Convert more NEON tests to use FileCheck.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83595 91177308-0d34-0410-b5e6-96231b3b80d8

9 files changed, 139 insertions, 46 deletions

diff --git a/test/CodeGen/ARM/vraddhn.ll b/test/CodeGen/ARM/vraddhn.ll
index b3f2f0d..40e5cea 100644
--- a/test/CodeGen/ARM/vraddhn.ll
+++ b/test/CodeGen/ARM/vraddhn.ll
@@ -1,9 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vraddhn\\.i16} %t | count 1
-; RUN: grep {vraddhn\\.i32} %t | count 1
-; RUN: grep {vraddhn\\.i64} %t | count 1
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vraddhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vraddhni16:
+;CHECK: vraddhn.i16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vraddhn.v8i8(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -11,6 +10,8 @@ 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: vraddhn.i32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vraddhn.v4i16(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -18,6 +19,8 @@ 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: vraddhn.i64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vraddhn.v2i32(<2 x i64> %tmp1, <2 x i64> %tmp2)
diff --git a/test/CodeGen/ARM/vrecpe.ll b/test/CodeGen/ARM/vrecpe.ll
index a97054f..72e904e 100644
--- a/test/CodeGen/ARM/vrecpe.ll
+++ b/test/CodeGen/ARM/vrecpe.ll
@@ -1,26 +1,32 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrecpe\\.u32} %t | count 2
-; RUN: grep {vrecpe\\.f32} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <2 x i32> @vrecpei32(<2 x i32>* %A) nounwind {
+;CHECK: vrecpei32:
+;CHECK: vrecpe.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i32> @llvm.arm.neon.vrecpe.v2i32(<2 x i32> %tmp1)
 	ret <2 x i32> %tmp2
 }
 
 define <4 x i32> @vrecpeQi32(<4 x i32>* %A) nounwind {
+;CHECK: vrecpeQi32:
+;CHECK: vrecpe.u32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vrecpe.v4i32(<4 x i32> %tmp1)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x float> @vrecpef32(<2 x float>* %A) nounwind {
+;CHECK: vrecpef32:
+;CHECK: vrecpe.f32
 	%tmp1 = load <2 x float>* %A
 	%tmp2 = call <2 x float> @llvm.arm.neon.vrecpe.v2f32(<2 x float> %tmp1)
 	ret <2 x float> %tmp2
 }
 
 define <4 x float> @vrecpeQf32(<4 x float>* %A) nounwind {
+;CHECK: vrecpeQf32:
+;CHECK: vrecpe.f32
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = call <4 x float> @llvm.arm.neon.vrecpe.v4f32(<4 x float> %tmp1)
 	ret <4 x float> %tmp2
diff --git a/test/CodeGen/ARM/vrecps.ll b/test/CodeGen/ARM/vrecps.ll
index 5ddd60b..71e8f86 100644
--- a/test/CodeGen/ARM/vrecps.ll
+++ b/test/CodeGen/ARM/vrecps.ll
@@ -1,7 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrecps\\.f32} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <2 x float> @vrecpsf32(<2 x float>* %A, <2 x float>* %B) nounwind {
+;CHECK: vrecpsf32:
+;CHECK: vrecps.f32
 	%tmp1 = load <2 x float>* %A
 	%tmp2 = load <2 x float>* %B
 	%tmp3 = call <2 x float> @llvm.arm.neon.vrecps.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
@@ -9,6 +10,8 @@ define <2 x float> @vrecpsf32(<2 x float>* %A, <2 x float>* %B) nounwind {
 }
 
