1 ; RUN: opt %s -loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -S | FileCheck %s
2 ; RUN: opt %s -loop-vectorize -force-vector-interleave=2 -force-vector-width=4 -S | FileCheck %s -check-prefix=UNROLL
4 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
6 @b = common global [1000 x i32] zeroinitializer, align 16
7 @c = common global [1000 x i32] zeroinitializer, align 16
8 @a = common global [1000 x i32] zeroinitializer, align 16
10 ; Generate min.iters.check to skip the vector loop and jump to scalar.ph directly when loop iteration number is less than VF * UF.
12 ; CHECK: %min.iters.check = icmp ult i64 %N, 4
13 ; CHECK: br i1 %min.iters.check, label %scalar.ph, label %vector.ph
15 ; UNROLL: %min.iters.check = icmp ult i64 %N, 8
16 ; UNROLL: br i1 %min.iters.check, label %scalar.ph, label %vector.ph
18 define void @foo(i64 %N) {
20 %cmp.8 = icmp sgt i64 %N, 0
21 br i1 %cmp.8, label %for.body.preheader, label %for.end
23 for.body.preheader: ; preds = %entry
26 for.body: ; preds = %for.body, %for.body.preheader
27 %i.09 = phi i64 [ %inc, %for.body ], [ 0, %for.body.preheader ]
28 %arrayidx = getelementptr inbounds [1000 x i32], [1000 x i32]* @b, i64 0, i64 %i.09
29 %tmp = load i32, i32* %arrayidx, align 4
30 %arrayidx1 = getelementptr inbounds [1000 x i32], [1000 x i32]* @c, i64 0, i64 %i.09
31 %tmp1 = load i32, i32* %arrayidx1, align 4
32 %add = add nsw i32 %tmp1, %tmp
33 %arrayidx2 = getelementptr inbounds [1000 x i32], [1000 x i32]* @a, i64 0, i64 %i.09
34 store i32 %add, i32* %arrayidx2, align 4
35 %inc = add nuw nsw i64 %i.09, 1
36 %exitcond = icmp eq i64 %inc, %N
37 br i1 %exitcond, label %for.end.loopexit, label %for.body
39 for.end.loopexit: ; preds = %for.body
42 for.end: ; preds = %for.end.loopexit, %entry