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*%v. = load i8, ptr %in."
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 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
19 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
20 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
21 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
22 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
23 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
24 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
25 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
26 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
27 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
28 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
29 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
30 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
31 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
32 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
33 ; SSE2: LV: Found an estimated cost of 56 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
34 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i8, ptr %in1, align 1
35 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i8, ptr %in2, align 1
36 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i8, ptr %in3, align 1
37 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i8, ptr %in4, align 1
38 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v5 = load i8, ptr %in5, align 1
39 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v6 = load i8, ptr %in6, align 1
40 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v7 = load i8, ptr %in7, align 1
41 ; SSE2: LV: Found an estimated cost of 120 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
42 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i8, ptr %in1, align 1
43 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i8, ptr %in2, align 1
44 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i8, ptr %in3, align 1
45 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i8, ptr %in4, align 1
46 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v5 = load i8, ptr %in5, align 1
47 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v6 = load i8, ptr %in6, align 1
48 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v7 = load i8, ptr %in7, align 1
49 ; SSE2: LV: Found an estimated cost of 248 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
50 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i8, ptr %in1, align 1
51 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i8, ptr %in2, align 1
52 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i8, ptr %in3, align 1
53 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i8, ptr %in4, align 1
54 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v5 = load i8, ptr %in5, align 1
55 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v6 = load i8, ptr %in6, align 1
56 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v7 = load i8, ptr %in7, align 1
57 ; SSE2: LV: Found an estimated cost of 504 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
58 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i8, ptr %in1, align 1
59 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i8, ptr %in2, align 1
60 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i8, ptr %in3, align 1
61 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i8, ptr %in4, align 1
62 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v5 = load i8, ptr %in5, align 1
63 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v6 = load i8, ptr %in6, align 1
64 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v7 = load i8, ptr %in7, align 1
67 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
68 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
69 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
70 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
71 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
72 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
73 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
74 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
75 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
76 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
77 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
78 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
79 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
80 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
81 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
82 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
83 ; AVX1: LV: Found an estimated cost of 33 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
84 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i8, ptr %in1, align 1
85 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i8, ptr %in2, align 1
86 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i8, ptr %in3, align 1
87 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i8, ptr %in4, align 1
88 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v5 = load i8, ptr %in5, align 1
89 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v6 = load i8, ptr %in6, align 1
90 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v7 = load i8, ptr %in7, align 1
91 ; AVX1: LV: Found an estimated cost of 66 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
92 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i8, ptr %in1, align 1
93 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i8, ptr %in2, align 1
94 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i8, ptr %in3, align 1
95 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i8, ptr %in4, align 1
96 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v5 = load i8, ptr %in5, align 1
97 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v6 = load i8, ptr %in6, align 1
98 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v7 = load i8, ptr %in7, align 1
99 ; AVX1: LV: Found an estimated cost of 132 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
100 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i8, ptr %in1, align 1
101 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i8, ptr %in2, align 1
102 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i8, ptr %in3, align 1
103 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i8, ptr %in4, align 1
104 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v5 = load i8, ptr %in5, align 1
105 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v6 = load i8, ptr %in6, align 1
106 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v7 = load i8, ptr %in7, align 1
107 ; AVX1: LV: Found an estimated cost of 264 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
108 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i8, ptr %in1, align 1
109 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i8, ptr %in2, align 1
110 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i8, ptr %in3, align 1
111 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i8, ptr %in4, align 1
112 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v5 = load i8, ptr %in5, align 1
113 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v6 = load i8, ptr %in6, align 1
114 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v7 = load i8, ptr %in7, align 1
115 ; AVX1: LV: Found an estimated cost of 536 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
116 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i8, ptr %in1, align 1
117 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i8, ptr %in2, align 1
118 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i8, ptr %in3, align 1
119 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i8, ptr %in4, align 1
120 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v5 = load i8, ptr %in5, align 1
121 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v6 = load i8, ptr %in6, align 1
122 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v7 = load i8, ptr %in7, align 1
125 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
126 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
127 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
