>;
   }
-  
+
   // With SSE 4.1, insertps/blendi are preferred to movsd, so match those too.
   let Predicates = [UseSSE41] in {
     // extracted scalar math op with insert via insertps
@@ -3203,7 +3205,7 @@
           FR32:$src))), (iPTR 0))),
       (!cast<I>("V"#OpcPrefix#SSrr_Int) v4f32:$dst,
           (COPY_TO_REGCLASS FR32:$src, VR128))>;
- 
+
     // extracted scalar math op with insert via blend
     def : Pat<(v4f32 (X86Blendi (v4f32 VR128:$dst), (v4f32 (scalar_to_vector
           (Op (f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
@@ -3251,7 +3253,7 @@
           FR64:$src))), (i8 1))),
       (!cast<I>(OpcPrefix#SDrr_Int) v2f64:$dst,
           (COPY_TO_REGCLASS FR64:$src, VR128))>;
-          
+
     // vector math op with insert via blend
     def : Pat<(v2f64 (X86Blendi (v2f64 VR128:$dst),
           (Op (v2f64 VR128:$dst), (v2f64 VR128:$src)), (i8 1))),
@@ -3345,17 +3347,17 @@
                           ValueType vt, ValueType ScalarVT,
                           X86MemOperand x86memop, Operand vec_memop,
                           ComplexPattern mem_cpat, Intrinsic Intr,
-                          SDNode OpNode, OpndItins itins, Predicate target,
-                          string Suffix> {
+                          SDNode OpNode, Domain d, OpndItins itins,
+                          Predicate target, string Suffix> {
   let hasSideEffects = 0 in {
   def r : I<opc, MRMSrcReg, (outs RC:$dst), (ins RC:$src1),
               !strconcat(OpcodeStr, "\t{$src1, $dst|$dst, $src1}"),
-            [(set RC:$dst, (OpNode RC:$src1))], itins.rr>, Sched<[itins.Sched]>,
+            [(set RC:$dst, (OpNode RC:$src1))], itins.rr, d>, Sched<[itins.Sched]>,
             Requires<[target]>;
   let mayLoad = 1 in
   def m : I<opc, MRMSrcMem, (outs RC:$dst), (ins x86memop:$src1),
             !strconcat(OpcodeStr, "\t{$src1, $dst|$dst, $src1}"),
-            [(set RC:$dst, (OpNode (load addr:$src1)))], itins.rm>,
+            [(set RC:$dst, (OpNode (load addr:$src1)))], itins.rm, d>,
             Sched<[itins.Sched.Folded, ReadAfterLd]>,
             Requires<[target, OptForSize]>;
 
@@ -3378,7 +3380,7 @@
   // because the high elements of the destination are unchanged in SSE.
   def : Pat<(Intr VR128:$src),
             (!cast<Instruction>(NAME#Suffix##r_Int) VR128:$src, VR128:$src)>;
-  def : Pat<(Intr (load addr:$src)), 
+  def : Pat<(Intr (load addr:$src)),
             (vt (COPY_TO_REGCLASS(!cast<Instruction>(NAME#Suffix##m)
                                       addr:$src), VR128))>;
    def : Pat<(Intr mem_cpat:$src),
@@ -3391,24 +3393,24 @@
                           ValueType vt, ValueType ScalarVT,
                           X86MemOperand x86memop, Operand vec_memop,
                           ComplexPattern mem_cpat,
-                          Intrinsic Intr, SDNode OpNode, OpndItins itins,
-                          Predicate target, string Suffix> {
+                          Intrinsic Intr, SDNode OpNode, Domain d,
+                          OpndItins itins, Predicate target, string Suffix> {
   let hasSideEffects = 0 in {
   def r : I<opc, MRMSrcReg, (outs RC:$dst), (ins RC:$src1, RC:$src2),
             !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
-            [], itins.rr>, Sched<[itins.Sched]>;
-  let mayLoad = 1 in 
+            [], itins.rr, d>, Sched<[itins.Sched]>;
+  let mayLoad = 1 in
   def m : I<opc, MRMSrcMem, (outs RC:$dst), (ins RC:$src1, x86memop:$src2),
              !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
-            [], itins.rm>, Sched<[itins.Sched.Folded, ReadAfterLd]>;
+            [], itins.rm, d>, Sched<[itins.Sched.Folded, ReadAfterLd]>;
   let isCodeGenOnly = 1 in {
   // todo: uncomment when all r_Int forms will be added to X86InstrInfo.cpp
-  //def r_Int : I<opc, MRMSrcReg, (outs VR128:$dst), 
+  //def r_Int : I<opc, MRMSrcReg, (outs VR128:$dst),
   //              (ins VR128:$src1, VR128:$src2),
   //           !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
   //          []>, Sched<[itins.Sched.Folded]>;
   let mayLoad = 1 in
-  def m_Int : I<opc, MRMSrcMem, (outs VR128:$dst), 
+  def m_Int : I<opc, MRMSrcMem, (outs VR128:$dst),
                 (ins VR128:$src1, vec_memop:$src2),
              !strconcat(OpcodeStr, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"),
             []>, Sched<[itins.Sched.Folded, ReadAfterLd]>;
@@ -3419,7 +3421,7 @@
    def : Pat<(OpNode RC:$src),  (!cast<Instruction>("V"#NAME#Suffix##r)
                                 (ScalarVT (IMPLICIT_DEF)), RC:$src)>;
 
