1 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2 ; RUN: opt < %s -mtriple=x86_64-unknown -passes=slp-vectorizer -S | FileCheck %s --check-prefix=SSE
3 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=slm -passes=slp-vectorizer -S | FileCheck %s --check-prefix=SLM
4 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -mattr=+prefer-128-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX128
5 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -mattr=-prefer-128-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX256
6 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -mattr=+prefer-128-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX128
7 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -mattr=-prefer-128-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX256
8 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=knl -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX512
9 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=-prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX512
10 ; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=+prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefix=AVX256
12 @a64 = common global [8 x i64] zeroinitializer, align 64
13 @b64 = common global [8 x i64] zeroinitializer, align 64
14 @c64 = common global [8 x i64] zeroinitializer, align 64
15 @a32 = common global [16 x i32] zeroinitializer, align 64
16 @b32 = common global [16 x i32] zeroinitializer, align 64
17 @c32 = common global [16 x i32] zeroinitializer, align 64
18 @a16 = common global [32 x i16] zeroinitializer, align 64
19 @b16 = common global [32 x i16] zeroinitializer, align 64
20 @c16 = common global [32 x i16] zeroinitializer, align 64
21 @a8 = common global [64 x i8] zeroinitializer, align 64
22 @b8 = common global [64 x i8] zeroinitializer, align 64
23 @c8 = common global [64 x i8] zeroinitializer, align 64
25 define void @mul_v8i64() {
26 ; SSE-LABEL: @mul_v8i64(
27 ; SSE-NEXT: [[A0:%.*]] = load i64, ptr @a64, align 8
28 ; SSE-NEXT: [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8
29 ; SSE-NEXT: [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8
30 ; SSE-NEXT: [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8
31 ; SSE-NEXT: [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8
32 ; SSE-NEXT: [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8
33 ; SSE-NEXT: [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8
34 ; SSE-NEXT: [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8
35 ; SSE-NEXT: [[B0:%.*]] = load i64, ptr @b64, align 8
36 ; SSE-NEXT: [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8
37 ; SSE-NEXT: [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8
38 ; SSE-NEXT: [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8
39 ; SSE-NEXT: [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8
40 ; SSE-NEXT: [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8
41 ; SSE-NEXT: [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8
42 ; SSE-NEXT: [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8
43 ; SSE-NEXT: [[R0:%.*]] = mul i64 [[A0]], [[B0]]
44 ; SSE-NEXT: [[R1:%.*]] = mul i64 [[A1]], [[B1]]
45 ; SSE-NEXT: [[R2:%.*]] = mul i64 [[A2]], [[B2]]
46 ; SSE-NEXT: [[R3:%.*]] = mul i64 [[A3]], [[B3]]
47 ; SSE-NEXT: [[R4:%.*]] = mul i64 [[A4]], [[B4]]
48 ; SSE-NEXT: [[R5:%.*]] = mul i64 [[A5]], [[B5]]
49 ; SSE-NEXT: [[R6:%.*]] = mul i64 [[A6]], [[B6]]
50 ; SSE-NEXT: [[R7:%.*]] = mul i64 [[A7]], [[B7]]
51 ; SSE-NEXT: store i64 [[R0]], ptr @c64, align 8
52 ; SSE-NEXT: store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8
53 ; SSE-NEXT: store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8
54 ; SSE-NEXT: store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8
55 ; SSE-NEXT: store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8
56 ; SSE-NEXT: store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8
57 ; SSE-NEXT: store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8
58 ; SSE-NEXT: store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8
61 ; SLM-LABEL: @mul_v8i64(
62 ; SLM-NEXT: [[A0:%.*]] = load i64, ptr @a64, align 8
63 ; SLM-NEXT: [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8
64 ; SLM-NEXT: [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8
65 ; SLM-NEXT: [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8
66 ; SLM-NEXT: [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8
67 ; SLM-NEXT: [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8
68 ; SLM-NEXT: [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8
69 ; SLM-NEXT: [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8
70 ; SLM-NEXT: [[B0:%.*]] = load i64, ptr @b64, align 8
71 ; SLM-NEXT: [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8
72 ; SLM-NEXT: [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8
73 ; SLM-NEXT: [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8
74 ; SLM-NEXT: [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8
75 ; SLM-NEXT: [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8
76 ; SLM-NEXT: [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8
77 ; SLM-NEXT: [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8
78 ; SLM-NEXT: [[R0:%.*]] = mul i64 [[A0]], [[B0]]
79 ; SLM-NEXT: [[R1:%.*]] = mul i64 [[A1]], [[B1]]
80 ; SLM-NEXT: [[R2:%.*]] = mul i64 [[A2]], [[B2]]
81 ; SLM-NEXT: [[R3:%.*]] = mul i64 [[A3]], [[B3]]
82 ; SLM-NEXT: [[R4:%.*]] = mul i64 [[A4]], [[B4]]
83 ; SLM-NEXT: [[R5:%.*]] = mul i64 [[A5]], [[B5]]
84 ; SLM-NEXT: [[R6:%.*]] = mul i64 [[A6]], [[B6]]
85 ; SLM-NEXT: [[R7:%.*]] = mul i64 [[A7]], [[B7]]
86 ; SLM-NEXT: store i64 [[R0]], ptr @c64, align 8
87 ; SLM-NEXT: store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8
88 ; SLM-NEXT: store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8
