1; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --filter "LV: Found an estimated cost of [0-9]+ for VF [0-9]+ For instruction:\s*%v0 = load float, float\* %in0, align 4"
2; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512
3; REQUIRES: asserts
4
5target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
6target triple = "x86_64-unknown-linux-gnu"
7
8@A = global [1024 x float] zeroinitializer, align 128
9@B = global [1024 x i8] zeroinitializer, align 128
10
11define void @test() {
12; CHECK-LABEL: 'test'
13; CHECK:  LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, float* %in0, align 4
14; CHECK:  LV: Found an estimated cost of 3 for VF 2 For instruction: %v0 = load float, float* %in0, align 4
15; CHECK:  LV: Found an estimated cost of 3 for VF 4 For instruction: %v0 = load float, float* %in0, align 4
16; CHECK:  LV: Found an estimated cost of 3 for VF 8 For instruction: %v0 = load float, float* %in0, align 4
17; CHECK:  LV: Found an estimated cost of 5 for VF 16 For instruction: %v0 = load float, float* %in0, align 4
18; CHECK:  LV: Found an estimated cost of 22 for VF 32 For instruction: %v0 = load float, float* %in0, align 4
19; CHECK:  LV: Found an estimated cost of 92 for VF 64 For instruction: %v0 = load float, float* %in0, align 4
20;
21entry:
22  br label %for.body
23
24for.body:
25  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
26
27  %iv.0 = add nuw nsw i64 %iv, 0
28  %iv.1 = add nuw nsw i64 %iv, 1
29
30  %in0 = getelementptr inbounds [1024 x float], [1024 x float]* @A, i64 0, i64 %iv.0
31  %in1 = getelementptr inbounds [1024 x float], [1024 x float]* @A, i64 0, i64 %iv.1
32
33  %v0 = load float, float* %in0
34  %v1 = load float, float* %in1
35
36  %reduce.add.0 = fadd float %v0, %v1
37
38  %reduce.add.0.narrow = fptoui float %reduce.add.0 to i8
39
40  %out = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.0
41  store i8 %reduce.add.0.narrow, i8* %out
42
43  %iv.next = add nuw nsw i64 %iv.0, 2
44  %cmp = icmp ult i64 %iv.next, 1024
45  br i1 %cmp, label %for.body, label %for.cond.cleanup
46
47for.cond.cleanup:
48  ret void
49}
50