-   def : Pat<(vt (OpNode mem_cpat:$src)), 
+   def : Pat<(vt (OpNode mem_cpat:$src)),
              (!cast<Instruction>("V"#NAME#Suffix##m_Int) (vt (IMPLICIT_DEF)),
                                   mem_cpat:$src)>;
 
@@ -3428,14 +3430,14 @@
    //                 (VT (IMPLICIT_DEF)), VR128:$src)>;
    def : Pat<(Intr VR128:$src),
              (vt (COPY_TO_REGCLASS(
-             !cast<Instruction>("V"#NAME#Suffix##r) (ScalarVT (IMPLICIT_DEF)), 
+             !cast<Instruction>("V"#NAME#Suffix##r) (ScalarVT (IMPLICIT_DEF)),
                     (ScalarVT (COPY_TO_REGCLASS VR128:$src, RC))), VR128))>;
    def : Pat<(Intr mem_cpat:$src),
              (!cast<Instruction>("V"#NAME#Suffix##m_Int)
                     (vt (IMPLICIT_DEF)), mem_cpat:$src)>;
   }
   let Predicates = [target, OptForSize] in
-  def : Pat<(ScalarVT (OpNode (load addr:$src))), 
+  def : Pat<(ScalarVT (OpNode (load addr:$src))),
             (!cast<Instruction>("V"#NAME#Suffix##m) (ScalarVT (IMPLICIT_DEF)),
              addr:$src)>;
 }
@@ -3557,11 +3559,11 @@
   defm SS        :  sse_fp_unop_s<opc, OpcodeStr##ss, FR32, v4f32, f32, f32mem,
                       ssmem, sse_load_f32,
                       !cast<Intrinsic>("int_x86_sse_"##OpcodeStr##_ss), OpNode,
-                      itins, UseSSE1, "SS">, XS;
+                      SSEPackedSingle, itins, UseSSE1, "SS">, XS;
   defm V#NAME#SS  : avx_fp_unop_s<opc, "v"#OpcodeStr##ss, FR32, v4f32, f32,
                       f32mem, ssmem, sse_load_f32,
                       !cast<Intrinsic>("int_x86_sse_"##OpcodeStr##_ss), OpNode,
-                      itins, UseAVX, "SS">, XS, VEX_4V, VEX_LIG;
+                      SSEPackedSingle, itins, UseAVX, "SS">, XS, VEX_4V, VEX_LIG;
 }
 
 multiclass sse2_fp_unop_s<bits<8> opc, string OpcodeStr, SDNode OpNode,
@@ -3569,11 +3571,12 @@
   defm SD         : sse_fp_unop_s<opc, OpcodeStr##sd, FR64, v2f64, f64, f64mem,
                          sdmem, sse_load_f64,
                          !cast<Intrinsic>("int_x86_sse2_"##OpcodeStr##_sd),
-                         OpNode, itins, UseSSE2, "SD">, XD;
+                         OpNode, SSEPackedDouble, itins, UseSSE2, "SD">, XD;
   defm V#NAME#SD  : avx_fp_unop_s<opc, "v"#OpcodeStr##sd, FR64, v2f64, f64,
                          f64mem, sdmem, sse_load_f64,
                          !cast<Intrinsic>("int_x86_sse2_"##OpcodeStr##_sd),
-                         OpNode, itins, UseAVX, "SD">, XD, VEX_4V, VEX_LIG;
+                         OpNode, SSEPackedDouble, itins, UseAVX, "SD">,
+                         XD, VEX_4V, VEX_LIG;
 }
 
 // Square root.
Index: llvm/trunk/test/CodeGen/X86/sink-hoist.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/sink-hoist.ll
+++ llvm/trunk/test/CodeGen/X86/sink-hoist.ll
@@ -9,7 +9,7 @@
 ; CHECK-NEXT: testb $1, %dil
 ; CHECK-NEXT: jne
 ; CHECK-NEXT: divsd
-; CHECK-NEXT: movaps
+; CHECK-NEXT: movapd
 ; CHECK-NEXT: ret
 ; CHECK:      divsd
 
