X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=test%2FCodeGen%2FARM%2Fvadd.ll;h=86b0d029701830436c9dc92da094f5e0f9dbbfbf;hb=e62fbca6b7dbb40ed453cc6fb002ef6be7f5c3e6;hp=a830e968ff7885a3f443b9a950181ce8b635eb29;hpb=103b4a571ef01e4717c3c6d9db6506a3abd6cc0b;p=oota-llvm.git

diff --git a/test/CodeGen/ARM/vadd.ll b/test/CodeGen/ARM/vadd.ll
index a830e968ff7..86b0d029701 100644
--- a/test/CodeGen/ARM/vadd.ll
+++ b/test/CodeGen/ARM/vadd.ll
@@ -1,7 +1,7 @@
-; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
+; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
 
 define <8 x i8> @vaddi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
-;CHECK: vaddi8:
+;CHECK-LABEL: vaddi8:
 ;CHECK: vadd.i8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
@@ -10,7 +10,7 @@ define <8 x i8> @vaddi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vaddi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
-;CHECK: vaddi16:
+;CHECK-LABEL: vaddi16:
 ;CHECK: vadd.i16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
@@ -19,7 +19,7 @@ define <4 x i16> @vaddi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vaddi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
-;CHECK: vaddi32:
+;CHECK-LABEL: vaddi32:
 ;CHECK: vadd.i32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
@@ -28,7 +28,7 @@ define <2 x i32> @vaddi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <1 x i64> @vaddi64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
-;CHECK: vaddi64:
+;CHECK-LABEL: vaddi64:
 ;CHECK: vadd.i64
 	%tmp1 = load <1 x i64>* %A
 	%tmp2 = load <1 x i64>* %B
@@ -37,7 +37,7 @@ define <1 x i64> @vaddi64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
 }
 
 define <2 x float> @vaddf32(<2 x float>* %A, <2 x float>* %B) nounwind {
-;CHECK: vaddf32:
+;CHECK-LABEL: vaddf32:
 ;CHECK: vadd.f32
 	%tmp1 = load <2 x float>* %A
 	%tmp2 = load <2 x float>* %B
@@ -46,7 +46,7 @@ define <2 x float> @vaddf32(<2 x float>* %A, <2 x float>* %B) nounwind {
 }
 
 define <16 x i8> @vaddQi8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
-;CHECK: vaddQi8:
+;CHECK-LABEL: vaddQi8:
 ;CHECK: vadd.i8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = load <16 x i8>* %B
@@ -55,7 +55,7 @@ define <16 x i8> @vaddQi8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vaddQi16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
-;CHECK: vaddQi16:
+;CHECK-LABEL: vaddQi16:
 ;CHECK: vadd.i16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
@@ -64,7 +64,7 @@ define <8 x i16> @vaddQi16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vaddQi32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
-;CHECK: vaddQi32:
+;CHECK-LABEL: vaddQi32:
 ;CHECK: vadd.i32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
@@ -73,7 +73,7 @@ define <4 x i32> @vaddQi32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i64> @vaddQi64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
-;CHECK: vaddQi64:
+;CHECK-LABEL: vaddQi64:
 ;CHECK: vadd.i64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
@@ -82,7 +82,7 @@ define <2 x i64> @vaddQi64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
 }
 
 define <4 x float> @vaddQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
-;CHECK: vaddQf32:
+;CHECK-LABEL: vaddQf32:
 ;CHECK: vadd.f32
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = load <4 x float>* %B
@@ -90,39 +90,8 @@ define <4 x float> @vaddQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
 	ret <4 x float> %tmp3
 }
 
-define <8 x i8> @vaddhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
-;CHECK: vaddhni16:
-;CHECK: vaddhn.i16
-	%tmp1 = load <8 x i16>* %A
-	%tmp2 = load <8 x i16>* %B
-	%tmp3 = call <8 x i8> @llvm.arm.neon.vaddhn.v8i8(<8 x i16> %tmp1, <8 x i16> %tmp2)
-	ret <8 x i8> %tmp3
-}
-
-define <4 x i16> @vaddhni32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
-;CHECK: vaddhni32:
-;CHECK: vaddhn.i32
-	%tmp1 = load <4 x i32>* %A
-	%tmp2 = load <4 x i32>* %B
-	%tmp3 = call <4 x i16> @llvm.arm.neon.vaddhn.v4i16(<4 x i32> %tmp1, <4 x i32> %tmp2)
-	ret <4 x i16> %tmp3
-}
-
-define <2 x i32> @vaddhni64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
-;CHECK: vaddhni64:
-;CHECK: vaddhn.i64
-	%tmp1 = load <2 x i64>* %A
-	%tmp2 = load <2 x i64>* %B
-	%tmp3 = call <2 x i32> @llvm.arm.neon.vaddhn.v2i32(<2 x i64> %tmp1, <2 x i64> %tmp2)
-	ret <2 x i32> %tmp3
-}
-
-declare <8 x i8>  @llvm.arm.neon.vaddhn.v8i8(<8 x i16>, <8 x i16>) nounwind readnone
-declare <4 x i16> @llvm.arm.neon.vaddhn.v4i16(<4 x i32>, <4 x i32>) nounwind readnone
-declare <2 x i32> @llvm.arm.neon.vaddhn.v2i32(<2 x i64>, <2 x i64>) nounwind readnone
-
 define <8 x i8> @vraddhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
-;CHECK: vraddhni16:
+;CHECK-LABEL: vraddhni16:
 ;CHECK: vraddhn.i16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
@@ -131,7 +100,7 @@ define <8 x i8> @vraddhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i16> @vraddhni32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
-;CHECK: vraddhni32:
+;CHECK-LABEL: vraddhni32:
 ;CHECK: vraddhn.i32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
@@ -140,7 +109,7 @@ define <4 x i16> @vraddhni32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i32> @vraddhni64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
-;CHECK: vraddhni64:
+;CHECK-LABEL: vraddhni64:
 ;CHECK: vraddhn.i64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
@@ -152,8 +121,35 @@ declare <8 x i8>  @llvm.arm.neon.vraddhn.v8i8(<8 x i16>, <8 x i16>) nounwind rea
 declare <4 x i16> @llvm.arm.neon.vraddhn.v4i16(<4 x i32>, <4 x i32>) nounwind readnone
 declare <2 x i32> @llvm.arm.neon.vraddhn.v2i32(<2 x i64>, <2 x i64>) nounwind readnone
 