 define <4 x float> @vrecpsQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
+;CHECK: vrecpsQf32:
+;CHECK: vrecps.f32
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = load <4 x float>* %B
 	%tmp3 = call <4 x float> @llvm.arm.neon.vrecps.v4f32(<4 x float> %tmp1, <4 x float> %tmp2)
diff --git a/test/CodeGen/ARM/vrhadd.ll b/test/CodeGen/ARM/vrhadd.ll
index 6fa9f5e..d47dfc4 100644
--- a/test/CodeGen/ARM/vrhadd.ll
+++ b/test/CodeGen/ARM/vrhadd.ll
@@ -1,12 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrhadd\\.s8} %t | count 2
-; RUN: grep {vrhadd\\.s16} %t | count 2
-; RUN: grep {vrhadd\\.s32} %t | count 2
-; RUN: grep {vrhadd\\.u8} %t | count 2
-; RUN: grep {vrhadd\\.u16} %t | count 2
-; RUN: grep {vrhadd\\.u32} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vrhadds8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vrhadds8:
+;CHECK: vrhadd.s8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vrhadds.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -14,6 +10,8 @@ define <8 x i8> @vrhadds8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vrhadds16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vrhadds16:
+;CHECK: vrhadd.s16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vrhadds.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -21,6 +19,8 @@ define <4 x i16> @vrhadds16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vrhadds32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vrhadds32:
+;CHECK: vrhadd.s32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vrhadds.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -28,6 +28,8 @@ define <2 x i32> @vrhadds32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <8 x i8> @vrhaddu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vrhaddu8:
+;CHECK: vrhadd.u8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vrhaddu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -35,6 +37,8 @@ define <8 x i8> @vrhaddu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vrhaddu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vrhaddu16:
+;CHECK: vrhadd.u16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vrhaddu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -42,6 +46,8 @@ define <4 x i16> @vrhaddu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vrhaddu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vrhaddu32:
+;CHECK: vrhadd.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vrhaddu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -49,6 +55,8 @@ define <2 x i32> @vrhaddu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <16 x i8> @vrhaddQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vrhaddQs8:
+;CHECK: vrhadd.s8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = load <16 x i8>* %B
 	%tmp3 = call <16 x i8> @llvm.arm.neon.vrhadds.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -56,6 +64,8 @@ define <16 x i8> @vrhaddQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vrhaddQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vrhaddQs16:
+;CHECK: vrhadd.s16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i16> @llvm.arm.neon.vrhadds.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -63,6 +73,8 @@ define <8 x i16> @vrhaddQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vrhaddQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vrhaddQs32:
+;CHECK: vrhadd.s32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i32> @llvm.arm.neon.vrhadds.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -70,6 +82,8 @@ define <4 x i32> @vrhaddQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <16 x i8> @vrhaddQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vrhaddQu8:
+;CHECK: vrhadd.u8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = load <16 x i8>* %B
 	%tmp3 = call <16 x i8> @llvm.arm.neon.vrhaddu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -77,6 +91,8 @@ define <16 x i8> @vrhaddQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vrhaddQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vrhaddQu16:
+;CHECK: vrhadd.u16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i16> @llvm.arm.neon.vrhaddu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -84,6 +100,8 @@ define <8 x i16> @vrhaddQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vrhaddQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vrhaddQu32:
+;CHECK: vrhadd.u32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i32> @llvm.arm.neon.vrhaddu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
diff --git a/test/CodeGen/ARM/vrshl.ll b/test/CodeGen/ARM/vrshl.ll
index df051e9..48fd388 100644
--- a/test/CodeGen/ARM/vrshl.ll
+++ b/test/CodeGen/ARM/vrshl.ll
@@ -1,22 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrshl\\.s8} %t | count 2
-; RUN: grep {vrshl\\.s16} %t | count 2
-; RUN: grep {vrshl\\.s32} %t | count 2
-; RUN: grep {vrshl\\.s64} %t | count 2
-; RUN: grep {vrshl\\.u8} %t | count 2
-; RUN: grep {vrshl\\.u16} %t | count 2
-; RUN: grep {vrshl\\.u32} %t | count 2
-; RUN: grep {vrshl\\.u64} %t | count 2
-; RUN: grep {vrshr\\.s8} %t | count 2
-; RUN: grep {vrshr\\.s16} %t | count 2
-; RUN: grep {vrshr\\.s32} %t | count 2
-; RUN: grep {vrshr\\.s64} %t | count 2
-; RUN: grep {vrshr\\.u8} %t | count 2
-; RUN: grep {vrshr\\.u16} %t | count 2
-; RUN: grep {vrshr\\.u32} %t | count 2
-; RUN: grep {vrshr\\.u64} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vrshls8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vrshls8:
+;CHECK: vrshl.s8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vrshifts.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -24,6 +10,8 @@ define <8 x i8> @vrshls8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vrshls16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vrshls16:
+;CHECK: vrshl.s16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vrshifts.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -31,6 +19,8 @@ define <4 x i16> @vrshls16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vrshls32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vrshls32:
+;CHECK: vrshl.s32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vrshifts.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -38,6 +28,8 @@ define <2 x i32> @vrshls32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <1 x i64> @vrshls64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+;CHECK: vrshls64:
+;CHECK: vrshl.s64
 	%tmp1 = load <1 x i64>* %A
 	%tmp2 = load <1 x i64>* %B
 	%tmp3 = call <1 x i64> @llvm.arm.neon.vrshifts.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2)
@@ -45,6 +37,8 @@ define <1 x i64> @vrshls64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
 }
 