128 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
129 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
130 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
131 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
132 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
133 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
134 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
135 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
136 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
137 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
138 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
139 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
140 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
141 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
142 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v1 = load i8, ptr %in1, align 1
143 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v2 = load i8, ptr %in2, align 1
144 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v3 = load i8, ptr %in3, align 1
145 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v4 = load i8, ptr %in4, align 1
146 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v5 = load i8, ptr %in5, align 1
147 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v6 = load i8, ptr %in6, align 1
148 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v7 = load i8, ptr %in7, align 1
149 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
150 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v1 = load i8, ptr %in1, align 1
151 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v2 = load i8, ptr %in2, align 1
152 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v3 = load i8, ptr %in3, align 1
153 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v4 = load i8, ptr %in4, align 1
154 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v5 = load i8, ptr %in5, align 1
155 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v6 = load i8, ptr %in6, align 1
156 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v7 = load i8, ptr %in7, align 1
157 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
158 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v1 = load i8, ptr %in1, align 1
159 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v2 = load i8, ptr %in2, align 1
160 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v3 = load i8, ptr %in3, align 1
161 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v4 = load i8, ptr %in4, align 1
162 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v5 = load i8, ptr %in5, align 1
163 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v6 = load i8, ptr %in6, align 1
164 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v7 = load i8, ptr %in7, align 1
165 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
166 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v1 = load i8, ptr %in1, align 1
167 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v2 = load i8, ptr %in2, align 1
168 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v3 = load i8, ptr %in3, align 1
169 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v4 = load i8, ptr %in4, align 1
170 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v5 = load i8, ptr %in5, align 1
171 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v6 = load i8, ptr %in6, align 1
172 ; AVX2: LV: Found an estimated cost of 32 for VF 16 For instruction: %v7 = load i8, ptr %in7, align 1
173 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
174 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v1 = load i8, ptr %in1, align 1
175 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v2 = load i8, ptr %in2, align 1
176 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v3 = load i8, ptr %in3, align 1
177 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v4 = load i8, ptr %in4, align 1
178 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v5 = load i8, ptr %in5, align 1
179 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v6 = load i8, ptr %in6, align 1
180 ; AVX2: LV: Found an estimated cost of 65 for VF 32 For instruction: %v7 = load i8, ptr %in7, align 1
182 ; AVX512DQ-LABEL: 'test'
183 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
184 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
185 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
186 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
187 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
188 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
189 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
190 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
191 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
192 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
193 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
194 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
195 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
196 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
197 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
198 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
199 ; AVX512DQ: LV: Found an estimated cost of 33 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
200 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i8, ptr %in1, align 1
201 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i8, ptr %in2, align 1
202 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i8, ptr %in3, align 1
203 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i8, ptr %in4, align 1
204 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v5 = load i8, ptr %in5, align 1
205 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v6 = load i8, ptr %in6, align 1
206 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v7 = load i8, ptr %in7, align 1
207 ; AVX512DQ: LV: Found an estimated cost of 66 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
208 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i8, ptr %in1, align 1
209 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i8, ptr %in2, align 1
210 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i8, ptr %in3, align 1
211 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i8, ptr %in4, align 1
212 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v5 = load i8, ptr %in5, align 1
213 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v6 = load i8, ptr %in6, align 1
214 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v7 = load i8, ptr %in7, align 1
215 ; AVX512DQ: LV: Found an estimated cost of 132 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
216 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i8, ptr %in1, align 1
217 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i8, ptr %in2, align 1
218 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i8, ptr %in3, align 1
219 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i8, ptr %in4, align 1
220 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v5 = load i8, ptr %in5, align 1
221 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v6 = load i8, ptr %in6, align 1
222 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v7 = load i8, ptr %in7, align 1
223 ; AVX512DQ: LV: Found an estimated cost of 264 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
224 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i8, ptr %in1, align 1
225 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i8, ptr %in2, align 1
226 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i8, ptr %in3, align 1
227 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i8, ptr %in4, align 1
228 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v5 = load i8, ptr %in5, align 1
229 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v6 = load i8, ptr %in6, align 1
230 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v7 = load i8, ptr %in7, align 1
231 ; AVX512DQ: LV: Found an estimated cost of 536 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
232 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i8, ptr %in1, align 1
233 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i8, ptr %in2, align 1
