1; RUN: opt < %s  -loop-vectorize -force-vector-width=1 -force-target-num-scalar-regs=16 -force-target-max-scalar-interleave=8 -force-target-instruction-cost=1 -small-loop-cost=40 -dce -instcombine -S | FileCheck %s
2
3target 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"
4target triple = "x86_64-apple-macosx10.8.0"
5
6@a = common global [2048 x i32] zeroinitializer, align 16
7
8; This is the loop.
9;  for (i=0; i<n; i++){
10;    a[i] += i;
11;  }
12;CHECK-LABEL: @inc(
13;CHECK: load i32*
14;CHECK: load i32*
15;CHECK: load i32*
16;CHECK: load i32*
17;CHECK-NOT: load i32*
18;CHECK: add nsw i32
19;CHECK: add nsw i32
20;CHECK: add nsw i32
21;CHECK: add nsw i32
22;CHECK-NOT: add nsw i32
23;CHECK: store i32
24;CHECK: store i32
25;CHECK: store i32
26;CHECK: store i32
27;CHECK-NOT: store i32
28;CHECK: add i64 %{{.*}}, 4
29;CHECK: ret void
30define void @inc(i32 %n) nounwind uwtable noinline ssp {
31  %1 = icmp sgt i32 %n, 0
32  br i1 %1, label %.lr.ph, label %._crit_edge
33
34.lr.ph:                                           ; preds = %0, %.lr.ph
35  %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ]
36  %2 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv
37  %3 = load i32* %2, align 4
38  %4 = trunc i64 %indvars.iv to i32
39  %5 = add nsw i32 %3, %4
40  store i32 %5, i32* %2, align 4
41  %indvars.iv.next = add i64 %indvars.iv, 1
42  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
43  %exitcond = icmp eq i32 %lftr.wideiv, %n
44  br i1 %exitcond, label %._crit_edge, label %.lr.ph
45
46._crit_edge:                                      ; preds = %.lr.ph, %0
47  ret void
48}
49
50