1 ; RUN: llc < %s -mtriple=thumbv7-elf -mattr=+neon -arm-atomic-cfg-tidy=0 | FileCheck %s
4 %bar = type { float, float, float }
5 %baz = type { i32, [16 x %bar], [16 x float], [16 x i32], i8 }
6 %foo = type { <4 x float> }
7 %quux = type { i32 (...)**, %baz*, i32 }
8 %quuz = type { %quux, i32, %bar, [128 x i8], [16 x %foo], %foo, %foo, %foo }
10 declare <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8*, i32) nounwind readonly
12 define void @aaa(%quuz* %this, i8* %block) {
14 ; CHECK: bfc r4, #0, #4
15 ; CHECK: vst1.64 {{.*}}[{{.*}}:128]
16 ; CHECK: vld1.64 {{.*}}[{{.*}}:128]
18 %aligned_vec = alloca <4 x float>, align 16
19 %"alloca point" = bitcast i32 0 to i32
20 %vecptr = bitcast <4 x float>* %aligned_vec to i8*
21 %0 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* %vecptr, i32 1) nounwind
22 store float 6.300000e+01, float* undef, align 4
23 %1 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
24 store float 0.000000e+00, float* undef, align 4
25 %2 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind ; <<4 x float>> [#uses=1]
26 %ld3 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
27 store float 0.000000e+00, float* undef, align 4
28 %ld4 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
29 store float 0.000000e+00, float* undef, align 4
30 %ld5 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
31 store float 0.000000e+00, float* undef, align 4
32 %ld6 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
33 store float 0.000000e+00, float* undef, align 4
34 %ld7 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
35 store float 0.000000e+00, float* undef, align 4
36 %ld8 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
37 store float 0.000000e+00, float* undef, align 4
38 %ld9 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
39 store float 0.000000e+00, float* undef, align 4
40 %ld10 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
41 store float 0.000000e+00, float* undef, align 4
42 %ld11 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
43 store float 0.000000e+00, float* undef, align 4
44 %ld12 = call <4 x float> @llvm.arm.neon.vld1.v4f32.p0i8(i8* undef, i32 1) nounwind
45 store float 0.000000e+00, float* undef, align 4
46 %val173 = load <4 x float>, <4 x float>* undef ; <<4 x float>> [#uses=1]
49 bb4: ; preds = %bb193, %entry
50 %besterror.0.2264 = phi <4 x float> [ undef, %entry ], [ %besterror.0.0, %bb193 ] ; <<4 x float>> [#uses=2]
51 %part0.0.0261 = phi <4 x float> [ zeroinitializer, %entry ], [ %23, %bb193 ] ; <<4 x float>> [#uses=2]
52 %3 = fmul <4 x float> zeroinitializer, %0 ; <<4 x float>> [#uses=2]
53 %4 = fadd <4 x float> %3, %part0.0.0261 ; <<4 x float>> [#uses=1]
54 %5 = shufflevector <4 x float> %3, <4 x float> undef, <2 x i32> <i32 2, i32 3> ; <<2 x float>> [#uses=1]
55 %6 = shufflevector <2 x float> %5, <2 x float> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1> ; <<4 x float>> [#uses=1]
56 %7 = fmul <4 x float> %1, %1
57 %8 = fadd <4 x float> %7, <float 5.000000e-01, float 5.000000e-01, float 5.000000e-01, float 5.000000e-01> ; <<4 x float>> [#uses=1]
58 %9 = fptosi <4 x float> %8 to <4 x i32> ; <<4 x i32>> [#uses=1]
59 %10 = sitofp <4 x i32> %9 to <4 x float> ; <<4 x float>> [#uses=1]
60 %11 = fmul <4 x float> %10, %2 ; <<4 x float>> [#uses=1]
61 %12 = fmul <4 x float> %6, %6
62 %13 = fmul <4 x float> %11, %4 ; <<4 x float>> [#uses=1]
63 %14 = fsub <4 x float> %12, %13 ; <<4 x float>> [#uses=1]
64 %15 = fsub <4 x float> %14, %14
65 %16 = fmul <4 x float> %15, <float 2.000000e+00, float 2.000000e+00, float 2.000000e+00, float 2.000000e+00> ; <<4 x float>> [#uses=1]
66 %17 = fadd <4 x float> %16, %16
67 %18 = fmul <4 x float> %17, %val173 ; <<4 x float>> [#uses=1]
68 %19 = shufflevector <4 x float> %18, <4 x float> undef, <2 x i32> <i32 2, i32 3> ; <<2 x float>> [#uses=1]
69 %20 = shufflevector <2 x float> %19, <2 x float> undef, <4 x i32> zeroinitializer ; <<4 x float>> [#uses=1]
70 %tmp1 = fadd <4 x float> %20, %ld3
71 %tmp2 = fadd <4 x float> %tmp1, %ld4
72 %tmp3 = fadd <4 x float> %tmp2, %ld5
73 %tmp4 = fadd <4 x float> %tmp3, %ld6
74 %tmp5 = fadd <4 x float> %tmp4, %ld7
75 %tmp6 = fadd <4 x float> %tmp5, %ld8
76 %tmp7 = fadd <4 x float> %tmp6, %ld9
77 %tmp8 = fadd <4 x float> %tmp7, %ld10
78 %tmp9 = fadd <4 x float> %tmp8, %ld11
79 %21 = fadd <4 x float> %tmp9, %ld12
80 %22 = fcmp ogt <4 x float> %besterror.0.2264, %21 ; <<4 x i1>> [#uses=0]
81 %tmp = extractelement <4 x i1> %22, i32 0
82 br i1 %tmp, label %bb193, label %bb186
87 bb193: ; preds = %bb186, %bb4
88 %besterror.0.0 = phi <4 x float> [ %21, %bb186 ], [ %besterror.0.2264, %bb4 ] ; <<4 x float>> [#uses=1]
89 %23 = fadd <4 x float> %part0.0.0261, zeroinitializer ; <<4 x float>> [#uses=1]