1 ; RUN: llvm-as < %s | llc -march=arm -mattr=+neon | FileCheck %s
3 define <8 x i16> @vaddws8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
6 %tmp1 = load <8 x i16>* %A
7 %tmp2 = load <8 x i8>* %B
8 %tmp3 = call <8 x i16> @llvm.arm.neon.vaddws.v8i16(<8 x i16> %tmp1, <8 x i8> %tmp2)
12 define <4 x i32> @vaddws16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
15 %tmp1 = load <4 x i32>* %A
16 %tmp2 = load <4 x i16>* %B
17 %tmp3 = call <4 x i32> @llvm.arm.neon.vaddws.v4i32(<4 x i32> %tmp1, <4 x i16> %tmp2)
21 define <2 x i64> @vaddws32(<2 x i64>* %A, <2 x i32>* %B) nounwind {
24 %tmp1 = load <2 x i64>* %A
25 %tmp2 = load <2 x i32>* %B
26 %tmp3 = call <2 x i64> @llvm.arm.neon.vaddws.v2i64(<2 x i64> %tmp1, <2 x i32> %tmp2)
30 define <8 x i16> @vaddwu8(<8 x i16>* %A, <8 x i8>* %B) nounwind {
33 %tmp1 = load <8 x i16>* %A
34 %tmp2 = load <8 x i8>* %B
35 %tmp3 = call <8 x i16> @llvm.arm.neon.vaddwu.v8i16(<8 x i16> %tmp1, <8 x i8> %tmp2)
39 define <4 x i32> @vaddwu16(<4 x i32>* %A, <4 x i16>* %B) nounwind {
42 %tmp1 = load <4 x i32>* %A
43 %tmp2 = load <4 x i16>* %B
44 %tmp3 = call <4 x i32> @llvm.arm.neon.vaddwu.v4i32(<4 x i32> %tmp1, <4 x i16> %tmp2)
48 define <2 x i64> @vaddwu32(<2 x i64>* %A, <2 x i32>* %B) nounwind {
51 %tmp1 = load <2 x i64>* %A
52 %tmp2 = load <2 x i32>* %B
53 %tmp3 = call <2 x i64> @llvm.arm.neon.vaddwu.v2i64(<2 x i64> %tmp1, <2 x i32> %tmp2)
57 declare <8 x i16> @llvm.arm.neon.vaddws.v8i16(<8 x i16>, <8 x i8>) nounwind readnone
58 declare <4 x i32> @llvm.arm.neon.vaddws.v4i32(<4 x i32>, <4 x i16>) nounwind readnone
59 declare <2 x i64> @llvm.arm.neon.vaddws.v2i64(<2 x i64>, <2 x i32>) nounwind readnone
61 declare <8 x i16> @llvm.arm.neon.vaddwu.v8i16(<8 x i16>, <8 x i8>) nounwind readnone
62 declare <4 x i32> @llvm.arm.neon.vaddwu.v4i32(<4 x i32>, <4 x i16>) nounwind readnone
63 declare <2 x i64> @llvm.arm.neon.vaddwu.v2i64(<2 x i64>, <2 x i32>) nounwind readnone