89 ; SLM-NEXT: store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8
90 ; SLM-NEXT: store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8
91 ; SLM-NEXT: store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8
92 ; SLM-NEXT: store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8
93 ; SLM-NEXT: store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8
96 ; AVX128-LABEL: @mul_v8i64(
97 ; AVX128-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 8
98 ; AVX128-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 8
99 ; AVX128-NEXT: [[TMP3:%.*]] = mul <2 x i64> [[TMP1]], [[TMP2]]
100 ; AVX128-NEXT: store <2 x i64> [[TMP3]], ptr @c64, align 8
101 ; AVX128-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8
102 ; AVX128-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8
103 ; AVX128-NEXT: [[TMP6:%.*]] = mul <2 x i64> [[TMP4]], [[TMP5]]
104 ; AVX128-NEXT: store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8
105 ; AVX128-NEXT: [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8
106 ; AVX128-NEXT: [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8
107 ; AVX128-NEXT: [[TMP9:%.*]] = mul <2 x i64> [[TMP7]], [[TMP8]]
108 ; AVX128-NEXT: store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8
109 ; AVX128-NEXT: [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8
110 ; AVX128-NEXT: [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8
111 ; AVX128-NEXT: [[TMP12:%.*]] = mul <2 x i64> [[TMP10]], [[TMP11]]
112 ; AVX128-NEXT: store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8
113 ; AVX128-NEXT: ret void
115 ; AVX256-LABEL: @mul_v8i64(
116 ; AVX256-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8
117 ; AVX256-NEXT: [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8
118 ; AVX256-NEXT: [[TMP3:%.*]] = mul <4 x i64> [[TMP1]], [[TMP2]]
119 ; AVX256-NEXT: store <4 x i64> [[TMP3]], ptr @c64, align 8
120 ; AVX256-NEXT: [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8
121 ; AVX256-NEXT: [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8
122 ; AVX256-NEXT: [[TMP6:%.*]] = mul <4 x i64> [[TMP4]], [[TMP5]]
123 ; AVX256-NEXT: store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8
124 ; AVX256-NEXT: ret void
126 ; AVX512-LABEL: @mul_v8i64(
127 ; AVX512-NEXT: [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8
128 ; AVX512-NEXT: [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8
129 ; AVX512-NEXT: [[TMP3:%.*]] = mul <8 x i64> [[TMP1]], [[TMP2]]
130 ; AVX512-NEXT: store <8 x i64> [[TMP3]], ptr @c64, align 8
131 ; AVX512-NEXT: ret void
133 %a0 = load i64, ptr @a64, align 8
134 %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8
135 %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8
136 %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8
137 %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8
138 %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8
139 %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8
140 %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8
141 %b0 = load i64, ptr @b64, align 8
142 %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8
143 %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8
144 %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8
145 %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8
146 %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8
147 %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8
148 %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8
149 %r0 = mul i64 %a0, %b0
150 %r1 = mul i64 %a1, %b1
151 %r2 = mul i64 %a2, %b2
152 %r3 = mul i64 %a3, %b3
153 %r4 = mul i64 %a4, %b4
154 %r5 = mul i64 %a5, %b5
155 %r6 = mul i64 %a6, %b6
156 %r7 = mul i64 %a7, %b7
157 store i64 %r0, ptr @c64, align 8
158 store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8
159 store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8
160 store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8
161 store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8
162 store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8
163 store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8
164 store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8
168 define void @mul_v16i32() {
169 ; SSE-LABEL: @mul_v16i32(
170 ; SSE-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4
171 ; SSE-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4
172 ; SSE-NEXT: [[TMP3:%.*]] = mul <4 x i32> [[TMP1]], [[TMP2]]
173 ; SSE-NEXT: store <4 x i32> [[TMP3]], ptr @c32, align 4
174 ; SSE-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4
175 ; SSE-NEXT: [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4
176 ; SSE-NEXT: [[TMP6:%.*]] = mul <4 x i32> [[TMP4]], [[TMP5]]
177 ; SSE-NEXT: store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4
178 ; SSE-NEXT: [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4
179 ; SSE-NEXT: [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4
180 ; SSE-NEXT: [[TMP9:%.*]] = mul <4 x i32> [[TMP7]], [[TMP8]]
181 ; SSE-NEXT: store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4
182 ; SSE-NEXT: [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4
183 ; SSE-NEXT: [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4
184 ; SSE-NEXT: [[TMP12:%.*]] = mul <4 x i32> [[TMP10]], [[TMP11]]
185 ; SSE-NEXT: store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4
188 ; SLM-LABEL: @mul_v16i32(
189 ; SLM-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4
190 ; SLM-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4
191 ; SLM-NEXT: [[TMP3:%.*]] = mul <4 x i32> [[TMP1]], [[TMP2]]
192 ; SLM-NEXT: store <4 x i32> [[TMP3]], ptr @c32, align 4
193 ; SLM-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4
194 ; SLM-NEXT: [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4