 define <8 x i8> @vrshlu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vrshlu8:
+;CHECK: vrshl.u8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = load <8 x i8>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vrshiftu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -52,6 +46,8 @@ define <8 x i8> @vrshlu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vrshlu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vrshlu16:
+;CHECK: vrshl.u16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = load <4 x i16>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vrshiftu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -59,6 +55,8 @@ define <4 x i16> @vrshlu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vrshlu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vrshlu32:
+;CHECK: vrshl.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = load <2 x i32>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vrshiftu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -66,6 +64,8 @@ define <2 x i32> @vrshlu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <1 x i64> @vrshlu64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+;CHECK: vrshlu64:
+;CHECK: vrshl.u64
 	%tmp1 = load <1 x i64>* %A
 	%tmp2 = load <1 x i64>* %B
 	%tmp3 = call <1 x i64> @llvm.arm.neon.vrshiftu.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2)
@@ -73,6 +73,8 @@ define <1 x i64> @vrshlu64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
 }
 
 define <16 x i8> @vrshlQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vrshlQs8:
+;CHECK: vrshl.s8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = load <16 x i8>* %B
 	%tmp3 = call <16 x i8> @llvm.arm.neon.vrshifts.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -80,6 +82,8 @@ define <16 x i8> @vrshlQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vrshlQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vrshlQs16:
+;CHECK: vrshl.s16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i16> @llvm.arm.neon.vrshifts.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -87,6 +91,8 @@ define <8 x i16> @vrshlQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vrshlQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vrshlQs32:
+;CHECK: vrshl.s32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i32> @llvm.arm.neon.vrshifts.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -94,6 +100,8 @@ define <4 x i32> @vrshlQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i64> @vrshlQs64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+;CHECK: vrshlQs64:
+;CHECK: vrshl.s64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
 	%tmp3 = call <2 x i64> @llvm.arm.neon.vrshifts.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2)
@@ -101,6 +109,8 @@ define <2 x i64> @vrshlQs64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
 }
 
 define <16 x i8> @vrshlQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vrshlQu8:
+;CHECK: vrshl.u8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = load <16 x i8>* %B
 	%tmp3 = call <16 x i8> @llvm.arm.neon.vrshiftu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -108,6 +118,8 @@ define <16 x i8> @vrshlQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vrshlQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vrshlQu16:
+;CHECK: vrshl.u16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i16> @llvm.arm.neon.vrshiftu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -115,6 +127,8 @@ define <8 x i16> @vrshlQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vrshlQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vrshlQu32:
+;CHECK: vrshl.u32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i32> @llvm.arm.neon.vrshiftu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -122,6 +136,8 @@ define <4 x i32> @vrshlQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i64> @vrshlQu64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+;CHECK: vrshlQu64:
+;CHECK: vrshl.u64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
 	%tmp3 = call <2 x i64> @llvm.arm.neon.vrshiftu.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2)
@@ -129,96 +145,128 @@ define <2 x i64> @vrshlQu64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
 }
 
 define <8 x i8> @vrshrs8(<8 x i8>* %A) nounwind {
+;CHECK: vrshrs8:
+;CHECK: vrshr.s8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = call <8 x i8> @llvm.arm.neon.vrshifts.v8i8(<8 x i8> %tmp1, <8 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
 	ret <8 x i8> %tmp2
 }
 