234 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i8, ptr %in3, align 1
235 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i8, ptr %in4, align 1
236 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v5 = load i8, ptr %in5, align 1
237 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v6 = load i8, ptr %in6, align 1
238 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v7 = load i8, ptr %in7, align 1
239 ; AVX512DQ: LV: Found an estimated cost of 1080 for VF 64 For instruction: %v0 = load i8, ptr %in0, align 1
240 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v1 = load i8, ptr %in1, align 1
241 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v2 = load i8, ptr %in2, align 1
242 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v3 = load i8, ptr %in3, align 1
243 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v4 = load i8, ptr %in4, align 1
244 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v5 = load i8, ptr %in5, align 1
245 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v6 = load i8, ptr %in6, align 1
246 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v7 = load i8, ptr %in7, align 1
248 ; AVX512BW-LABEL: 'test'
249 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
250 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
251 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
252 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
253 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
254 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
255 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
256 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
257 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i8, ptr %in0, align 1
258 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i8, ptr %in1, align 1
259 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i8, ptr %in2, align 1
260 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i8, ptr %in3, align 1
261 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i8, ptr %in4, align 1
262 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v5 = load i8, ptr %in5, align 1
263 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v6 = load i8, ptr %in6, align 1
264 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v7 = load i8, ptr %in7, align 1
265 ; AVX512BW: LV: Found an estimated cost of 9 for VF 2 For instruction: %v0 = load i8, ptr %in0, align 1
266 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i8, ptr %in1, align 1
267 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i8, ptr %in2, align 1
268 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i8, ptr %in3, align 1
269 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i8, ptr %in4, align 1
270 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v5 = load i8, ptr %in5, align 1
271 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v6 = load i8, ptr %in6, align 1
272 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v7 = load i8, ptr %in7, align 1
273 ; AVX512BW: LV: Found an estimated cost of 33 for VF 4 For instruction: %v0 = load i8, ptr %in0, align 1
274 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i8, ptr %in1, align 1
275 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i8, ptr %in2, align 1
276 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i8, ptr %in3, align 1
277 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i8, ptr %in4, align 1
278 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v5 = load i8, ptr %in5, align 1
279 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v6 = load i8, ptr %in6, align 1
280 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v7 = load i8, ptr %in7, align 1
281 ; AVX512BW: LV: Found an estimated cost of 65 for VF 8 For instruction: %v0 = load i8, ptr %in0, align 1
282 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i8, ptr %in1, align 1
283 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i8, ptr %in2, align 1
284 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i8, ptr %in3, align 1
285 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i8, ptr %in4, align 1
286 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v5 = load i8, ptr %in5, align 1
287 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v6 = load i8, ptr %in6, align 1
288 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v7 = load i8, ptr %in7, align 1
289 ; AVX512BW: LV: Found an estimated cost of 158 for VF 16 For instruction: %v0 = load i8, ptr %in0, align 1
290 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i8, ptr %in1, align 1
291 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i8, ptr %in2, align 1
292 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i8, ptr %in3, align 1
293 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i8, ptr %in4, align 1
294 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v5 = load i8, ptr %in5, align 1
295 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v6 = load i8, ptr %in6, align 1
296 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v7 = load i8, ptr %in7, align 1
297 ; AVX512BW: LV: Found an estimated cost of 472 for VF 32 For instruction: %v0 = load i8, ptr %in0, align 1
298 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i8, ptr %in1, align 1
299 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i8, ptr %in2, align 1
300 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i8, ptr %in3, align 1
301 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i8, ptr %in4, align 1
302 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v5 = load i8, ptr %in5, align 1
303 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v6 = load i8, ptr %in6, align 1
304 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v7 = load i8, ptr %in7, align 1
305 ; AVX512BW: LV: Found an estimated cost of 1100 for VF 64 For instruction: %v0 = load i8, ptr %in0, align 1
306 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v1 = load i8, ptr %in1, align 1
307 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v2 = load i8, ptr %in2, align 1
308 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v3 = load i8, ptr %in3, align 1
309 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v4 = load i8, ptr %in4, align 1
310 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v5 = load i8, ptr %in5, align 1
311 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v6 = load i8, ptr %in6, align 1
312 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v7 = load i8, ptr %in7, align 1
318 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
320 %iv.0 = add nuw nsw i64 %iv, 0
321 %iv.1 = add nuw nsw i64 %iv, 1
322 %iv.2 = add nuw nsw i64 %iv, 2
323 %iv.3 = add nuw nsw i64 %iv, 3
324 %iv.4 = add nuw nsw i64 %iv, 4
325 %iv.5 = add nuw nsw i64 %iv, 5
326 %iv.6 = add nuw nsw i64 %iv, 6
327 %iv.7 = add nuw nsw i64 %iv, 7
329 %in0 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.0
330 %in1 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.1
331 %in2 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.2
332 %in3 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.3
333 %in4 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.4
334 %in5 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.5
335 %in6 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.6
336 %in7 = getelementptr inbounds [1024 x i8], ptr @A, i64 0, i64 %iv.7
338 %v0 = load i8, ptr %in0
339 %v1 = load i8, ptr %in1
340 %v2 = load i8, ptr %in2
341 %v3 = load i8, ptr %in3
342 %v4 = load i8, ptr %in4
343 %v5 = load i8, ptr %in5
344 %v6 = load i8, ptr %in6
345 %v7 = load i8, ptr %in7
347 %reduce.add.0 = add i8 %v0, %v1
348 %reduce.add.1 = add i8 %reduce.add.0, %v2
349 %reduce.add.2 = add i8 %reduce.add.1, %v3
350 %reduce.add.3 = add i8 %reduce.add.2, %v4
351 %reduce.add.4 = add i8 %reduce.add.3, %v5
352 %reduce.add.5 = add i8 %reduce.add.4, %v6
353 %reduce.add.6 = add i8 %reduce.add.5, %v7
355 %out = getelementptr inbounds [1024 x i8], ptr @B, i64 0, i64 %iv.0
356 store i8 %reduce.add.6, ptr %out
358 %iv.next = add nuw nsw i64 %iv.0, 8
359 %cmp = icmp ult i64 %iv.next, 1024
360 br i1 %cmp, label %for.body, label %for.cond.cleanup