1; RUN: llc -march=amdgcn < %s | FileCheck %s
2
3; Check we can compile this bugpoint-reduced test without an
4; infinite loop in TLI.SimplifyDemandedBits() due to failure
5; to use return value of TLO.DAG.UpdateNodeOperands()
6
7; Check that code was generated; we know there will be
8; a s_endpgm, so check for it.
9
10@0 = external unnamed_addr addrspace(3) global [462 x float], align 4
11
12; Function Attrs: nounwind readnone speculatable
13declare i32 @llvm.amdgcn.workitem.id.y() #0
14
15; Function Attrs: nounwind readnone speculatable
16declare i32 @llvm.amdgcn.workitem.id.x() #0
17
18; Function Attrs: nounwind readnone speculatable
19declare float @llvm.fmuladd.f32(float, float, float) #0
20
21; CHECK: s_endpgm
22define amdgpu_kernel void @foo(float addrspace(1)* noalias nocapture readonly %arg, float addrspace(1)* noalias nocapture readonly %arg1, float addrspace(1)* noalias nocapture %arg2, float %arg3) local_unnamed_addr !reqd_work_group_size !0 {
23bb:
24  %tmp = tail call i32 @llvm.amdgcn.workitem.id.y()
25  %tmp4 = tail call i32 @llvm.amdgcn.workitem.id.x()
26  %tmp5 = and i32 %tmp, 15
27  %tmp6 = mul nuw nsw i32 %tmp5, 21
28  %tmp7 = sub i32 %tmp6, 0
29  %tmp8 = add i32 %tmp7, 0
30  %tmp9 = add i32 %tmp8, 0
31  %tmp10 = getelementptr inbounds [462 x float], [462 x float] addrspace(3)* @0, i32 0, i32 0
32  br label %bb12
33
34bb11:                                             ; preds = %bb30
35  br i1 undef, label %bb37, label %bb38
36
37bb12:                                             ; preds = %bb30, %bb
38  br i1 false, label %.preheader, label %.loopexit145
39
40.loopexit145:                                     ; preds = %.preheader, %bb12
41  br label %bb13
42
43bb13:                                             ; preds = %.loopexit, %.loopexit145
44  %tmp14 = phi i32 [ %tmp5, %.loopexit145 ], [ %tmp20, %.loopexit ]
45  %tmp15 = add nsw i32 %tmp14, -3
46  %tmp16 = mul i32 %tmp14, 21
47  br i1 undef, label %bb17, label %.loopexit
48
49bb17:                                             ; preds = %bb13
50  %tmp18 = mul i32 %tmp15, 224
51  %tmp19 = add i32 undef, %tmp18
52  br label %bb21
53
54.loopexit:                                        ; preds = %bb21, %bb13
55  %tmp20 = add nuw nsw i32 %tmp14, 16
56  br i1 undef, label %bb13, label %bb26
57
58bb21:                                             ; preds = %bb21, %bb17
59  %tmp22 = phi i32 [ %tmp4, %bb17 ], [ %tmp25, %bb21 ]
60  %tmp23 = add i32 %tmp22, %tmp16
61  %tmp24 = getelementptr inbounds float, float addrspace(3)* %tmp10, i32 %tmp23
62  store float undef, float addrspace(3)* %tmp24, align 4
63  %tmp25 = add nuw i32 %tmp22, 8
64  br i1 undef, label %bb21, label %.loopexit
65
66bb26:                                             ; preds = %.loopexit
67  br label %bb31
68
69.preheader:                                       ; preds = %.preheader, %bb12
70  %tmp27 = phi i32 [ %tmp28, %.preheader ], [ undef, %bb12 ]
71  %tmp28 = add nuw i32 %tmp27, 128
72  %tmp29 = icmp ult i32 %tmp28, 1568
73  br i1 %tmp29, label %.preheader, label %.loopexit145
74
75bb30:                                             ; preds = %bb31
76  br i1 undef, label %bb11, label %bb12
77
78bb31:                                             ; preds = %bb31, %bb26
79  %tmp32 = phi i32 [ %tmp9, %bb26 ], [ undef, %bb31 ]
80  %tmp33 = getelementptr inbounds [462 x float], [462 x float] addrspace(3)* @0, i32 0, i32 %tmp32
81  %tmp34 = load float, float addrspace(3)* %tmp33, align 4
82  %tmp35 = tail call float @llvm.fmuladd.f32(float %tmp34, float undef, float undef)
83  %tmp36 = tail call float @llvm.fmuladd.f32(float undef, float undef, float %tmp35)
84  br i1 undef, label %bb30, label %bb31
85
86bb37:                                             ; preds = %bb11
87  br label %bb38
88
89bb38:                                             ; preds = %bb37, %bb11
90  ret void
91}
92
93attributes #0 = { nounwind readnone speculatable }
94
95!0 = !{i32 8, i32 16, i32 1}
96