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 i8, ptr %in0"
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=AVX512DQ
6 ; RUN: opt -passes=loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512vl,+avx512bw --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefix=AVX512BW
9 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
10 target triple = "x86_64-unknown-linux-gnu"
12 @A = global [1024 x i8] zeroinitializer, align 128
13 @B = global [1024 x i8] zeroinitializer, align 128
17 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
18 ; SSE2: LV: Found an estimated cost of 28 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
19 ; SSE2: LV: Found an estimated cost of 60 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
20 ; SSE2: LV: Found an estimated cost of 124 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
21 ; SSE2: LV: Found an estimated cost of 252 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
24 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
25 ; AVX1: LV: Found an estimated cost of 17 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
26 ; AVX1: LV: Found an estimated cost of 33 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
27 ; AVX1: LV: Found an estimated cost of 66 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
28 ; AVX1: LV: Found an estimated cost of 132 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
29 ; AVX1: LV: Found an estimated cost of 268 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
32 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
33 ; AVX2: LV: Found an estimated cost of 5 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
34 ; AVX2: LV: Found an estimated cost of 5 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
35 ; AVX2: LV: Found an estimated cost of 13 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
36 ; AVX2: LV: Found an estimated cost of 26 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
37 ; AVX2: LV: Found an estimated cost of 60 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
39 ; AVX512DQ-LABEL: 'test'
40 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
41 ; AVX512DQ: LV: Found an estimated cost of 5 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
42 ; AVX512DQ: LV: Found an estimated cost of 5 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
43 ; AVX512DQ: LV: Found an estimated cost of 13 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
44 ; AVX512DQ: LV: Found an estimated cost of 25 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
45 ; AVX512DQ: LV: Found an estimated cost of 58 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
46 ; AVX512DQ: LV: Found an estimated cost of 540 for VF 64 For instruction: %v0 = load i8, ptr %in0, align 1
48 ; AVX512BW-LABEL: 'test'
49 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
50 ; AVX512BW: LV: Found an estimated cost of 5 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
51 ; AVX512BW: LV: Found an estimated cost of 5 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
52 ; AVX512BW: LV: Found an estimated cost of 17 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
53 ; AVX512BW: LV: Found an estimated cost of 33 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
54 ; AVX512BW: LV: Found an estimated cost of 80 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
55 ; AVX512BW: LV: Found an estimated cost of 238 for VF 64 For instruction: %v0 = load i8, ptr %in0, align 1
61 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
63 %iv.0 = add nuw nsw i64 %iv, 0
64 %iv.1 = add nuw nsw i64 %iv, 1
65 %iv.2 = add nuw nsw i64 %iv, 2
66 %iv.3 = add nuw nsw i64 %iv, 3
68 %in0 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.0
69 %in1 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.1
70 %in2 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.2
71 %in3 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.3
73 %v0 = load i8, ptr %in0
74 %v1 = load i8, ptr %in1
75 %v2 = load i8, ptr %in2
76 %v3 = load i8, ptr %in3
78 %reduce.add.0 = add i8 %v0, %v1
79 %reduce.add.1 = add i8 %reduce.add.0, %v2
80 %reduce.add.2 = add i8 %reduce.add.1, %v3
82 %out = getelementptr inbounds [1024 x i8], ptr @B, i64 0, i64 %iv.0
83 store i8 %reduce.add.2, ptr %out
85 %iv.next = add nuw nsw i64 %iv.0, 4
86 %cmp = icmp ult i64 %iv.next, 1024
87 br i1 %cmp, label %for.body, label %for.cond.cleanup