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 i16, 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 i16] 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 i16, ptr %in0, align 2
18 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
19 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
20 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
21 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
22 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
23 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
24 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
25 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
26 ; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
27 ; SSE2: LV: Found an estimated cost of 22 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 2
28 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i16, ptr %in1, align 2
29 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i16, ptr %in2, align 2
30 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i16, ptr %in3, align 2
31 ; SSE2: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i16, ptr %in4, align 2
32 ; SSE2: LV: Found an estimated cost of 43 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 2
33 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i16, ptr %in1, align 2
34 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i16, ptr %in2, align 2
35 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i16, ptr %in3, align 2
36 ; SSE2: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i16, ptr %in4, align 2
37 ; SSE2: LV: Found an estimated cost of 85 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 2
38 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i16, ptr %in1, align 2
39 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i16, ptr %in2, align 2
40 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i16, ptr %in3, align 2
41 ; SSE2: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i16, ptr %in4, align 2
42 ; SSE2: LV: Found an estimated cost of 170 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 2
43 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i16, ptr %in1, align 2
44 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i16, ptr %in2, align 2
45 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i16, ptr %in3, align 2
46 ; SSE2: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i16, ptr %in4, align 2
49 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
50 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
51 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
52 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
53 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
54 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
55 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
56 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
57 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
58 ; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
59 ; AVX1: LV: Found an estimated cost of 25 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 2
60 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i16, ptr %in1, align 2
61 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i16, ptr %in2, align 2
62 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i16, ptr %in3, align 2
63 ; AVX1: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i16, ptr %in4, align 2
64 ; AVX1: LV: Found an estimated cost of 43 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 2
65 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i16, ptr %in1, align 2
66 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i16, ptr %in2, align 2
67 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i16, ptr %in3, align 2
68 ; AVX1: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i16, ptr %in4, align 2
69 ; AVX1: LV: Found an estimated cost of 85 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 2
70 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i16, ptr %in1, align 2
71 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i16, ptr %in2, align 2
72 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i16, ptr %in3, align 2
73 ; AVX1: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i16, ptr %in4, align 2
74 ; AVX1: LV: Found an estimated cost of 175 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 2
75 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i16, ptr %in1, align 2
76 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i16, ptr %in2, align 2
77 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i16, ptr %in3, align 2
78 ; AVX1: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i16, ptr %in4, align 2
79 ; AVX1: LV: Found an estimated cost of 350 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 2
80 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i16, ptr %in1, align 2
81 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i16, ptr %in2, align 2
82 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i16, ptr %in3, align 2
83 ; AVX1: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i16, ptr %in4, align 2
86 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
87 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
88 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
89 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
90 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
91 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
92 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
93 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
94 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
95 ; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
96 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 2
97 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v1 = load i16, ptr %in1, align 2
98 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v2 = load i16, ptr %in2, align 2
99 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v3 = load i16, ptr %in3, align 2
100 ; AVX2: LV: Found an estimated cost of 4 for VF 2 For instruction: %v4 = load i16, ptr %in4, align 2
101 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 2
102 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v1 = load i16, ptr %in1, align 2
103 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v2 = load i16, ptr %in2, align 2
104 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v3 = load i16, ptr %in3, align 2
105 ; AVX2: LV: Found an estimated cost of 8 for VF 4 For instruction: %v4 = load i16, ptr %in4, align 2
106 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 2
107 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v1 = load i16, ptr %in1, align 2
108 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v2 = load i16, ptr %in2, align 2
109 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v3 = load i16, ptr %in3, align 2
110 ; AVX2: LV: Found an estimated cost of 16 for VF 8 For instruction: %v4 = load i16, ptr %in4, align 2
111 ; AVX2: LV: Found an estimated cost of 33 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 2
112 ; AVX2: LV: Found an estimated cost of 33 for VF 16 For instruction: %v1 = load i16, ptr %in1, align 2
113 ; AVX2: LV: Found an estimated cost of 33 for VF 16 For instruction: %v2 = load i16, ptr %in2, align 2
114 ; AVX2: LV: Found an estimated cost of 33 for VF 16 For instruction: %v3 = load i16, ptr %in3, align 2
115 ; AVX2: LV: Found an estimated cost of 33 for VF 16 For instruction: %v4 = load i16, ptr %in4, align 2
116 ; AVX2: LV: Found an estimated cost of 66 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 2
117 ; AVX2: LV: Found an estimated cost of 66 for VF 32 For instruction: %v1 = load i16, ptr %in1, align 2
118 ; AVX2: LV: Found an estimated cost of 66 for VF 32 For instruction: %v2 = load i16, ptr %in2, align 2
119 ; AVX2: LV: Found an estimated cost of 66 for VF 32 For instruction: %v3 = load i16, ptr %in3, align 2
120 ; AVX2: LV: Found an estimated cost of 66 for VF 32 For instruction: %v4 = load i16, ptr %in4, align 2
122 ; AVX512DQ-LABEL: 'test'
123 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
