Inline asm mult-alt constraint tests.

[oota-llvm.git] / test / CodeGen / X86 / vec_shuffle.ll

diff --git a/test/CodeGen/X86/vec_shuffle.ll b/test/CodeGen/X86/vec_shuffle.ll
index f27a6e870070995c47b6dd72ba7b98e6c3a6bedb..2a48de22098faf34a276fab16c01b5373df37292 100644 (file)
--- a/test/CodeGen/X86/vec_shuffle.ll
+++ b/test/CodeGen/X86/vec_shuffle.ll
@@ -1,68 +1,45 @@
-; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 | grep shufp | wc -l | grep 1
-; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 | grep movhpd | wc -l | grep 1
-; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 | grep pshufhw | wc -l | grep 2
-; RUN: llvm-as < %s | llc -march=x86 -mattr=+sse2 | grep pshuflw | wc -l | grep 1
+; RUN: llc < %s -march=x86 -mcpu=core2 -o %t
+; RUN: grep movq    %t | count 1
+; RUN: grep pshufd  %t | count 1
+; RUN: grep movupd  %t | count 1
+; RUN: grep pshufhw %t | count 1
 
-void %test_v4sf(<4 x float>* %P, float %X, float %Y) {
-       %tmp = insertelement <4 x float> zeroinitializer, float %X, uint 0
-       %tmp2 = insertelement <4 x float> %tmp, float %X, uint 1
-       %tmp4 = insertelement <4 x float> %tmp2, float %Y, uint 2
-       %tmp6 = insertelement <4 x float> %tmp4, float %Y, uint 3
+define void @test_v4sf(<4 x float>* %P, float %X, float %Y) nounwind {
+       %tmp = insertelement <4 x float> zeroinitializer, float %X, i32 0               ; <<4 x float>> [#uses=1]
+       %tmp2 = insertelement <4 x float> %tmp, float %X, i32 1         ; <<4 x float>> [#uses=1]
+       %tmp4 = insertelement <4 x float> %tmp2, float %Y, i32 2                ; <<4 x float>> [#uses=1]
+       %tmp6 = insertelement <4 x float> %tmp4, float %Y, i32 3                ; <<4 x float>> [#uses=1]
        store <4 x float> %tmp6, <4 x float>* %P
        ret void
 }
 
-void %test_v2sd(<2 x double>* %P, double %X, double %Y) {
-       %tmp = insertelement <2 x double> zeroinitializer, double %X, uint 0
-       %tmp2 = insertelement <2 x double> %tmp, double %Y, uint 1
+define void @test_v2sd(<2 x double>* %P, double %X, double %Y) nounwind {
+       %tmp = insertelement <2 x double> zeroinitializer, double %X, i32 0             ; <<2 x double>> [#uses=1]
+       %tmp2 = insertelement <2 x double> %tmp, double %Y, i32 1               ; <<2 x double>> [#uses=1]
        store <2 x double> %tmp2, <2 x double>* %P
        ret void
 }
 