+define <8 x i8> @vaddhni16_natural(<8 x i16> %A, <8 x i16> %B) nounwind {
+; CHECK-LABEL: vaddhni16_natural:
+; CHECK: vaddhn.i16
+  %sum = add <8 x i16> %A, %B
+  %shift = lshr <8 x i16> %sum, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
+  %trunc = trunc <8 x i16> %shift to <8 x i8>
+  ret <8 x i8> %trunc
+}
+
+define <4 x i16> @vaddhni32_natural(<4 x i32> %A, <4 x i32> %B) nounwind {
+; CHECK-LABEL: vaddhni32_natural:
+; CHECK: vaddhn.i32
+  %sum = add <4 x i32> %A, %B
+  %shift = lshr <4 x i32> %sum, <i32 16, i32 16, i32 16, i32 16>
+  %trunc = trunc <4 x i32> %shift to <4 x i16>
+  ret <4 x i16> %trunc
+}
+
+define <2 x i32> @vaddhni64_natural(<2 x i64> %A, <2 x i64> %B) nounwind {
+; CHECK-LABEL: vaddhni64_natural:
+; CHECK: vaddhn.i64
+  %sum = add <2 x i64> %A, %B
+  %shift = lshr <2 x i64> %sum, <i64 32, i64 32>
+  %trunc = trunc <2 x i64> %shift to <2 x i32>
+  ret <2 x i32> %trunc
+}
+
 define <8 x i16> @vaddls8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
-;CHECK: vaddls8:
+;CHECK-LABEL: vaddls8:
 ;CHECK: vaddl.s8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
@@ -164,7 +160,7 @@ define <8 x i16> @vaddls8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i32> @vaddls16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
-;CHECK: vaddls16:
+;CHECK-LABEL: vaddls16:
 ;CHECK: vaddl.s16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
@@ -175,7 +171,7 @@ define <4 x i32> @vaddls16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i64> @vaddls32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
-;CHECK: vaddls32:
+;CHECK-LABEL: vaddls32:
 ;CHECK: vaddl.s32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
@@ -186,7 +182,7 @@ define <2 x i64> @vaddls32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <8 x i16> @vaddlu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
-;CHECK: vaddlu8:
+;CHECK-LABEL: vaddlu8:
 ;CHECK: vaddl.u8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
@@ -197,7 +193,7 @@ define <8 x i16> @vaddlu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i32> @vaddlu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
-;CHECK: vaddlu16:
+;CHECK-LABEL: vaddlu16:
 ;CHECK: vaddl.u16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
@@ -208,7 +204,7 @@ define <4 x i32> @vaddlu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i64> @vaddlu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
-;CHECK: vaddlu32:
+;CHECK-LABEL: vaddlu32:
 ;CHECK: vaddl.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
@@ -219,7 +215,7 @@ define <2 x i64> @vaddlu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <8 x i16> @vaddws8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
-;CHECK: vaddws8:
+;CHECK-LABEL: vaddws8:
 ;CHECK: vaddw.s8
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i8>* %B
@@ -229,7 +225,7 @@ define <8 x i16> @vaddws8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i32> @vaddws16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
-;CHECK: vaddws16:
+;CHECK-LABEL: vaddws16:
 ;CHECK: vaddw.s16
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i16>* %B
@@ -239,7 +235,7 @@ define <4 x i32> @vaddws16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i64> @vaddws32(<2 x i64>* %A, <2 x i32>* %B) nounwind {
-;CHECK: vaddws32:
+;CHECK-LABEL: vaddws32:
 ;CHECK: vaddw.s32
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i32>* %B
@@ -249,7 +245,7 @@ define <2 x i64> @vaddws32(<2 x i64>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <8 x i16> @vaddwu8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
-;CHECK: vaddwu8:
+;CHECK-LABEL: vaddwu8:
 ;CHECK: vaddw.u8
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i8>* %B
@@ -259,7 +255,7 @@ define <8 x i16> @vaddwu8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i32> @vaddwu16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
-;CHECK: vaddwu16:
+;CHECK-LABEL: vaddwu16:
 ;CHECK: vaddw.u16
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i16>* %B
@@ -269,7 +265,7 @@ define <4 x i32> @vaddwu16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i64> @vaddwu32(<2 x i64>* %A, <2 x i32>* %B) nounwind {
-;CHECK: vaddwu32:
+;CHECK-LABEL: vaddwu32:
 ;CHECK: vaddw.u32
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i32>* %B