1; RUN: llc < %s -march=arm -mattr=+neon -disable-arm-fast-isel -O0 | FileCheck %s 2 3define void @vst3i8(i8* %A, <8 x i8>* %B) nounwind { 4;CHECK: vst3i8: 5;Check the alignment value. Max for this instruction is 64 bits: 6;This test runs at -O0 so do not check for specific register numbers. 7;CHECK: vst3.8 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}, :64] 8 %tmp1 = load <8 x i8>* %B 9 call void @llvm.arm.neon.vst3.v8i8(i8* %A, <8 x i8> %tmp1, <8 x i8> %tmp1, <8 x i8> %tmp1, i32 32) 10 ret void 11} 12 13define void @vst3i16(i16* %A, <4 x i16>* %B) nounwind { 14;CHECK: vst3i16: 15;CHECK: vst3.16 16 %tmp0 = bitcast i16* %A to i8* 17 %tmp1 = load <4 x i16>* %B 18 call void @llvm.arm.neon.vst3.v4i16(i8* %tmp0, <4 x i16> %tmp1, <4 x i16> %tmp1, <4 x i16> %tmp1, i32 1) 19 ret void 20} 21 22define void @vst3i32(i32* %A, <2 x i32>* %B) nounwind { 23;CHECK: vst3i32: 24;CHECK: vst3.32 25 %tmp0 = bitcast i32* %A to i8* 26 %tmp1 = load <2 x i32>* %B 27 call void @llvm.arm.neon.vst3.v2i32(i8* %tmp0, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1) 28 ret void 29} 30 31;Check for a post-increment updating store. 32define void @vst3i32_update(i32** %ptr, <2 x i32>* %B) nounwind { 33;CHECK: vst3i32_update: 34;CHECK: vst3.32 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}]! 35 %A = load i32** %ptr 36 %tmp0 = bitcast i32* %A to i8* 37 %tmp1 = load <2 x i32>* %B 38 call void @llvm.arm.neon.vst3.v2i32(i8* %tmp0, <2 x i32> %tmp1, <2 x i32> %tmp1, <2 x i32> %tmp1, i32 1) 39 %tmp2 = getelementptr i32* %A, i32 6 40 store i32* %tmp2, i32** %ptr 41 ret void 42} 43 44define void @vst3f(float* %A, <2 x float>* %B) nounwind { 45;CHECK: vst3f: 46;CHECK: vst3.32 47 %tmp0 = bitcast float* %A to i8* 48 %tmp1 = load <2 x float>* %B 49 call void @llvm.arm.neon.vst3.v2f32(i8* %tmp0, <2 x float> %tmp1, <2 x float> %tmp1, <2 x float> %tmp1, i32 1) 50 ret void 51} 52 53define void @vst3i64(i64* %A, <1 x i64>* %B) nounwind { 54;CHECK: vst3i64: 55;Check the alignment value. Max for this instruction is 64 bits: 56;This test runs at -O0 so do not check for specific register numbers. 57;CHECK: vst1.64 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}, :64] 58 %tmp0 = bitcast i64* %A to i8* 59 %tmp1 = load <1 x i64>* %B 60 call void @llvm.arm.neon.vst3.v1i64(i8* %tmp0, <1 x i64> %tmp1, <1 x i64> %tmp1, <1 x i64> %tmp1, i32 16) 61 ret void 62} 63 64define void @vst3Qi8(i8* %A, <16 x i8>* %B) nounwind { 65;CHECK: vst3Qi8: 66;Check the alignment value. Max for this instruction is 64 bits: 67;This test runs at -O0 so do not check for specific register numbers. 68;CHECK: vst3.8 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}, :64]! 69;CHECK: vst3.8 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}, :64] 70 %tmp1 = load <16 x i8>* %B 71 call void @llvm.arm.neon.vst3.v16i8(i8* %A, <16 x i8> %tmp1, <16 x i8> %tmp1, <16 x i8> %tmp1, i32 32) 72 ret void 73} 74 75define void @vst3Qi16(i16* %A, <8 x i16>* %B) nounwind { 76;CHECK: vst3Qi16: 77;CHECK: vst3.16 78;CHECK: vst3.16 79 %tmp0 = bitcast i16* %A to i8* 80 %tmp1 = load <8 x i16>* %B 81 call void @llvm.arm.neon.vst3.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1) 82 ret void 83} 84 85;Check for a post-increment updating store. 86define void @vst3Qi16_update(i16** %ptr, <8 x i16>* %B) nounwind { 87;CHECK: vst3Qi16_update: 88;CHECK: vst3.16 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}]! 89;CHECK: vst3.16 {d{{.*}}, d{{.*}}, d{{.*}}}, [r{{.*}}]! 90 %A = load i16** %ptr 91 %tmp0 = bitcast i16* %A to i8* 92 %tmp1 = load <8 x i16>* %B 93 call void @llvm.arm.neon.vst3.v8i16(i8* %tmp0, <8 x i16> %tmp1, <8 x i16> %tmp1, <8 x i16> %tmp1, i32 1) 94 %tmp2 = getelementptr i16* %A, i32 24 95 store i16* %tmp2, i16** %ptr 96 ret void 97} 98 99define void @vst3Qi32(i32* %A, <4 x i32>* %B) nounwind { 100;CHECK: vst3Qi32: 101;CHECK: vst3.32 102;CHECK: vst3.32 103 %tmp0 = bitcast i32* %A to i8* 104 %tmp1 = load <4 x i32>* %B 105 call void @llvm.arm.neon.vst3.v4i32(i8* %tmp0, <4 x i32> %tmp1, <4 x i32> %tmp1, <4 x i32> %tmp1, i32 1) 106 ret void 107} 108 109define void @vst3Qf(float* %A, <4 x float>* %B) nounwind { 110;CHECK: vst3Qf: 111;CHECK: vst3.32 112;CHECK: vst3.32 113 %tmp0 = bitcast float* %A to i8* 114 %tmp1 = load <4 x float>* %B 115 call void @llvm.arm.neon.vst3.v4f32(i8* %tmp0, <4 x float> %tmp1, <4 x float> %tmp1, <4 x float> %tmp1, i32 1) 116 ret void 117} 118 119declare void @llvm.arm.neon.vst3.v8i8(i8*, <8 x i8>, <8 x i8>, <8 x i8>, i32) nounwind 120declare void @llvm.arm.neon.vst3.v4i16(i8*, <4 x i16>, <4 x i16>, <4 x i16>, i32) nounwind 121declare void @llvm.arm.neon.vst3.v2i32(i8*, <2 x i32>, <2 x i32>, <2 x i32>, i32) nounwind 122declare void @llvm.arm.neon.vst3.v2f32(i8*, <2 x float>, <2 x float>, <2 x float>, i32) nounwind 123declare void @llvm.arm.neon.vst3.v1i64(i8*, <1 x i64>, <1 x i64>, <1 x i64>, i32) nounwind 124 125declare void @llvm.arm.neon.vst3.v16i8(i8*, <16 x i8>, <16 x i8>, <16 x i8>, i32) nounwind 126declare void @llvm.arm.neon.vst3.v8i16(i8*, <8 x i16>, <8 x i16>, <8 x i16>, i32) nounwind 127declare void @llvm.arm.neon.vst3.v4i32(i8*, <4 x i32>, <4 x i32>, <4 x i32>, i32) nounwind 128declare void @llvm.arm.neon.vst3.v4f32(i8*, <4 x float>, <4 x float>, <4 x float>, i32) nounwind 129