Split a store of a VMOVDRR into two integer stores to avoid mixing NEON and ARM

[oota-llvm.git] / test / CodeGen / ARM / vqsub.ll

diff --git a/test/CodeGen/ARM/vqsub.ll b/test/CodeGen/ARM/vqsub.ll
index 07052f78dbea28eda3ba0eddea2d253e6554e56d..4231fca37e370036167505df26893491c6be04ad 100644 (file)
--- a/test/CodeGen/ARM/vqsub.ll
+++ b/test/CodeGen/ARM/vqsub.ll
@@ -1,14 +1,8 @@
-; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
-; RUN: grep {vqsub\\.s8} %t | count 2
-; RUN: grep {vqsub\\.s16} %t | count 2
-; RUN: grep {vqsub\\.s32} %t | count 2
-; RUN: grep {vqsub\\.s64} %t | count 2
-; RUN: grep {vqsub\\.u8} %t | count 2
-; RUN: grep {vqsub\\.u16} %t | count 2
-; RUN: grep {vqsub\\.u32} %t | count 2
-; RUN: grep {vqsub\\.u64} %t | count 2
+; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
 
 define <8 x i8> @vqsubs8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vqsubs8:
+;CHECK: vqsub.s8
        %tmp1 = load <8 x i8>* %A
        %tmp2 = load <8 x i8>* %B
        %tmp3 = call <8 x i8> @llvm.arm.neon.vqsubs.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -16,6 +10,8 @@ define <8 x i8> @vqsubs8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vqsubs16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vqsubs16:
+;CHECK: vqsub.s16
        %tmp1 = load <4 x i16>* %A
        %tmp2 = load <4 x i16>* %B
        %tmp3 = call <4 x i16> @llvm.arm.neon.vqsubs.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -23,6 +19,8 @@ define <4 x i16> @vqsubs16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vqsubs32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vqsubs32:
+;CHECK: vqsub.s32
        %tmp1 = load <2 x i32>* %A
        %tmp2 = load <2 x i32>* %B
        %tmp3 = call <2 x i32> @llvm.arm.neon.vqsubs.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -30,6 +28,8 @@ define <2 x i32> @vqsubs32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <1 x i64> @vqsubs64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+;CHECK: vqsubs64:
+;CHECK: vqsub.s64
        %tmp1 = load <1 x i64>* %A
        %tmp2 = load <1 x i64>* %B
        %tmp3 = call <1 x i64> @llvm.arm.neon.vqsubs.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2)
@@ -37,6 +37,8 @@ define <1 x i64> @vqsubs64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
 }
 
 define <8 x i8> @vqsubu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+;CHECK: vqsubu8:
+;CHECK: vqsub.u8
        %tmp1 = load <8 x i8>* %A
        %tmp2 = load <8 x i8>* %B
        %tmp3 = call <8 x i8> @llvm.arm.neon.vqsubu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
@@ -44,6 +46,8 @@ define <8 x i8> @vqsubu8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
 }
 
 define <4 x i16> @vqsubu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+;CHECK: vqsubu16:
+;CHECK: vqsub.u16
        %tmp1 = load <4 x i16>* %A
        %tmp2 = load <4 x i16>* %B
        %tmp3 = call <4 x i16> @llvm.arm.neon.vqsubu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
@@ -51,6 +55,8 @@ define <4 x i16> @vqsubu16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
 }
 
 define <2 x i32> @vqsubu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+;CHECK: vqsubu32:
+;CHECK: vqsub.u32
        %tmp1 = load <2 x i32>* %A
        %tmp2 = load <2 x i32>* %B
        %tmp3 = call <2 x i32> @llvm.arm.neon.vqsubu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
@@ -58,6 +64,8 @@ define <2 x i32> @vqsubu32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
 }
 
 define <1 x i64> @vqsubu64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+;CHECK: vqsubu64:
+;CHECK: vqsub.u64
        %tmp1 = load <1 x i64>* %A
        %tmp2 = load <1 x i64>* %B
        %tmp3 = call <1 x i64> @llvm.arm.neon.vqsubu.v1i64(<1 x i64> %tmp1, <1 x i64> %tmp2)
@@ -65,6 +73,8 @@ define <1 x i64> @vqsubu64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
 }
 
 define <16 x i8> @vqsubQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vqsubQs8:
+;CHECK: vqsub.s8
        %tmp1 = load <16 x i8>* %A
        %tmp2 = load <16 x i8>* %B
        %tmp3 = call <16 x i8> @llvm.arm.neon.vqsubs.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -72,6 +82,8 @@ define <16 x i8> @vqsubQs8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vqsubQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vqsubQs16:
+;CHECK: vqsub.s16
        %tmp1 = load <8 x i16>* %A
        %tmp2 = load <8 x i16>* %B
        %tmp3 = call <8 x i16> @llvm.arm.neon.vqsubs.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -79,6 +91,8 @@ define <8 x i16> @vqsubQs16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vqsubQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vqsubQs32:
+;CHECK: vqsub.s32
        %tmp1 = load <4 x i32>* %A
        %tmp2 = load <4 x i32>* %B
        %tmp3 = call <4 x i32> @llvm.arm.neon.vqsubs.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -86,6 +100,8 @@ define <4 x i32> @vqsubQs32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i64> @vqsubQs64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+;CHECK: vqsubQs64:
+;CHECK: vqsub.s64
        %tmp1 = load <2 x i64>* %A
        %tmp2 = load <2 x i64>* %B
        %tmp3 = call <2 x i64> @llvm.arm.neon.vqsubs.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2)
@@ -93,6 +109,8 @@ define <2 x i64> @vqsubQs64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
 }
 
 define <16 x i8> @vqsubQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+;CHECK: vqsubQu8:
+;CHECK: vqsub.u8
        %tmp1 = load <16 x i8>* %A
        %tmp2 = load <16 x i8>* %B
        %tmp3 = call <16 x i8> @llvm.arm.neon.vqsubu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
@@ -100,6 +118,8 @@ define <16 x i8> @vqsubQu8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
 }
 
 define <8 x i16> @vqsubQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+;CHECK: vqsubQu16:
+;CHECK: vqsub.u16
        %tmp1 = load <8 x i16>* %A
        %tmp2 = load <8 x i16>* %B
        %tmp3 = call <8 x i16> @llvm.arm.neon.vqsubu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
@@ -107,6 +127,8 @@ define <8 x i16> @vqsubQu16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
 }
 
 define <4 x i32> @vqsubQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+;CHECK: vqsubQu32:
+;CHECK: vqsub.u32
        %tmp1 = load <4 x i32>* %A
        %tmp2 = load <4 x i32>* %B
        %tmp3 = call <4 x i32> @llvm.arm.neon.vqsubu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
@@ -114,6 +136,8 @@ define <4 x i32> @vqsubQu32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
 }
 
 define <2 x i64> @vqsubQu64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+;CHECK: vqsubQu64:
+;CHECK: vqsub.u64
        %tmp1 = load <2 x i64>* %A
        %tmp2 = load <2 x i64>* %B
        %tmp3 = call <2 x i64> @llvm.arm.neon.vqsubu.v2i64(<2 x i64> %tmp1, <2 x i64> %tmp2)