 define <4 x i16> @vrshrs16(<4 x i16>* %A) nounwind {
+;CHECK: vrshrs16:
+;CHECK: vrshr.s16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = call <4 x i16> @llvm.arm.neon.vrshifts.v4i16(<4 x i16> %tmp1, <4 x i16> < i16 -16, i16 -16, i16 -16, i16 -16 >)
 	ret <4 x i16> %tmp2
 }
 
 define <2 x i32> @vrshrs32(<2 x i32>* %A) nounwind {
+;CHECK: vrshrs32:
+;CHECK: vrshr.s32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i32> @llvm.arm.neon.vrshifts.v2i32(<2 x i32> %tmp1, <2 x i32> < i32 -32, i32 -32 >)
 	ret <2 x i32> %tmp2
 }
 
 define <1 x i64> @vrshrs64(<1 x i64>* %A) nounwind {
+;CHECK: vrshrs64:
+;CHECK: vrshr.s64
 	%tmp1 = load <1 x i64>* %A
 	%tmp2 = call <1 x i64> @llvm.arm.neon.vrshifts.v1i64(<1 x i64> %tmp1, <1 x i64> < i64 -64 >)
 	ret <1 x i64> %tmp2
 }
 
 define <8 x i8> @vrshru8(<8 x i8>* %A) nounwind {
+;CHECK: vrshru8:
+;CHECK: vrshr.u8
 	%tmp1 = load <8 x i8>* %A
 	%tmp2 = call <8 x i8> @llvm.arm.neon.vrshiftu.v8i8(<8 x i8> %tmp1, <8 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
 	ret <8 x i8> %tmp2
 }
 
 define <4 x i16> @vrshru16(<4 x i16>* %A) nounwind {
+;CHECK: vrshru16:
+;CHECK: vrshr.u16
 	%tmp1 = load <4 x i16>* %A
 	%tmp2 = call <4 x i16> @llvm.arm.neon.vrshiftu.v4i16(<4 x i16> %tmp1, <4 x i16> < i16 -16, i16 -16, i16 -16, i16 -16 >)
 	ret <4 x i16> %tmp2
 }
 
 define <2 x i32> @vrshru32(<2 x i32>* %A) nounwind {
+;CHECK: vrshru32:
+;CHECK: vrshr.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i32> @llvm.arm.neon.vrshiftu.v2i32(<2 x i32> %tmp1, <2 x i32> < i32 -32, i32 -32 >)
 	ret <2 x i32> %tmp2
 }
 
 define <1 x i64> @vrshru64(<1 x i64>* %A) nounwind {
+;CHECK: vrshru64:
+;CHECK: vrshr.u64
 	%tmp1 = load <1 x i64>* %A
 	%tmp2 = call <1 x i64> @llvm.arm.neon.vrshiftu.v1i64(<1 x i64> %tmp1, <1 x i64> < i64 -64 >)
 	ret <1 x i64> %tmp2
 }
 
 define <16 x i8> @vrshrQs8(<16 x i8>* %A) nounwind {
+;CHECK: vrshrQs8:
+;CHECK: vrshr.s8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = call <16 x i8> @llvm.arm.neon.vrshifts.v16i8(<16 x i8> %tmp1, <16 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
 	ret <16 x i8> %tmp2
 }
 
 define <8 x i16> @vrshrQs16(<8 x i16>* %A) nounwind {
+;CHECK: vrshrQs16:
+;CHECK: vrshr.s16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = call <8 x i16> @llvm.arm.neon.vrshifts.v8i16(<8 x i16> %tmp1, <8 x i16> < i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16 >)
 	ret <8 x i16> %tmp2
 }
 
 define <4 x i32> @vrshrQs32(<4 x i32>* %A) nounwind {
+;CHECK: vrshrQs32:
+;CHECK: vrshr.s32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vrshifts.v4i32(<4 x i32> %tmp1, <4 x i32> < i32 -32, i32 -32, i32 -32, i32 -32 >)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x i64> @vrshrQs64(<2 x i64>* %A) nounwind {
+;CHECK: vrshrQs64:
+;CHECK: vrshr.s64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = call <2 x i64> @llvm.arm.neon.vrshifts.v2i64(<2 x i64> %tmp1, <2 x i64> < i64 -64, i64 -64 >)
 	ret <2 x i64> %tmp2
 }
 