195 ; SLM-NEXT: [[TMP6:%.*]] = mul <4 x i32> [[TMP4]], [[TMP5]]
196 ; SLM-NEXT: store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4
197 ; SLM-NEXT: [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4
198 ; SLM-NEXT: [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4
199 ; SLM-NEXT: [[TMP9:%.*]] = mul <4 x i32> [[TMP7]], [[TMP8]]
200 ; SLM-NEXT: store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4
201 ; SLM-NEXT: [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4
202 ; SLM-NEXT: [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4
203 ; SLM-NEXT: [[TMP12:%.*]] = mul <4 x i32> [[TMP10]], [[TMP11]]
204 ; SLM-NEXT: store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4
207 ; AVX128-LABEL: @mul_v16i32(
208 ; AVX128-NEXT: [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4
209 ; AVX128-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4
210 ; AVX128-NEXT: [[TMP3:%.*]] = mul <4 x i32> [[TMP1]], [[TMP2]]
211 ; AVX128-NEXT: store <4 x i32> [[TMP3]], ptr @c32, align 4
212 ; AVX128-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4
213 ; AVX128-NEXT: [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4
214 ; AVX128-NEXT: [[TMP6:%.*]] = mul <4 x i32> [[TMP4]], [[TMP5]]
215 ; AVX128-NEXT: store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4
216 ; AVX128-NEXT: [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4
217 ; AVX128-NEXT: [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4
218 ; AVX128-NEXT: [[TMP9:%.*]] = mul <4 x i32> [[TMP7]], [[TMP8]]
219 ; AVX128-NEXT: store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4
220 ; AVX128-NEXT: [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4
221 ; AVX128-NEXT: [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4
222 ; AVX128-NEXT: [[TMP12:%.*]] = mul <4 x i32> [[TMP10]], [[TMP11]]
223 ; AVX128-NEXT: store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4
224 ; AVX128-NEXT: ret void
226 ; AVX256-LABEL: @mul_v16i32(
227 ; AVX256-NEXT: [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4
228 ; AVX256-NEXT: [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4
229 ; AVX256-NEXT: [[TMP3:%.*]] = mul <8 x i32> [[TMP1]], [[TMP2]]
230 ; AVX256-NEXT: store <8 x i32> [[TMP3]], ptr @c32, align 4
231 ; AVX256-NEXT: [[TMP4:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4
232 ; AVX256-NEXT: [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4
233 ; AVX256-NEXT: [[TMP6:%.*]] = mul <8 x i32> [[TMP4]], [[TMP5]]
234 ; AVX256-NEXT: store <8 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4
235 ; AVX256-NEXT: ret void
237 ; AVX512-LABEL: @mul_v16i32(
238 ; AVX512-NEXT: [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4
239 ; AVX512-NEXT: [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4
240 ; AVX512-NEXT: [[TMP3:%.*]] = mul <16 x i32> [[TMP1]], [[TMP2]]
241 ; AVX512-NEXT: store <16 x i32> [[TMP3]], ptr @c32, align 4
242 ; AVX512-NEXT: ret void
244 %a0 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4
245 %a1 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4
246 %a2 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 4
247 %a3 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 4
248 %a4 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 4
249 %a5 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 4
250 %a6 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 4
251 %a7 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 4
252 %a8 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 4
253 %a9 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 4
254 %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4
255 %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4
256 %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4
257 %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4
258 %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4
259 %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4
260 %b0 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4
261 %b1 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4
262 %b2 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 4
263 %b3 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 4
264 %b4 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 4
265 %b5 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 4
266 %b6 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 4
267 %b7 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 4
268 %b8 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 4
269 %b9 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 4
270 %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4
271 %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4
272 %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4
273 %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4
274 %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4
275 %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4
276 %r0 = mul i32 %a0 , %b0
277 %r1 = mul i32 %a1 , %b1
278 %r2 = mul i32 %a2 , %b2
279 %r3 = mul i32 %a3 , %b3
280 %r4 = mul i32 %a4 , %b4
281 %r5 = mul i32 %a5 , %b5
282 %r6 = mul i32 %a6 , %b6
283 %r7 = mul i32 %a7 , %b7
284 %r8 = mul i32 %a8 , %b8
285 %r9 = mul i32 %a9 , %b9
286 %r10 = mul i32 %a10, %b10
287 %r11 = mul i32 %a11, %b11
288 %r12 = mul i32 %a12, %b12
289 %r13 = mul i32 %a13, %b13
290 %r14 = mul i32 %a14, %b14
291 %r15 = mul i32 %a15, %b15
292 store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 0 ), align 4
293 store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1 ), align 4
294 store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2 ), align 4
295 store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3 ), align 4
296 store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4 ), align 4
297 store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5 ), align 4
