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 i32 %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 i32] zeroinitializer, align 128
16 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 4
17 ; SSE2: LV: Found an estimated cost of 28 for VF 2 For instruction: store i32 %v3, ptr %out3, align 4
18 ; SSE2: LV: Found an estimated cost of 60 for VF 4 For instruction: store i32 %v3, ptr %out3, align 4
19 ; SSE2: LV: Found an estimated cost of 120 for VF 8 For instruction: store i32 %v3, ptr %out3, align 4
20 ; SSE2: LV: Found an estimated cost of 240 for VF 16 For instruction: store i32 %v3, ptr %out3, align 4
23 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 4
24 ; AVX1: LV: Found an estimated cost of 18 for VF 2 For instruction: store i32 %v3, ptr %out3, align 4
25 ; AVX1: LV: Found an estimated cost of 36 for VF 4 For instruction: store i32 %v3, ptr %out3, align 4
26 ; AVX1: LV: Found an estimated cost of 76 for VF 8 For instruction: store i32 %v3, ptr %out3, align 4
27 ; AVX1: LV: Found an estimated cost of 152 for VF 16 For instruction: store i32 %v3, ptr %out3, align 4
28 ; AVX1: LV: Found an estimated cost of 304 for VF 32 For instruction: store i32 %v3, ptr %out3, align 4
31 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 4
32 ; AVX2: LV: Found an estimated cost of 6 for VF 2 For instruction: store i32 %v3, ptr %out3, align 4
33 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: store i32 %v3, ptr %out3, align 4
34 ; AVX2: LV: Found an estimated cost of 20 for VF 8 For instruction: store i32 %v3, ptr %out3, align 4
35 ; AVX2: LV: Found an estimated cost of 40 for VF 16 For instruction: store i32 %v3, ptr %out3, align 4
36 ; AVX2: LV: Found an estimated cost of 80 for VF 32 For instruction: store i32 %v3, ptr %out3, align 4
38 ; AVX512-LABEL: 'test'
39 ; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v3, ptr %out3, align 4
40 ; AVX512: LV: Found an estimated cost of 5 for VF 2 For instruction: store i32 %v3, ptr %out3, align 4
41 ; AVX512: LV: Found an estimated cost of 5 for VF 4 For instruction: store i32 %v3, ptr %out3, align 4
42 ; AVX512: LV: Found an estimated cost of 11 for VF 8 For instruction: store i32 %v3, ptr %out3, align 4
43 ; AVX512: LV: Found an estimated cost of 22 for VF 16 For instruction: store i32 %v3, ptr %out3, align 4
44 ; AVX512: LV: Found an estimated cost of 44 for VF 32 For instruction: store i32 %v3, ptr %out3, align 4
45 ; AVX512: LV: Found an estimated cost of 88 for VF 64 For instruction: store i32 %v3, ptr %out3, align 4
51 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
53 %iv.0 = add nuw nsw i64 %iv, 0
54 %iv.1 = add nuw nsw i64 %iv, 1
55 %iv.2 = add nuw nsw i64 %iv, 2
56 %iv.3 = add nuw nsw i64 %iv, 3
58 %in = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.0
59 %v.narrow = load i8, ptr %in
61 %v = zext i8 %v.narrow to i32
68 %out0 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.0
69 %out1 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.1
70 %out2 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.2
71 %out3 = getelementptr inbounds [1024 x i32], ptr @B, i64 0, i64 %iv.3
73 store i32 %v0, ptr %out0
74 store i32 %v1, ptr %out1
75 store i32 %v2, ptr %out2
76 store i32 %v3, ptr %out3
78 %iv.next = add nuw nsw i64 %iv.0, 4
79 %cmp = icmp ult i64 %iv.next, 1024
80 br i1 %cmp, label %for.body, label %for.cond.cleanup