@@ -28,7 +28,7 @@
 ; CHECK-NEXT: testb $1, %dil
 ; CHECK-NEXT: je
 ; CHECK:      divsd
-; CHECK:      movaps
+; CHECK:      movapd
 ; CHECK:      ret
 define double @split(double %x, double %y, i1 %c) nounwind {
   %a = fdiv double %x, 3.2
Index: llvm/trunk/test/CodeGen/X86/sse-minmax.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/sse-minmax.ll
+++ llvm/trunk/test/CodeGen/X86/sse-minmax.ll
@@ -805,7 +805,7 @@
 ; CHECK-LABEL: clampTo3k_a:
 ; CHECK-NEXT: movsd {{[^,]*}}, %xmm1
 ; CHECK-NEXT: minsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: movapd %xmm1, %xmm0
 ; CHECK-NEXT: ret
 ; UNSAFE-LABEL: clampTo3k_a:
 ; UNSAFE-NEXT: minsd {{[^,]*}}, %xmm0
@@ -813,7 +813,7 @@
 ; FINITE-LABEL: clampTo3k_a:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: minsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_a(double %x) nounwind readnone {
 entry:
@@ -831,7 +831,7 @@
 ; FINITE-LABEL: clampTo3k_b:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: minsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_b(double %x) nounwind readnone {
 entry:
@@ -843,7 +843,7 @@
 ; CHECK-LABEL: clampTo3k_c:
 ; CHECK-NEXT: movsd {{[^,]*}}, %xmm1
 ; CHECK-NEXT: maxsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: movapd %xmm1, %xmm0
 ; CHECK-NEXT: ret
 ; UNSAFE-LABEL: clampTo3k_c:
 ; UNSAFE-NEXT: maxsd {{[^,]*}}, %xmm0
@@ -851,7 +851,7 @@
 ; FINITE-LABEL: clampTo3k_c:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: maxsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_c(double %x) nounwind readnone {
 entry:
@@ -869,7 +869,7 @@
 ; FINITE-LABEL: clampTo3k_d:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: maxsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_d(double %x) nounwind readnone {
 entry:
@@ -881,7 +881,7 @@
 ; CHECK-LABEL: clampTo3k_e:
 ; CHECK-NEXT: movsd {{[^,]*}}, %xmm1
 ; CHECK-NEXT: maxsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: movapd %xmm1, %xmm0
 ; CHECK-NEXT: ret
 ; UNSAFE-LABEL: clampTo3k_e:
 ; UNSAFE-NEXT: maxsd {{[^,]*}}, %xmm0
@@ -889,7 +889,7 @@
 ; FINITE-LABEL: clampTo3k_e:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: maxsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_e(double %x) nounwind readnone {
 entry:
@@ -907,7 +907,7 @@
 ; FINITE-LABEL: clampTo3k_f:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: maxsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_f(double %x) nounwind readnone {
 entry:
@@ -919,7 +919,7 @@
 ; CHECK-LABEL: clampTo3k_g:
 ; CHECK-NEXT: movsd {{[^,]*}}, %xmm1
 ; CHECK-NEXT: minsd %xmm0, %xmm1
-; CHECK-NEXT: movaps %xmm1, %xmm0
+; CHECK-NEXT: movapd %xmm1, %xmm0
 ; CHECK-NEXT: ret
 ; UNSAFE-LABEL: clampTo3k_g:
 ; UNSAFE-NEXT: minsd {{[^,]*}}, %xmm0
@@ -927,7 +927,7 @@
 ; FINITE-LABEL: clampTo3k_g:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: minsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_g(double %x) nounwind readnone {
 entry:
@@ -945,7 +945,7 @@
 ; FINITE-LABEL: clampTo3k_h:
 ; FINITE-NEXT: movsd {{[^,]*}}, %xmm1
 ; FINITE-NEXT: minsd %xmm0, %xmm1
-; FINITE-NEXT: movaps %xmm1, %xmm0
+; FINITE-NEXT: movapd %xmm1, %xmm0
 ; FINITE-NEXT: ret
 define double @clampTo3k_h(double %x) nounwind readnone {
 entry:
Index: llvm/trunk/test/CodeGen/X86/sse-scalar-fp-arith.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/sse-scalar-fp-arith.ll
+++ llvm/trunk/test/CodeGen/X86/sse-scalar-fp-arith.ll
@@ -76,6 +76,31 @@
   ret <4 x float> %3
 }
 