 define <16 x i8> @vrshrQu8(<16 x i8>* %A) nounwind {
+;CHECK: vrshrQu8:
+;CHECK: vrshr.u8
 	%tmp1 = load <16 x i8>* %A
 	%tmp2 = call <16 x i8> @llvm.arm.neon.vrshiftu.v16i8(<16 x i8> %tmp1, <16 x i8> < i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8, i8 -8 >)
 	ret <16 x i8> %tmp2
 }
 
 define <8 x i16> @vrshrQu16(<8 x i16>* %A) nounwind {
+;CHECK: vrshrQu16:
+;CHECK: vrshr.u16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = call <8 x i16> @llvm.arm.neon.vrshiftu.v8i16(<8 x i16> %tmp1, <8 x i16> < i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16, i16 -16 >)
 	ret <8 x i16> %tmp2
 }
 
 define <4 x i32> @vrshrQu32(<4 x i32>* %A) nounwind {
+;CHECK: vrshrQu32:
+;CHECK: vrshr.u32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vrshiftu.v4i32(<4 x i32> %tmp1, <4 x i32> < i32 -32, i32 -32, i32 -32, i32 -32 >)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x i64> @vrshrQu64(<2 x i64>* %A) nounwind {
+;CHECK: vrshrQu64:
+;CHECK: vrshr.u64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = call <2 x i64> @llvm.arm.neon.vrshiftu.v2i64(<2 x i64> %tmp1, <2 x i64> < i64 -64, i64 -64 >)
 	ret <2 x i64> %tmp2
diff --git a/test/CodeGen/ARM/vrshrn.ll b/test/CodeGen/ARM/vrshrn.ll
index 3dd21bc..b0dedef 100644
--- a/test/CodeGen/ARM/vrshrn.ll
+++ b/test/CodeGen/ARM/vrshrn.ll
@@ -1,21 +1,24 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrshrn\\.i16} %t | count 1
-; RUN: grep {vrshrn\\.i32} %t | count 1
-; RUN: grep {vrshrn\\.i64} %t | count 1
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vrshrns8(<8 x i16>* %A) nounwind {
+;CHECK: vrshrns8:
+;CHECK: vrshrn.i16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = call <8 x i8> @llvm.arm.neon.vrshiftn.v8i8(<8 x i16> %tmp1, <8 x i16> < i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8, i16 -8 >)
 	ret <8 x i8> %tmp2
 }
 
 define <4 x i16> @vrshrns16(<4 x i32>* %A) nounwind {
+;CHECK: vrshrns16:
+;CHECK: vrshrn.i32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = call <4 x i16> @llvm.arm.neon.vrshiftn.v4i16(<4 x i32> %tmp1, <4 x i32> < i32 -16, i32 -16, i32 -16, i32 -16 >)
 	ret <4 x i16> %tmp2
 }
 
 define <2 x i32> @vrshrns32(<2 x i64>* %A) nounwind {
+;CHECK: vrshrns32:
+;CHECK: vrshrn.i64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = call <2 x i32> @llvm.arm.neon.vrshiftn.v2i32(<2 x i64> %tmp1, <2 x i64> < i64 -32, i64 -32 >)
 	ret <2 x i32> %tmp2
diff --git a/test/CodeGen/ARM/vrsqrte.ll b/test/CodeGen/ARM/vrsqrte.ll
index 5eb9494..ef0aa8a 100644
--- a/test/CodeGen/ARM/vrsqrte.ll
+++ b/test/CodeGen/ARM/vrsqrte.ll
@@ -1,26 +1,32 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrsqrte\\.u32} %t | count 2
-; RUN: grep {vrsqrte\\.f32} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <2 x i32> @vrsqrtei32(<2 x i32>* %A) nounwind {
+;CHECK: vrsqrtei32:
+;CHECK: vrsqrte.u32
 	%tmp1 = load <2 x i32>* %A
 	%tmp2 = call <2 x i32> @llvm.arm.neon.vrsqrte.v2i32(<2 x i32> %tmp1)
 	ret <2 x i32> %tmp2
 }
 