298 store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6 ), align 4
299 store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7 ), align 4
300 store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8 ), align 4
301 store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9 ), align 4
302 store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4
303 store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4
304 store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4
305 store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4
306 store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4
307 store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4
311 define void @mul_v32i16() {
312 ; SSE-LABEL: @mul_v32i16(
313 ; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2
314 ; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2
315 ; SSE-NEXT: [[TMP3:%.*]] = mul <8 x i16> [[TMP1]], [[TMP2]]
316 ; SSE-NEXT: store <8 x i16> [[TMP3]], ptr @c16, align 2
317 ; SSE-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2
318 ; SSE-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2
319 ; SSE-NEXT: [[TMP6:%.*]] = mul <8 x i16> [[TMP4]], [[TMP5]]
320 ; SSE-NEXT: store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2
321 ; SSE-NEXT: [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2
322 ; SSE-NEXT: [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2
323 ; SSE-NEXT: [[TMP9:%.*]] = mul <8 x i16> [[TMP7]], [[TMP8]]
324 ; SSE-NEXT: store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2
325 ; SSE-NEXT: [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2
326 ; SSE-NEXT: [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2
327 ; SSE-NEXT: [[TMP12:%.*]] = mul <8 x i16> [[TMP10]], [[TMP11]]
328 ; SSE-NEXT: store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2
331 ; SLM-LABEL: @mul_v32i16(
332 ; SLM-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2
333 ; SLM-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2
334 ; SLM-NEXT: [[TMP3:%.*]] = mul <8 x i16> [[TMP1]], [[TMP2]]
335 ; SLM-NEXT: store <8 x i16> [[TMP3]], ptr @c16, align 2
336 ; SLM-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2
337 ; SLM-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2
338 ; SLM-NEXT: [[TMP6:%.*]] = mul <8 x i16> [[TMP4]], [[TMP5]]
339 ; SLM-NEXT: store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2
340 ; SLM-NEXT: [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2
341 ; SLM-NEXT: [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2
342 ; SLM-NEXT: [[TMP9:%.*]] = mul <8 x i16> [[TMP7]], [[TMP8]]
343 ; SLM-NEXT: store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2
344 ; SLM-NEXT: [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2
345 ; SLM-NEXT: [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2
346 ; SLM-NEXT: [[TMP12:%.*]] = mul <8 x i16> [[TMP10]], [[TMP11]]
347 ; SLM-NEXT: store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2
350 ; AVX128-LABEL: @mul_v32i16(
351 ; AVX128-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2
352 ; AVX128-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2
353 ; AVX128-NEXT: [[TMP3:%.*]] = mul <8 x i16> [[TMP1]], [[TMP2]]
354 ; AVX128-NEXT: store <8 x i16> [[TMP3]], ptr @c16, align 2
355 ; AVX128-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2
356 ; AVX128-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2
357 ; AVX128-NEXT: [[TMP6:%.*]] = mul <8 x i16> [[TMP4]], [[TMP5]]
358 ; AVX128-NEXT: store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2
359 ; AVX128-NEXT: [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2
360 ; AVX128-NEXT: [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2
361 ; AVX128-NEXT: [[TMP9:%.*]] = mul <8 x i16> [[TMP7]], [[TMP8]]
362 ; AVX128-NEXT: store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2
363 ; AVX128-NEXT: [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2
364 ; AVX128-NEXT: [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2
365 ; AVX128-NEXT: [[TMP12:%.*]] = mul <8 x i16> [[TMP10]], [[TMP11]]
366 ; AVX128-NEXT: store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2
367 ; AVX128-NEXT: ret void
369 ; AVX256-LABEL: @mul_v32i16(
370 ; AVX256-NEXT: [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 2
371 ; AVX256-NEXT: [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 2
372 ; AVX256-NEXT: [[TMP3:%.*]] = mul <16 x i16> [[TMP1]], [[TMP2]]
373 ; AVX256-NEXT: store <16 x i16> [[TMP3]], ptr @c16, align 2
374 ; AVX256-NEXT: [[TMP4:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2
375 ; AVX256-NEXT: [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2
376 ; AVX256-NEXT: [[TMP6:%.*]] = mul <16 x i16> [[TMP4]], [[TMP5]]
377 ; AVX256-NEXT: store <16 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2
378 ; AVX256-NEXT: ret void
380 ; AVX512-LABEL: @mul_v32i16(
381 ; AVX512-NEXT: [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2
382 ; AVX512-NEXT: [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2
383 ; AVX512-NEXT: [[TMP3:%.*]] = mul <32 x i16> [[TMP1]], [[TMP2]]
384 ; AVX512-NEXT: store <32 x i16> [[TMP3]], ptr @c16, align 2
385 ; AVX512-NEXT: ret void
387 %a0 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 2
388 %a1 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 2
389 %a2 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 2
390 %a3 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 2
391 %a4 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 2
392 %a5 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 2
393 %a6 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 2
394 %a7 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 2
395 %a8 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 2
