1 ; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
3 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
4 target triple = "x86_64-apple-macosx10.8.0"
6 define i32 @foo(ptr noalias nocapture %A, ptr noalias nocapture %B, i32 %start, i32 %end) nounwind uwtable ssp {
8 ;CHECK: cost of 1 {{.*}} icmp
9 %cmp7 = icmp slt i32 %start, %end
10 br i1 %cmp7, label %for.body.lr.ph, label %for.end
12 for.body.lr.ph: ; preds = %entry
13 ;CHECK: cost of 1 {{.*}} sext
14 %0 = sext i32 %start to i64
15 %1 = sub i32 %end, %start
16 %2 = zext i32 %1 to i64
17 %end.idx = add i64 %2, %0
18 ;CHECK: cost of 1 {{.*}} add
19 %n.vec = and i64 %2, 4294967288
20 %end.idx.rnd.down = add i64 %n.vec, %0
21 ;CHECK: cost of 1 {{.*}} icmp
22 %cmp.zero = icmp eq i64 %n.vec, 0
23 br i1 %cmp.zero, label %middle.block, label %vector.body
25 vector.body: ; preds = %for.body.lr.ph, %vector.body
26 %index = phi i64 [ %index.next, %vector.body ], [ %0, %for.body.lr.ph ]
27 %3 = add i64 %index, 2
28 %4 = getelementptr inbounds i32, ptr %B, i64 %3
29 ;CHECK: cost of 2 {{.*}} load
30 %5 = load <8 x i32>, ptr %4, align 4
31 ;CHECK: cost of 5 {{.*}} mul
32 %6 = mul nsw <8 x i32> %5, <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
33 %7 = getelementptr inbounds i32, ptr %A, i64 %index
34 ;CHECK: cost of 2 {{.*}} load
35 %8 = load <8 x i32>, ptr %7, align 4
36 ;CHECK: cost of 4 {{.*}} add
37 %9 = add nsw <8 x i32> %8, %6
38 ;CHECK: cost of 2 {{.*}} store
39 store <8 x i32> %9, ptr %7, align 4
40 %index.next = add i64 %index, 8
41 %10 = icmp eq i64 %index.next, %end.idx.rnd.down
42 ;CHECK: cost of 0 {{.*}} br
43 br i1 %10, label %middle.block, label %vector.body
45 middle.block: ; preds = %vector.body, %for.body.lr.ph
46 %cmp.n = icmp eq i64 %end.idx, %end.idx.rnd.down
47 br i1 %cmp.n, label %for.end, label %for.body
49 for.body: ; preds = %middle.block, %for.body
50 %indvars.iv = phi i64 [ %indvars.iv.next, %for.body ], [ %end.idx.rnd.down, %middle.block ]
51 %11 = add nsw i64 %indvars.iv, 2
52 %arrayidx = getelementptr inbounds i32, ptr %B, i64 %11
53 ;CHECK: cost of 1 {{.*}} load
54 %12 = load i32, ptr %arrayidx, align 4
55 ;CHECK: cost of 1 {{.*}} mul
56 %mul = mul nsw i32 %12, 5
57 %arrayidx2 = getelementptr inbounds i32, ptr %A, i64 %indvars.iv
58 ;CHECK: cost of 1 {{.*}} load
59 %13 = load i32, ptr %arrayidx2, align 4
60 %add3 = add nsw i32 %13, %mul
61 store i32 %add3, ptr %arrayidx2, align 4
62 %indvars.iv.next = add i64 %indvars.iv, 1
63 ;CHECK: cost of 0 {{.*}} trunc
64 %14 = trunc i64 %indvars.iv.next to i32
65 %cmp = icmp slt i32 %14, %end
66 ;CHECK: cost of 0 {{.*}} br
67 br i1 %cmp, label %for.body, label %for.end
69 for.end: ; preds = %middle.block, %for.body, %entry
70 ;CHECK: cost of 0 {{.*}} ret