 define <4 x i32> @vrsqrteQi32(<4 x i32>* %A) nounwind {
+;CHECK: vrsqrteQi32:
+;CHECK: vrsqrte.u32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = call <4 x i32> @llvm.arm.neon.vrsqrte.v4i32(<4 x i32> %tmp1)
 	ret <4 x i32> %tmp2
 }
 
 define <2 x float> @vrsqrtef32(<2 x float>* %A) nounwind {
+;CHECK: vrsqrtef32:
+;CHECK: vrsqrte.f32
 	%tmp1 = load <2 x float>* %A
 	%tmp2 = call <2 x float> @llvm.arm.neon.vrsqrte.v2f32(<2 x float> %tmp1)
 	ret <2 x float> %tmp2
 }
 
 define <4 x float> @vrsqrteQf32(<4 x float>* %A) nounwind {
+;CHECK: vrsqrteQf32:
+;CHECK: vrsqrte.f32
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = call <4 x float> @llvm.arm.neon.vrsqrte.v4f32(<4 x float> %tmp1)
 	ret <4 x float> %tmp2
diff --git a/test/CodeGen/ARM/vrsqrts.ll b/test/CodeGen/ARM/vrsqrts.ll
index 46a4ce9..99f4714 100644
--- a/test/CodeGen/ARM/vrsqrts.ll
+++ b/test/CodeGen/ARM/vrsqrts.ll
@@ -1,7 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrsqrts\\.f32} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <2 x float> @vrsqrtsf32(<2 x float>* %A, <2 x float>* %B) nounwind {
+;CHECK: vrsqrtsf32:
+;CHECK: vrsqrts.f32
 	%tmp1 = load <2 x float>* %A
 	%tmp2 = load <2 x float>* %B
 	%tmp3 = call <2 x float> @llvm.arm.neon.vrsqrts.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
@@ -9,6 +10,8 @@ define <2 x float> @vrsqrtsf32(<2 x float>* %A, <2 x float>* %B) nounwind {
 }
 
 define <4 x float> @vrsqrtsQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
+;CHECK: vrsqrtsQf32:
+;CHECK: vrsqrts.f32
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = load <4 x float>* %B
 	%tmp3 = call <4 x float> @llvm.arm.neon.vrsqrts.v4f32(<4 x float> %tmp1, <4 x float> %tmp2)
diff --git a/test/CodeGen/ARM/vrsubhn.ll b/test/CodeGen/ARM/vrsubhn.ll
index 4691783..cd7b995 100644
--- a/test/CodeGen/ARM/vrsubhn.ll
+++ b/test/CodeGen/ARM/vrsubhn.ll
@@ -1,9 +1,8 @@
-; RUN: llc < %s -march=arm -mattr=+neon > %t
-; RUN: grep {vrsubhn\\.i16} %t | count 1
-; RUN: grep {vrsubhn\\.i32} %t | count 1
-; RUN: grep {vrsubhn\\.i64} %t | count 1
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vrsubhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vrsubhni16:
+;CHECK: vrsubhn.i16
 	%tmp1 = load <8 x i16>* %A
 	%tmp2 = load <8 x i16>* %B
 	%tmp3 = call <8 x i8> @llvm.arm.neon.vrsubhn.v8i8(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -11,6 +10,8 @@ define <8 x i8> @vrsubhni16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i16> @vrsubhni32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vrsubhni32:
+;CHECK: vrsubhn.i32
 	%tmp1 = load <4 x i32>* %A
 	%tmp2 = load <4 x i32>* %B
 	%tmp3 = call <4 x i16> @llvm.arm.neon.vrsubhn.v4i16(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -18,6 +19,8 @@ define <4 x i16> @vrsubhni32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i32> @vrsubhni64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+;CHECK: vrsubhni64:
+;CHECK: vrsubhn.i64
 	%tmp1 = load <2 x i64>* %A
 	%tmp2 = load <2 x i64>* %B
 	%tmp3 = call <2 x i32> @llvm.arm.neon.vrsubhn.v2i32(<2 x i64> %tmp1, <2 x i64> %tmp2)