Add support for ARM's Advanced SIMD (NEON) instruction set.

This is still a work in progress but most of the NEON instruction set
is supported.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73919 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 76 insertions, 0 deletions

diff --git a/test/CodeGen/ARM/vsub.ll b/test/CodeGen/ARM/vsub.ll
new file mode 100644
index 0000000000..85dea41835
--- /dev/null
+++ b/test/CodeGen/ARM/vsub.ll
@@ -0,0 +1,76 @@
+; RUN: llvm-as < %s | llc -march=arm -mattr=+neon > %t
+; RUN: grep {vsub\\.i8} %t | count 2
+; RUN: grep {vsub\\.i16} %t | count 2
+; RUN: grep {vsub\\.i32} %t | count 2
+; RUN: grep {vsub\\.i64} %t | count 2
+; RUN: grep {vsub\\.f32} %t | count 2
+
+define <8 x i8> @vsubi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+	%tmp1 = load <8 x i8>* %A
+	%tmp2 = load <8 x i8>* %B
+	%tmp3 = sub <8 x i8> %tmp1, %tmp2
+	ret <8 x i8> %tmp3
+}
+
+define <4 x i16> @vsubi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+	%tmp1 = load <4 x i16>* %A
+	%tmp2 = load <4 x i16>* %B
+	%tmp3 = sub <4 x i16> %tmp1, %tmp2
+	ret <4 x i16> %tmp3
+}
+
+define <2 x i32> @vsubi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+	%tmp1 = load <2 x i32>* %A
+	%tmp2 = load <2 x i32>* %B
+	%tmp3 = sub <2 x i32> %tmp1, %tmp2
+	ret <2 x i32> %tmp3
+}
+
+define <1 x i64> @vsubi64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+	%tmp1 = load <1 x i64>* %A
+	%tmp2 = load <1 x i64>* %B
+	%tmp3 = sub <1 x i64> %tmp1, %tmp2
+	ret <1 x i64> %tmp3
+}
+
+define <2 x float> @vsubf32(<2 x float>* %A, <2 x float>* %B) nounwind {
+	%tmp1 = load <2 x float>* %A
+	%tmp2 = load <2 x float>* %B
+	%tmp3 = sub <2 x float> %tmp1, %tmp2
+	ret <2 x float> %tmp3
+}
+
+define <16 x i8> @vsubQi8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+	%tmp1 = load <16 x i8>* %A
+	%tmp2 = load <16 x i8>* %B
+	%tmp3 = sub <16 x i8> %tmp1, %tmp2
+	ret <16 x i8> %tmp3
+}
+
+define <8 x i16> @vsubQi16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+	%tmp1 = load <8 x i16>* %A
+	%tmp2 = load <8 x i16>* %B
+	%tmp3 = sub <8 x i16> %tmp1, %tmp2
+	ret <8 x i16> %tmp3
+}
+
+define <4 x i32> @vsubQi32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+	%tmp1 = load <4 x i32>* %A
+	%tmp2 = load <4 x i32>* %B
+	%tmp3 = sub <4 x i32> %tmp1, %tmp2
+	ret <4 x i32> %tmp3
+}
+
+define <2 x i64> @vsubQi64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+	%tmp1 = load <2 x i64>* %A
+	%tmp2 = load <2 x i64>* %B
+	%tmp3 = sub <2 x i64> %tmp1, %tmp2
+	ret <2 x i64> %tmp3
+}
+
+define <4 x float> @vsubQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
+	%tmp1 = load <4 x float>* %A
+	%tmp2 = load <4 x float>* %B
+	%tmp3 = sub <4 x float> %tmp1, %tmp2
+	ret <4 x float> %tmp3
+}