124 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
125 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
126 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
127 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
128 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
129 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
130 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
131 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
132 ; AVX512DQ: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
133 ; AVX512DQ: LV: Found an estimated cost of 25 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 2
134 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i16, ptr %in1, align 2
135 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i16, ptr %in2, align 2
136 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i16, ptr %in3, align 2
137 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i16, ptr %in4, align 2
138 ; AVX512DQ: LV: Found an estimated cost of 45 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 2
139 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i16, ptr %in1, align 2
140 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i16, ptr %in2, align 2
141 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i16, ptr %in3, align 2
142 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i16, ptr %in4, align 2
143 ; AVX512DQ: LV: Found an estimated cost of 85 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 2
144 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i16, ptr %in1, align 2
145 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i16, ptr %in2, align 2
146 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i16, ptr %in3, align 2
147 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i16, ptr %in4, align 2
148 ; AVX512DQ: LV: Found an estimated cost of 175 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 2
149 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i16, ptr %in1, align 2
150 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i16, ptr %in2, align 2
151 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i16, ptr %in3, align 2
152 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i16, ptr %in4, align 2
153 ; AVX512DQ: LV: Found an estimated cost of 355 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 2
154 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i16, ptr %in1, align 2
155 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i16, ptr %in2, align 2
156 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i16, ptr %in3, align 2
157 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i16, ptr %in4, align 2
158 ; AVX512DQ: LV: Found an estimated cost of 710 for VF 64 For instruction: %v0 = load i16, ptr %in0, align 2
159 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v1 = load i16, ptr %in1, align 2
160 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v2 = load i16, ptr %in2, align 2
161 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v3 = load i16, ptr %in3, align 2
162 ; AVX512DQ: LV: Found an estimated cost of 0 for VF 64 For instruction: %v4 = load i16, ptr %in4, align 2
164 ; AVX512BW-LABEL: 'test'
165 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
166 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
167 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
168 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
169 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
170 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i16, ptr %in0, align 2
171 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v1 = load i16, ptr %in1, align 2
172 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v2 = load i16, ptr %in2, align 2
173 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v3 = load i16, ptr %in3, align 2
174 ; AVX512BW: LV: Found an estimated cost of 1 for VF 1 For instruction: %v4 = load i16, ptr %in4, align 2
175 ; AVX512BW: LV: Found an estimated cost of 11 for VF 2 For instruction: %v0 = load i16, ptr %in0, align 2
176 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v1 = load i16, ptr %in1, align 2
177 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v2 = load i16, ptr %in2, align 2
178 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v3 = load i16, ptr %in3, align 2
179 ; AVX512BW: LV: Found an estimated cost of 0 for VF 2 For instruction: %v4 = load i16, ptr %in4, align 2
180 ; AVX512BW: LV: Found an estimated cost of 11 for VF 4 For instruction: %v0 = load i16, ptr %in0, align 2
181 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v1 = load i16, ptr %in1, align 2
182 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v2 = load i16, ptr %in2, align 2
183 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v3 = load i16, ptr %in3, align 2
184 ; AVX512BW: LV: Found an estimated cost of 0 for VF 4 For instruction: %v4 = load i16, ptr %in4, align 2
185 ; AVX512BW: LV: Found an estimated cost of 14 for VF 8 For instruction: %v0 = load i16, ptr %in0, align 2
186 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v1 = load i16, ptr %in1, align 2
187 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v2 = load i16, ptr %in2, align 2
188 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v3 = load i16, ptr %in3, align 2
189 ; AVX512BW: LV: Found an estimated cost of 0 for VF 8 For instruction: %v4 = load i16, ptr %in4, align 2
190 ; AVX512BW: LV: Found an estimated cost of 28 for VF 16 For instruction: %v0 = load i16, ptr %in0, align 2
191 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v1 = load i16, ptr %in1, align 2
192 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v2 = load i16, ptr %in2, align 2
193 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v3 = load i16, ptr %in3, align 2
194 ; AVX512BW: LV: Found an estimated cost of 0 for VF 16 For instruction: %v4 = load i16, ptr %in4, align 2
195 ; AVX512BW: LV: Found an estimated cost of 55 for VF 32 For instruction: %v0 = load i16, ptr %in0, align 2
196 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v1 = load i16, ptr %in1, align 2
197 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v2 = load i16, ptr %in2, align 2
198 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v3 = load i16, ptr %in3, align 2
199 ; AVX512BW: LV: Found an estimated cost of 0 for VF 32 For instruction: %v4 = load i16, ptr %in4, align 2
200 ; AVX512BW: LV: Found an estimated cost of 235 for VF 64 For instruction: %v0 = load i16, ptr %in0, align 2
201 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v1 = load i16, ptr %in1, align 2
202 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v2 = load i16, ptr %in2, align 2
203 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v3 = load i16, ptr %in3, align 2
204 ; AVX512BW: LV: Found an estimated cost of 0 for VF 64 For instruction: %v4 = load i16, ptr %in4, align 2
210 %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
212 %iv.0 = add nuw nsw i64 %iv, 0
213 %iv.1 = add nuw nsw i64 %iv, 1
214 %iv.2 = add nuw nsw i64 %iv, 2
215 %iv.3 = add nuw nsw i64 %iv, 3
216 %iv.4 = add nuw nsw i64 %iv, 4
218 %in0 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.0
219 %in1 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.1
220 %in2 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.2
221 %in3 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.3
222 %in4 = getelementptr inbounds [1024 x i16], ptr @A, i64 0, i64 %iv.4
224 %v0 = load i16, ptr %in0
225 %v1 = load i16, ptr %in1
226 %v2 = load i16, ptr %in2
227 %v3 = load i16, ptr %in3
228 %v4 = load i16, ptr %in4
230 %reduce.add.0 = add i16 %v0, %v1
231 %reduce.add.1 = add i16 %reduce.add.0, %v2
232 %reduce.add.2 = add i16 %reduce.add.1, %v3
233 %reduce.add.3 = add i16 %reduce.add.2, %v4
235 %reduce.add.3.narrow = trunc i16 %reduce.add.3 to i8
237 %out = getelementptr inbounds [1024 x i8], ptr @B, i64 0, i64 %iv.0
238 store i8 %reduce.add.3.narrow, ptr %out
240 %iv.next = add nuw nsw i64 %iv.0, 5
241 %cmp = icmp ult i64 %iv.next, 1024
242 br i1 %cmp, label %for.body, label %for.cond.cleanup