396 %a9 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 2
397 %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2
398 %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2
399 %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2
400 %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2
401 %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2
402 %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2
403 %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2
404 %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2
405 %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2
406 %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2
407 %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2
408 %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2
409 %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2
410 %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2
411 %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2
412 %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2
413 %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2
414 %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2
415 %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2
416 %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2
417 %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2
418 %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2
419 %b0 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 2
420 %b1 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 2
421 %b2 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 2
422 %b3 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 2
423 %b4 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 2
424 %b5 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 2
425 %b6 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 2
426 %b7 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 2
427 %b8 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 2
428 %b9 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 2
429 %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2
430 %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2
431 %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2
432 %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2
433 %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2
434 %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2
435 %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2
436 %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2
437 %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2
438 %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2
439 %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2
440 %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2
441 %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2
442 %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2
443 %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2
444 %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2
445 %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2
446 %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2
447 %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2
448 %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2
449 %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2
450 %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2
451 %r0 = mul i16 %a0 , %b0
452 %r1 = mul i16 %a1 , %b1
453 %r2 = mul i16 %a2 , %b2
454 %r3 = mul i16 %a3 , %b3
455 %r4 = mul i16 %a4 , %b4
456 %r5 = mul i16 %a5 , %b5
457 %r6 = mul i16 %a6 , %b6
458 %r7 = mul i16 %a7 , %b7
459 %r8 = mul i16 %a8 , %b8
460 %r9 = mul i16 %a9 , %b9
461 %r10 = mul i16 %a10, %b10
462 %r11 = mul i16 %a11, %b11
463 %r12 = mul i16 %a12, %b12
464 %r13 = mul i16 %a13, %b13
465 %r14 = mul i16 %a14, %b14
466 %r15 = mul i16 %a15, %b15
467 %r16 = mul i16 %a16, %b16
468 %r17 = mul i16 %a17, %b17
469 %r18 = mul i16 %a18, %b18
470 %r19 = mul i16 %a19, %b19
471 %r20 = mul i16 %a20, %b20
472 %r21 = mul i16 %a21, %b21
473 %r22 = mul i16 %a22, %b22
474 %r23 = mul i16 %a23, %b23
475 %r24 = mul i16 %a24, %b24
476 %r25 = mul i16 %a25, %b25
477 %r26 = mul i16 %a26, %b26
478 %r27 = mul i16 %a27, %b27
479 %r28 = mul i16 %a28, %b28
480 %r29 = mul i16 %a29, %b29
481 %r30 = mul i16 %a30, %b30
482 %r31 = mul i16 %a31, %b31
483 store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 0 ), align 2
484 store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 1 ), align 2
485 store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 2 ), align 2
486 store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 3 ), align 2
487 store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 4 ), align 2
488 store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 5 ), align 2
489 store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 6 ), align 2
490 store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 7 ), align 2
491 store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8 ), align 2
492 store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 9 ), align 2
493 store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 10), align 2
494 store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 11), align 2
495 store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 12), align 2
496 store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 13), align 2
497 store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 14), align 2
498 store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 15), align 2
499 store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2
500 store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 17), align 2
501 store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 18), align 2
502 store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 19), align 2
