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*store i64 %v3, ptr %out3"
2 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefix=SSE2
3 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefix=AVX1
4 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefix=AVX2
5 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefix=AVX512
8 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
9 target triple = "x86_64-unknown-linux-gnu"
11 @A = global [1024 x i8] zeroinitializer, align 128
12 @B = global [1024 x i64] zeroinitializer, align 128
16 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v3, ptr %out3, align 8
17 ; SSE2: LV: Found an estimated cost of 28 for VF 2 For instruction: store i64 %v3, ptr %out3, align 8
18 ; SSE2: LV: Found an estimated cost of 56 for VF 4 For instruction: store i64 %v3, ptr %out3, align 8
19 ; SSE2: LV: Found an estimated cost of 112 for VF 8 For instruction: store i64 %v3, ptr %out3, align 8
22 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v3, ptr %out3, align 8
23 ; AVX1: LV: Found an estimated cost of 20 for VF 2 For instruction: store i64 %v3, ptr %out3, align 8
24 ; AVX1: LV: Found an estimated cost of 44 for VF 4 For instruction: store i64 %v3, ptr %out3, align 8
25 ; AVX1: LV: Found an estimated cost of 88 for VF 8 For instruction: store i64 %v3, ptr %out3, align 8
26 ; AVX1: LV: Found an estimated cost of 176 for VF 16 For instruction: store i64 %v3, ptr %out3, align 8
29 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v3, ptr %out3, align 8
30 ; AVX2: LV: Found an estimated cost of 8 for VF 2 For instruction: store i64 %v3, ptr %out3, align 8
31 ; AVX2: LV: Found an estimated cost of 12 for VF 4 For instruction: store i64 %v3, ptr %out3, align 8
32 ; AVX2: LV: Found an estimated cost of 28 for VF 8 For instruction: store i64 %v3, ptr %out3, align 8
33 ; AVX2: LV: Found an estimated cost of 56 for VF 16 For instruction: store i64 %v3, ptr %out3, align 8
35 ; AVX512-LABEL: 'test'
36 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i64 %v3, ptr %out3, align 8
37 ; AVX512: LV: Found an estimated cost of 5 for VF 2 For instruction: store i64 %v3, ptr %out3, align 8
38 ; AVX512: LV: Found an estimated cost of 11 for VF 4 For instruction: store i64 %v3, ptr %out3, align 8
39 ; AVX512: LV: Found an estimated cost of 22 for VF 8 For instruction: store i64 %v3, ptr %out3, align 8
40 ; AVX512: LV: Found an estimated cost of 44 for VF 16 For instruction: store i64 %v3, ptr %out3, align 8
41 ; AVX512: LV: Found an estimated cost of 88 for VF 32 For instruction: store i64 %v3, ptr %out3, align 8
47 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
49 %iv.0 = add nuw nsw i64 %iv, 0
50 %iv.1 = add nuw nsw i64 %iv, 1
51 %iv.2 = add nuw nsw i64 %iv, 2
52 %iv.3 = add nuw nsw i64 %iv, 3
54 %in = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.0
55 %v.narrow = load i8, ptr %in
57 %v = zext i8 %v.narrow to i64
64 %out0 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.0
65 %out1 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.1
66 %out2 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.2
67 %out3 = getelementptr inbounds [1024 x i64], ptr @B, i64 0, i64 %iv.3
69 store i64 %v0, ptr %out0
70 store i64 %v1, ptr %out1
71 store i64 %v2, ptr %out2
72 store i64 %v3, ptr %out3
74 %iv.next = add nuw nsw i64 %iv.0, 4
75 %cmp = icmp ult i64 %iv.next, 1024
76 br i1 %cmp, label %for.body, label %for.cond.cleanup