+define <4 x float> @test_sqrt_ss(<4 x float> %a) {
+; SSE2-LABEL: test_sqrt_ss:
+; SSE2:       # BB#0:
+; SSE2-NEXT:   sqrtss %xmm0, %xmm1
+; SSE2-NEXT:   movss %xmm1, %xmm0
+; SSE2-NEXT:   retq
+;
+; SSE41-LABEL: test_sqrt_ss:
+; SSE41:       # BB#0:
+; SSE41-NEXT:  sqrtss %xmm0, %xmm1
+; SSE41-NEXT:  blendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
+; SSE41-NEXT:  retq
+;
+; AVX-LABEL: test_sqrt_ss:
+; AVX:       # BB#0:
+; AVX-NEXT:    vsqrtss %xmm0, %xmm0, %xmm1
+; AVX-NEXT:    vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
+; AVX-NEXT:    retq
+  %1 = extractelement <4 x float> %a, i32 0
+  %2 = call float @llvm.sqrt.f32(float %1)
+  %3 = insertelement <4 x float> %a, float %2, i32 0
+  ret <4 x float> %3
+}
+declare float @llvm.sqrt.f32(float)
+
 define <2 x double> @test_add_sd(<2 x double> %a, <2 x double> %b) {
 ; SSE-LABEL: test_add_sd:
 ; SSE:       # BB#0:
@@ -144,6 +169,25 @@
   ret <2 x double> %3
 }
 
+define <2 x double> @test_sqrt_sd(<2 x double> %a) {
+; SSE-LABEL: test_sqrt_sd:
+; SSE:       # BB#0:
+; SSE-NEXT:    sqrtsd %xmm0, %xmm1
+; SSE-NEXT:    movsd %xmm1, %xmm0
+; SSE-NEXT:    retq
+;
+; AVX-LABEL: test_sqrt_sd:
+; AVX:       # BB#0:
+; AVX-NEXT:    vsqrtsd %xmm0, %xmm0, %xmm1
+; AVX-NEXT:    vmovsd %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    retq
+  %1 = extractelement <2 x double> %a, i32 0
+  %2 = call double @llvm.sqrt.f64(double %1)
+  %3 = insertelement <2 x double> %a, double %2, i32 0
+  ret <2 x double> %3
+}
+declare double @llvm.sqrt.f64(double)
+
 define <4 x float> @test2_add_ss(<4 x float> %a, <4 x float> %b) {
 ; SSE-LABEL: test2_add_ss:
 ; SSE:       # BB#0:
@@ -220,7 +264,7 @@
 ; SSE-LABEL: test2_add_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    addsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test2_add_sd:
@@ -238,7 +282,7 @@
 ; SSE-LABEL: test2_sub_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    subsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test2_sub_sd:
@@ -256,7 +300,7 @@
 ; SSE-LABEL: test2_mul_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    mulsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test2_mul_sd:
@@ -274,7 +318,7 @@
 ; SSE-LABEL: test2_div_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    divsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: test2_div_sd:
@@ -371,7 +415,7 @@
 }
 
 ; With SSE4.1 or greater, the shuffles in the following tests may
-; be lowered to X86Blendi nodes. 
+; be lowered to X86Blendi nodes.
 
 define <4 x float> @blend_add_ss(<4 x float> %a, float %b) {
 ; SSE-LABEL: blend_add_ss:
@@ -708,7 +752,7 @@
 ; SSE-LABEL: insert_test2_add_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    addsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test2_add_sd:
@@ -724,7 +768,7 @@
 ; SSE-LABEL: insert_test2_sub_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    subsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test2_sub_sd:
@@ -740,7 +784,7 @@
 ; SSE-LABEL: insert_test2_mul_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    mulsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test2_mul_sd:
@@ -756,7 +800,7 @@
 ; SSE-LABEL: insert_test2_div_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    divsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test2_div_sd:
@@ -956,7 +1000,7 @@
 ; SSE-LABEL: insert_test4_add_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    addsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test4_add_sd:
@@ -972,7 +1016,7 @@
 ; SSE-LABEL: insert_test4_sub_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    subsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test4_sub_sd:
@@ -988,7 +1032,7 @@
 ; SSE-LABEL: insert_test4_mul_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    mulsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test4_mul_sd:
@@ -1004,7 +1048,7 @@
 ; SSE-LABEL: insert_test4_div_sd:
 ; SSE:       # BB#0:
 ; SSE-NEXT:    divsd %xmm0, %xmm1
-; SSE-NEXT:    movaps %xmm1, %xmm0
+; SSE-NEXT:    movapd %xmm1, %xmm0
 ; SSE-NEXT:    retq
 ;
 ; AVX-LABEL: insert_test4_div_sd: