1; REQUIRES: asserts
2; RUN: opt < %s -loop-vectorize -force-vector-width=2 -force-vector-interleave=1 -instcombine -debug-only=loop-vectorize -disable-output -print-after=instcombine -enable-new-pm=0 2>&1 | FileCheck %s
3; RUN: opt < %s -passes='function(loop-vectorize,instcombine)' -force-vector-width=2 -force-vector-interleave=1 -debug-only=loop-vectorize -disable-output -print-after=instcombine 2>&1 | FileCheck %s
4
5target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
6
7; CHECK-LABEL: phi_two_incoming_values
8; CHECK:       LV: Found an estimated cost of 1 for VF 2 For instruction: %i = phi i64 [ %i.next, %if.end ], [ 0, %entry ]
9; CHECK:       LV: Found an estimated cost of 1 for VF 2 For instruction: %tmp5 = phi i32 [ %tmp1, %for.body ], [ %tmp4, %if.then ]
10; CHECK:       vector.body:
11; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
12; CHECK:         [[WIDE_LOAD:%.*]] = load <2 x i32>, <2 x i32>* {{.*}}
13; CHECK:         [[TMP5:%.*]] = icmp sgt <2 x i32> [[WIDE_LOAD]], zeroinitializer
14; CHECK-NEXT:    [[TMP6:%.*]] = zext <2 x i1> [[TMP5]] to <2 x i32>
15; CHECK-NEXT:    [[PREDPHI:%.*]] = add <2 x i32> [[WIDE_LOAD]], [[TMP6]]
16; CHECK:         store <2 x i32> [[PREDPHI]], <2 x i32>* {{.*}}
17; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
18;
19define void @phi_two_incoming_values(i32* %a, i32* %b, i64 %n) {
20entry:
21  br label %for.body
22
23for.body:
24  %i = phi i64 [ %i.next, %if.end ], [ 0, %entry ]
25  %tmp0 = getelementptr inbounds i32, i32* %a, i64 %i
26  %tmp1 = load i32, i32* %tmp0, align 4
27  %tmp2 = getelementptr inbounds i32, i32* %b, i64 %i
28  %tmp3 = icmp sgt i32 %tmp1, 0
29  br i1 %tmp3, label %if.then, label %if.end
30
31if.then:
32  %tmp4 = add i32 %tmp1, 1
33  br label %if.end
34
35if.end:
36  %tmp5 = phi i32 [ %tmp1, %for.body ], [ %tmp4, %if.then ]
37  store i32 %tmp5, i32* %tmp2, align 4
38  %i.next = add i64 %i, 1
39  %cond = icmp eq i64 %i, %n
40  br i1 %cond, label %for.end, label %for.body
41
42for.end:
43  ret void
44}
45
46; CHECK-LABEL: phi_three_incoming_values
47; CHECK:       LV: Found an estimated cost of 1 for VF 2 For instruction: %i = phi i64 [ %i.next, %if.end ], [ 0, %entry ]
48; CHECK:       LV: Found an estimated cost of 2 for VF 2 For instruction: %tmp8 = phi i32 [ 9, %for.body ], [ 3, %if.then ], [ %tmp7, %if.else ]
49; CHECK:       vector.body:
50; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %vector.ph ], [ [[INDEX_NEXT:%.*]], %vector.body ]
51; CHECK:         [[PREDPHI:%.*]] = select <2 x i1> {{.*}}, <2 x i32> <i32 3, i32 3>, <2 x i32> <i32 9, i32 9>
52; CHECK:         [[PREDPHI7:%.*]] = select <2 x i1> {{.*}}, <2 x i32> {{.*}}, <2 x i32> [[PREDPHI]]
53; CHECK:         store <2 x i32> [[PREDPHI7]], <2 x i32>* {{.*}}
54; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
55;
56define void @phi_three_incoming_values(i32* %a, i32* %b, i64 %n) {
57entry:
58  br label %for.body
59
60for.body:
61  %i = phi i64 [ %i.next, %if.end ], [ 0, %entry ]
62  %tmp0 = getelementptr inbounds i32, i32* %a, i64 %i
63  %tmp1 = load i32, i32* %tmp0, align 4
64  %tmp2 = getelementptr inbounds i32, i32* %b, i64 %i
65  %tmp3 = load i32, i32* %tmp2, align 4
66  %tmp4 = icmp sgt i32 %tmp1, %tmp3
67  br i1 %tmp4, label %if.then, label %if.end
68
69if.then:
70  %tmp5 = icmp sgt i32 %tmp1, 19
71  br i1 %tmp5, label %if.end, label %if.else
72
73if.else:
74  %tmp6 = icmp slt i32 %tmp3, 4
75  %tmp7 = select i1 %tmp6, i32 4, i32 5
76  br label %if.end
77
78if.end:
79  %tmp8 = phi i32 [ 9, %for.body ], [ 3, %if.then ], [ %tmp7, %if.else ]
80  store i32 %tmp8, i32* %tmp0, align 4
81  %i.next = add i64 %i, 1
82  %cond = icmp eq i64 %i, %n
83  br i1 %cond, label %for.end, label %for.body
84
85for.end:
86  ret void
87}
88