503 store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 20), align 2
504 store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 21), align 2
505 store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 22), align 2
506 store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 23), align 2
507 store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2
508 store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 25), align 2
509 store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 26), align 2
510 store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 27), align 2
511 store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 28), align 2
512 store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 29), align 2
513 store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 30), align 2
514 store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 31), align 2
518 define void @mul_v64i8() {
519 ; SSE-LABEL: @mul_v64i8(
520 ; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1
521 ; SSE-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1
522 ; SSE-NEXT: [[TMP3:%.*]] = mul <16 x i8> [[TMP1]], [[TMP2]]
523 ; SSE-NEXT: store <16 x i8> [[TMP3]], ptr @c8, align 1
524 ; SSE-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1
525 ; SSE-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1
526 ; SSE-NEXT: [[TMP6:%.*]] = mul <16 x i8> [[TMP4]], [[TMP5]]
527 ; SSE-NEXT: store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1
528 ; SSE-NEXT: [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1
529 ; SSE-NEXT: [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1
530 ; SSE-NEXT: [[TMP9:%.*]] = mul <16 x i8> [[TMP7]], [[TMP8]]
531 ; SSE-NEXT: store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1
532 ; SSE-NEXT: [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1
533 ; SSE-NEXT: [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1
534 ; SSE-NEXT: [[TMP12:%.*]] = mul <16 x i8> [[TMP10]], [[TMP11]]
535 ; SSE-NEXT: store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1
538 ; SLM-LABEL: @mul_v64i8(
539 ; SLM-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1
540 ; SLM-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1
541 ; SLM-NEXT: [[TMP3:%.*]] = mul <16 x i8> [[TMP1]], [[TMP2]]
542 ; SLM-NEXT: store <16 x i8> [[TMP3]], ptr @c8, align 1
543 ; SLM-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1
544 ; SLM-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1
545 ; SLM-NEXT: [[TMP6:%.*]] = mul <16 x i8> [[TMP4]], [[TMP5]]
546 ; SLM-NEXT: store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1
547 ; SLM-NEXT: [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1
548 ; SLM-NEXT: [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1
549 ; SLM-NEXT: [[TMP9:%.*]] = mul <16 x i8> [[TMP7]], [[TMP8]]
550 ; SLM-NEXT: store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1
551 ; SLM-NEXT: [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1
552 ; SLM-NEXT: [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1
553 ; SLM-NEXT: [[TMP12:%.*]] = mul <16 x i8> [[TMP10]], [[TMP11]]
554 ; SLM-NEXT: store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1
557 ; AVX128-LABEL: @mul_v64i8(
558 ; AVX128-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1
559 ; AVX128-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1
560 ; AVX128-NEXT: [[TMP3:%.*]] = mul <16 x i8> [[TMP1]], [[TMP2]]
561 ; AVX128-NEXT: store <16 x i8> [[TMP3]], ptr @c8, align 1
562 ; AVX128-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1
563 ; AVX128-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1
564 ; AVX128-NEXT: [[TMP6:%.*]] = mul <16 x i8> [[TMP4]], [[TMP5]]
565 ; AVX128-NEXT: store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1
566 ; AVX128-NEXT: [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1
567 ; AVX128-NEXT: [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1
568 ; AVX128-NEXT: [[TMP9:%.*]] = mul <16 x i8> [[TMP7]], [[TMP8]]
569 ; AVX128-NEXT: store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1
570 ; AVX128-NEXT: [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1
571 ; AVX128-NEXT: [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1
572 ; AVX128-NEXT: [[TMP12:%.*]] = mul <16 x i8> [[TMP10]], [[TMP11]]
573 ; AVX128-NEXT: store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1
574 ; AVX128-NEXT: ret void
576 ; AVX256-LABEL: @mul_v64i8(
577 ; AVX256-NEXT: [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 1
578 ; AVX256-NEXT: [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 1
579 ; AVX256-NEXT: [[TMP3:%.*]] = mul <32 x i8> [[TMP1]], [[TMP2]]
580 ; AVX256-NEXT: store <32 x i8> [[TMP3]], ptr @c8, align 1
581 ; AVX256-NEXT: [[TMP4:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1
582 ; AVX256-NEXT: [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1
583 ; AVX256-NEXT: [[TMP6:%.*]] = mul <32 x i8> [[TMP4]], [[TMP5]]
584 ; AVX256-NEXT: store <32 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1
585 ; AVX256-NEXT: ret void
587 ; AVX512-LABEL: @mul_v64i8(
588 ; AVX512-NEXT: [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1
589 ; AVX512-NEXT: [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1
590 ; AVX512-NEXT: [[TMP3:%.*]] = mul <64 x i8> [[TMP1]], [[TMP2]]
591 ; AVX512-NEXT: store <64 x i8> [[TMP3]], ptr @c8, align 1
592 ; AVX512-NEXT: ret void
594 %a0 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 1
595 %a1 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 1
596 %a2 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 1
597 %a3 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 1
598 %a4 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 1
599 %a5 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 1
600 %a6 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 1