-void %test_v8i16(<2 x long>* %res, <2 x long>* %A) {
-       %tmp = load <2 x long>* %A
-       %tmp = cast <2 x long> %tmp to <8 x short>
-       %tmp = extractelement <8 x short> %tmp, uint 0
-       %tmp1 = extractelement <8 x short> %tmp, uint 1
-       %tmp2 = extractelement <8 x short> %tmp, uint 2
-       %tmp3 = extractelement <8 x short> %tmp, uint 3
-       %tmp4 = extractelement <8 x short> %tmp, uint 6
-       %tmp5 = extractelement <8 x short> %tmp, uint 5
-       %tmp6 = extractelement <8 x short> %tmp, uint 4
-       %tmp7 = extractelement <8 x short> %tmp, uint 7
-       %tmp8 = insertelement <8 x short> undef, short %tmp, uint 0
-       %tmp9 = insertelement <8 x short> %tmp8, short %tmp1, uint 1
-       %tmp10 = insertelement <8 x short> %tmp9, short %tmp2, uint 2
-       %tmp11 = insertelement <8 x short> %tmp10, short %tmp3, uint 3
-       %tmp12 = insertelement <8 x short> %tmp11, short %tmp4, uint 4
-       %tmp13 = insertelement <8 x short> %tmp12, short %tmp5, uint 5
-       %tmp14 = insertelement <8 x short> %tmp13, short %tmp6, uint 6
-       %tmp15 = insertelement <8 x short> %tmp14, short %tmp7, uint 7
-       %tmp15 = cast <8 x short> %tmp15 to <2 x long>
-       store <2 x long> %tmp15, <2 x long>* %res
-       ret void
-}
-
-void %test_v8i16_2(<2 x long>* %res, <2 x long>* %A) {
-       %tmp = load <2 x long>* %A
-       %tmp = cast <2 x long> %tmp to <8 x short>
-       %tmp0 = extractelement <8 x short> %tmp, uint 0
-       %tmp1 = extractelement <8 x short> %tmp, uint 1
-       %tmp2 = extractelement <8 x short> %tmp, uint 2
-       %tmp3 = extractelement <8 x short> %tmp, uint 3
-       %tmp4 = extractelement <8 x short> %tmp, uint 4
-       %tmp5 = extractelement <8 x short> %tmp, uint 5
-       %tmp6 = extractelement <8 x short> %tmp, uint 6
-       %tmp7 = extractelement <8 x short> %tmp, uint 7
-       %tmp8 = insertelement <8 x short> undef, short %tmp2, uint 0
-       %tmp9 = insertelement <8 x short> %tmp8, short %tmp1, uint 1
-       %tmp10 = insertelement <8 x short> %tmp9, short %tmp0, uint 2
-       %tmp11 = insertelement <8 x short> %tmp10, short %tmp3, uint 3
-       %tmp12 = insertelement <8 x short> %tmp11, short %tmp6, uint 4
-       %tmp13 = insertelement <8 x short> %tmp12, short %tmp5, uint 5
-       %tmp14 = insertelement <8 x short> %tmp13, short %tmp4, uint 6
-       %tmp15 = insertelement <8 x short> %tmp14, short %tmp7, uint 7
-       %tmp15 = cast <8 x short> %tmp15 to <2 x long>
-       store <2 x long> %tmp15, <2 x long>* %res
+define void @test_v8i16(<2 x i64>* %res, <2 x i64>* %A) nounwind {
+       %tmp = load <2 x i64>* %A               ; <<2 x i64>> [#uses=1]
+       %tmp.upgrd.1 = bitcast <2 x i64> %tmp to <8 x i16>              ; <<8 x i16>> [#uses=8]
+       %tmp.upgrd.2 = extractelement <8 x i16> %tmp.upgrd.1, i32 0             ; <i16> [#uses=1]
+       %tmp1 = extractelement <8 x i16> %tmp.upgrd.1, i32 1            ; <i16> [#uses=1]
+       %tmp2 = extractelement <8 x i16> %tmp.upgrd.1, i32 2            ; <i16> [#uses=1]
+       %tmp3 = extractelement <8 x i16> %tmp.upgrd.1, i32 3            ; <i16> [#uses=1]
+       %tmp4 = extractelement <8 x i16> %tmp.upgrd.1, i32 6            ; <i16> [#uses=1]
+       %tmp5 = extractelement <8 x i16> %tmp.upgrd.1, i32 5            ; <i16> [#uses=1]
+       %tmp6 = extractelement <8 x i16> %tmp.upgrd.1, i32 4            ; <i16> [#uses=1]
+       %tmp7 = extractelement <8 x i16> %tmp.upgrd.1, i32 7            ; <i16> [#uses=1]
+       %tmp8 = insertelement <8 x i16> undef, i16 %tmp.upgrd.2, i32 0          ; <<8 x i16>> [#uses=1]
+       %tmp9 = insertelement <8 x i16> %tmp8, i16 %tmp1, i32 1         ; <<8 x i16>> [#uses=1]
+       %tmp10 = insertelement <8 x i16> %tmp9, i16 %tmp2, i32 2                ; <<8 x i16>> [#uses=1]
+       %tmp11 = insertelement <8 x i16> %tmp10, i16 %tmp3, i32 3               ; <<8 x i16>> [#uses=1]
+       %tmp12 = insertelement <8 x i16> %tmp11, i16 %tmp4, i32 4               ; <<8 x i16>> [#uses=1]
+       %tmp13 = insertelement <8 x i16> %tmp12, i16 %tmp5, i32 5               ; <<8 x i16>> [#uses=1]
+       %tmp14 = insertelement <8 x i16> %tmp13, i16 %tmp6, i32 6               ; <<8 x i16>> [#uses=1]
+       %tmp15 = insertelement <8 x i16> %tmp14, i16 %tmp7, i32 7               ; <<8 x i16>> [#uses=1]
+       %tmp15.upgrd.3 = bitcast <8 x i16> %tmp15 to <2 x i64>          ; <<2 x i64>> [#uses=1]
+       store <2 x i64> %tmp15.upgrd.3, <2 x i64>* %res
        ret void
 }