1 ; RUN: opt -S -force-vector-width=2 -force-vector-interleave=1 -loop-vectorize -verify-loop-info -simplifycfg < %s | FileCheck %s
3 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
5 ; Test no-predication of instructions that are provably safe, e.g. dividing by
7 define void @test(i32* nocapture %asd, i32* nocapture %aud,
8 i32* nocapture %asr, i32* nocapture %aur,
9 i32* nocapture %asd0, i32* nocapture %aud0,
10 i32* nocapture %asr0, i32* nocapture %aur0
15 for.cond.cleanup: ; preds = %if.end
20 ; CHECK: %{{.*}} = sdiv <2 x i32> %{{.*}}, <i32 11, i32 11>
21 ; CHECK: %{{.*}} = udiv <2 x i32> %{{.*}}, <i32 13, i32 13>
22 ; CHECK: %{{.*}} = srem <2 x i32> %{{.*}}, <i32 17, i32 17>
23 ; CHECK: %{{.*}} = urem <2 x i32> %{{.*}}, <i32 19, i32 19>
24 ; CHECK-NOT: %{{.*}} = sdiv <2 x i32> %{{.*}}, <i32 0, i32 0>
25 ; CHECK-NOT: %{{.*}} = udiv <2 x i32> %{{.*}}, <i32 0, i32 0>
26 ; CHECK-NOT: %{{.*}} = srem <2 x i32> %{{.*}}, <i32 0, i32 0>
27 ; CHECK-NOT: %{{.*}} = urem <2 x i32> %{{.*}}, <i32 0, i32 0>
29 for.body: ; preds = %if.end, %entry
30 %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %if.end ]
31 %isd = getelementptr inbounds i32, i32* %asd, i64 %indvars.iv
32 %iud = getelementptr inbounds i32, i32* %aud, i64 %indvars.iv
33 %isr = getelementptr inbounds i32, i32* %asr, i64 %indvars.iv
34 %iur = getelementptr inbounds i32, i32* %aur, i64 %indvars.iv
35 %lsd = load i32, i32* %isd, align 4
36 %lud = load i32, i32* %iud, align 4
37 %lsr = load i32, i32* %isr, align 4
38 %lur = load i32, i32* %iur, align 4
39 %psd = add nsw i32 %lsd, 23
40 %pud = add nsw i32 %lud, 24
41 %psr = add nsw i32 %lsr, 25
42 %pur = add nsw i32 %lur, 26
43 %isd0 = getelementptr inbounds i32, i32* %asd0, i64 %indvars.iv
44 %iud0 = getelementptr inbounds i32, i32* %aud0, i64 %indvars.iv
45 %isr0 = getelementptr inbounds i32, i32* %asr0, i64 %indvars.iv
46 %iur0 = getelementptr inbounds i32, i32* %aur0, i64 %indvars.iv
47 %lsd0 = load i32, i32* %isd0, align 4
48 %lud0 = load i32, i32* %iud0, align 4
49 %lsr0 = load i32, i32* %isr0, align 4
50 %lur0 = load i32, i32* %iur0, align 4
51 %psd0 = add nsw i32 %lsd, 27
52 %pud0 = add nsw i32 %lud, 28
53 %psr0 = add nsw i32 %lsr, 29
54 %pur0 = add nsw i32 %lur, 30
55 %cmp1 = icmp slt i32 %lsd, 100
56 br i1 %cmp1, label %if.then, label %if.end
58 if.then: ; preds = %for.body
59 %rsd = sdiv i32 %psd, 11
60 %rud = udiv i32 %pud, 13
61 %rsr = srem i32 %psr, 17
62 %rur = urem i32 %pur, 19
63 %rsd0 = sdiv i32 %psd0, 0
64 %rud0 = udiv i32 %pud0, 0
65 %rsr0 = srem i32 %psr0, 0
66 %rur0 = urem i32 %pur0, 0
69 if.end: ; preds = %if.then, %for.body
70 %ysd.0 = phi i32 [ %rsd, %if.then ], [ %psd, %for.body ]
71 %yud.0 = phi i32 [ %rud, %if.then ], [ %pud, %for.body ]
72 %ysr.0 = phi i32 [ %rsr, %if.then ], [ %psr, %for.body ]
73 %yur.0 = phi i32 [ %rur, %if.then ], [ %pur, %for.body ]
74 %ysd0.0 = phi i32 [ %rsd0, %if.then ], [ %psd0, %for.body ]
75 %yud0.0 = phi i32 [ %rud0, %if.then ], [ %pud0, %for.body ]
76 %ysr0.0 = phi i32 [ %rsr0, %if.then ], [ %psr0, %for.body ]
77 %yur0.0 = phi i32 [ %rur0, %if.then ], [ %pur0, %for.body ]
78 store i32 %ysd.0, i32* %isd, align 4
79 store i32 %yud.0, i32* %iud, align 4
80 store i32 %ysr.0, i32* %isr, align 4
81 store i32 %yur.0, i32* %iur, align 4
82 store i32 %ysd0.0, i32* %isd0, align 4
83 store i32 %yud0.0, i32* %iud0, align 4
84 store i32 %ysr0.0, i32* %isr0, align 4
85 store i32 %yur0.0, i32* %iur0, align 4
86 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
87 %exitcond = icmp eq i64 %indvars.iv.next, 128
88 br i1 %exitcond, label %for.cond.cleanup, label %for.body