601 %a7 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 1
602 %a8 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 1
603 %a9 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 1
604 %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 1
605 %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 1
606 %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 1
607 %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 1
608 %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 1
609 %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 1
610 %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1
611 %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 1
612 %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 1
613 %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 1
614 %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 1
615 %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 1
616 %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 1
617 %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 1
618 %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 1
619 %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 1
620 %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 1
621 %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 1
622 %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 1
623 %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 1
624 %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 1
625 %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 1
626 %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1
627 %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 1
628 %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 1
629 %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 1
630 %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 1
631 %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 1
632 %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 1
633 %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 1
634 %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 1
635 %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 1
636 %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 1
637 %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 1
638 %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 1
639 %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 1
640 %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 1
641 %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 1
642 %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1
643 %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 1
644 %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 1
645 %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 1
646 %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 1
647 %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 1
648 %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 1
649 %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 1
650 %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 1
651 %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 1
652 %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 1
653 %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 1
654 %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 1
655 %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 1
656 %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 1
657 %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 1
658 %b0 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 1
659 %b1 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 1
660 %b2 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 1
661 %b3 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 1
662 %b4 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 1
663 %b5 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 1
664 %b6 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 1
665 %b7 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 1
666 %b8 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 1
667 %b9 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 1
668 %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 1
669 %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 1
670 %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 1
671 %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 1
672 %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 1
673 %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 1
674 %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1
675 %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 1
676 %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 1
677 %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 1
678 %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 1
679 %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 1
680 %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 1
681 %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 1
682 %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 1
683 %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 1
684 %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 1
685 %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 1
686 %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 1
687 %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 1
688 %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 1
689 %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 1
690 %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1
691 %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 1
692 %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 1
693 %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 1
694 %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 1
