1 ; RUN: llc < %s -march=arm -mattr=+neon | FileCheck %s
3 %struct.__neon_int8x8x4_t = type { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> }
4 %struct.__neon_int16x4x4_t = type { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }
5 %struct.__neon_int32x2x4_t = type { <2 x i32>, <2 x i32>, <2 x i32>, <2 x i32> }
6 %struct.__neon_float32x2x4_t = type { <2 x float>, <2 x float>, <2 x float>, <2 x float> }
8 %struct.__neon_int8x16x4_t = type { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> }
9 %struct.__neon_int16x8x4_t = type { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }
10 %struct.__neon_int32x4x4_t = type { <4 x i32>, <4 x i32>, <4 x i32>, <4 x i32> }
11 %struct.__neon_float32x4x4_t = type { <4 x float>, <4 x float>, <4 x float>, <4 x float> }
13 define <8 x i8> @vld4i8(i8* %A) nounwind {
16 %tmp1 = call %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4.v8i8(i8* %A)
17 %tmp2 = extractvalue %struct.__neon_int8x8x4_t %tmp1, 0
18 %tmp3 = extractvalue %struct.__neon_int8x8x4_t %tmp1, 2
19 %tmp4 = add <8 x i8> %tmp2, %tmp3
23 define <4 x i16> @vld4i16(i16* %A) nounwind {
26 %tmp1 = call %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4.v4i16(i16* %A)
27 %tmp2 = extractvalue %struct.__neon_int16x4x4_t %tmp1, 0
28 %tmp3 = extractvalue %struct.__neon_int16x4x4_t %tmp1, 2
29 %tmp4 = add <4 x i16> %tmp2, %tmp3
33 define <2 x i32> @vld4i32(i32* %A) nounwind {
36 %tmp1 = call %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4.v2i32(i32* %A)
37 %tmp2 = extractvalue %struct.__neon_int32x2x4_t %tmp1, 0
38 %tmp3 = extractvalue %struct.__neon_int32x2x4_t %tmp1, 2
39 %tmp4 = add <2 x i32> %tmp2, %tmp3
43 define <2 x float> @vld4f(float* %A) nounwind {
46 %tmp1 = call %struct.__neon_float32x2x4_t @llvm.arm.neon.vld4.v2f32(float* %A)
47 %tmp2 = extractvalue %struct.__neon_float32x2x4_t %tmp1, 0
48 %tmp3 = extractvalue %struct.__neon_float32x2x4_t %tmp1, 2
49 %tmp4 = add <2 x float> %tmp2, %tmp3
53 define <16 x i8> @vld4Qi8(i8* %A) nounwind {
57 %tmp1 = call %struct.__neon_int8x16x4_t @llvm.arm.neon.vld4.v16i8(i8* %A)
58 %tmp2 = extractvalue %struct.__neon_int8x16x4_t %tmp1, 0
59 %tmp3 = extractvalue %struct.__neon_int8x16x4_t %tmp1, 2
60 %tmp4 = add <16 x i8> %tmp2, %tmp3
64 define <8 x i16> @vld4Qi16(i16* %A) nounwind {
68 %tmp1 = call %struct.__neon_int16x8x4_t @llvm.arm.neon.vld4.v8i16(i16* %A)
69 %tmp2 = extractvalue %struct.__neon_int16x8x4_t %tmp1, 0
70 %tmp3 = extractvalue %struct.__neon_int16x8x4_t %tmp1, 2
71 %tmp4 = add <8 x i16> %tmp2, %tmp3
75 define <4 x i32> @vld4Qi32(i32* %A) nounwind {
79 %tmp1 = call %struct.__neon_int32x4x4_t @llvm.arm.neon.vld4.v4i32(i32* %A)
80 %tmp2 = extractvalue %struct.__neon_int32x4x4_t %tmp1, 0
81 %tmp3 = extractvalue %struct.__neon_int32x4x4_t %tmp1, 2
82 %tmp4 = add <4 x i32> %tmp2, %tmp3
86 define <4 x float> @vld4Qf(float* %A) nounwind {
90 %tmp1 = call %struct.__neon_float32x4x4_t @llvm.arm.neon.vld4.v4f32(float* %A)
91 %tmp2 = extractvalue %struct.__neon_float32x4x4_t %tmp1, 0
92 %tmp3 = extractvalue %struct.__neon_float32x4x4_t %tmp1, 2
93 %tmp4 = add <4 x float> %tmp2, %tmp3
97 declare %struct.__neon_int8x8x4_t @llvm.arm.neon.vld4.v8i8(i8*) nounwind readonly
98 declare %struct.__neon_int16x4x4_t @llvm.arm.neon.vld4.v4i16(i8*) nounwind readonly
99 declare %struct.__neon_int32x2x4_t @llvm.arm.neon.vld4.v2i32(i8*) nounwind readonly
100 declare %struct.__neon_float32x2x4_t @llvm.arm.neon.vld4.v2f32(i8*) nounwind readonly
102 declare %struct.__neon_int8x16x4_t @llvm.arm.neon.vld4.v16i8(i8*) nounwind readonly
103 declare %struct.__neon_int16x8x4_t @llvm.arm.neon.vld4.v8i16(i8*) nounwind readonly
104 declare %struct.__neon_int32x4x4_t @llvm.arm.neon.vld4.v4i32(i8*) nounwind readonly
105 declare %struct.__neon_float32x4x4_t @llvm.arm.neon.vld4.v4f32(i8*) nounwind readonly