1 ; RUN: opt -S -debug-only=loop-vectorize -loop-vectorize -instcombine < %s 2>&1 | FileCheck %s
4 target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
5 target triple = "aarch64--linux-gnueabi"
7 @AB = common global [1024 x i8] zeroinitializer, align 4
8 @CD = common global [1024 x i8] zeroinitializer, align 4
10 define void @test_byte_interleaved_cost(i8 %C, i8 %D) {
14 ; 8xi8 and 16xi8 are valid i8 vector types, so the cost of the interleaved
17 ; CHECK: LV: Found an estimated cost of 2 for VF 8 For instruction: %tmp = load i8, i8* %arrayidx0, align 4
18 ; CHECK: LV: Found an estimated cost of 2 for VF 16 For instruction: %tmp = load i8, i8* %arrayidx0, align 4
20 for.body: ; preds = %for.body, %entry
21 %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
22 %arrayidx0 = getelementptr inbounds [1024 x i8], [1024 x i8]* @AB, i64 0, i64 %indvars.iv
23 %tmp = load i8, i8* %arrayidx0, align 4
24 %tmp1 = or i64 %indvars.iv, 1
25 %arrayidx1 = getelementptr inbounds [1024 x i8], [1024 x i8]* @AB, i64 0, i64 %tmp1
26 %tmp2 = load i8, i8* %arrayidx1, align 4
27 %add = add nsw i8 %tmp, %C
28 %mul = mul nsw i8 %tmp2, %D
29 %arrayidx2 = getelementptr inbounds [1024 x i8], [1024 x i8]* @CD, i64 0, i64 %indvars.iv
30 store i8 %add, i8* %arrayidx2, align 4
31 %arrayidx3 = getelementptr inbounds [1024 x i8], [1024 x i8]* @CD, i64 0, i64 %tmp1
32 store i8 %mul, i8* %arrayidx3, align 4
33 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 2
34 %cmp = icmp slt i64 %indvars.iv.next, 1024
35 br i1 %cmp, label %for.body, label %for.end
37 for.end: ; preds = %for.body