695 %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 1
696 %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 1
697 %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 1
698 %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 1
699 %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 1
700 %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 1
701 %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 1
702 %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 1
703 %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 1
704 %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 1
705 %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 1
706 %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1
707 %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 1
708 %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 1
709 %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 1
710 %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 1
711 %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 1
712 %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 1
713 %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 1
714 %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 1
715 %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 1
716 %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 1
717 %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 1
718 %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 1
719 %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 1
720 %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 1
721 %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 1
722 %r0 = mul i8 %a0 , %b0
723 %r1 = mul i8 %a1 , %b1
724 %r2 = mul i8 %a2 , %b2
725 %r3 = mul i8 %a3 , %b3
726 %r4 = mul i8 %a4 , %b4
727 %r5 = mul i8 %a5 , %b5
728 %r6 = mul i8 %a6 , %b6
729 %r7 = mul i8 %a7 , %b7
730 %r8 = mul i8 %a8 , %b8
731 %r9 = mul i8 %a9 , %b9
732 %r10 = mul i8 %a10, %b10
733 %r11 = mul i8 %a11, %b11
734 %r12 = mul i8 %a12, %b12
735 %r13 = mul i8 %a13, %b13
736 %r14 = mul i8 %a14, %b14
737 %r15 = mul i8 %a15, %b15
738 %r16 = mul i8 %a16, %b16
739 %r17 = mul i8 %a17, %b17
740 %r18 = mul i8 %a18, %b18
741 %r19 = mul i8 %a19, %b19
742 %r20 = mul i8 %a20, %b20
743 %r21 = mul i8 %a21, %b21
744 %r22 = mul i8 %a22, %b22
745 %r23 = mul i8 %a23, %b23
746 %r24 = mul i8 %a24, %b24
747 %r25 = mul i8 %a25, %b25
748 %r26 = mul i8 %a26, %b26
749 %r27 = mul i8 %a27, %b27
750 %r28 = mul i8 %a28, %b28
751 %r29 = mul i8 %a29, %b29
752 %r30 = mul i8 %a30, %b30
753 %r31 = mul i8 %a31, %b31
754 %r32 = mul i8 %a32, %b32
755 %r33 = mul i8 %a33, %b33
756 %r34 = mul i8 %a34, %b34
757 %r35 = mul i8 %a35, %b35
758 %r36 = mul i8 %a36, %b36
759 %r37 = mul i8 %a37, %b37
760 %r38 = mul i8 %a38, %b38
761 %r39 = mul i8 %a39, %b39
762 %r40 = mul i8 %a40, %b40
763 %r41 = mul i8 %a41, %b41
764 %r42 = mul i8 %a42, %b42
765 %r43 = mul i8 %a43, %b43
766 %r44 = mul i8 %a44, %b44
767 %r45 = mul i8 %a45, %b45
768 %r46 = mul i8 %a46, %b46
769 %r47 = mul i8 %a47, %b47
770 %r48 = mul i8 %a48, %b48
771 %r49 = mul i8 %a49, %b49
772 %r50 = mul i8 %a50, %b50
773 %r51 = mul i8 %a51, %b51
774 %r52 = mul i8 %a52, %b52
775 %r53 = mul i8 %a53, %b53
776 %r54 = mul i8 %a54, %b54
777 %r55 = mul i8 %a55, %b55
778 %r56 = mul i8 %a56, %b56
779 %r57 = mul i8 %a57, %b57
780 %r58 = mul i8 %a58, %b58
781 %r59 = mul i8 %a59, %b59
782 %r60 = mul i8 %a60, %b60
783 %r61 = mul i8 %a61, %b61
784 %r62 = mul i8 %a62, %b62
785 %r63 = mul i8 %a63, %b63
786 store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 0 ), align 1
787 store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 1 ), align 1
788 store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 2 ), align 1
789 store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 3 ), align 1
790 store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 4 ), align 1
791 store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 5 ), align 1
792 store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 6 ), align 1
793 store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 7 ), align 1
794 store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 8 ), align 1
795 store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 9 ), align 1
796 store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 10), align 1
797 store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 11), align 1
798 store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 12), align 1
799 store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 13), align 1
800 store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 14), align 1
801 store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 15), align 1
802 store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1
803 store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 17), align 1
804 store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 18), align 1
805 store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 19), align 1
806 store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 20), align 1
807 store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 21), align 1
808 store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 22), align 1
809 store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 23), align 1
810 store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 24), align 1
811 store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 25), align 1
812 store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 26), align 1
813 store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 27), align 1
814 store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 28), align 1
815 store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 29), align 1
816 store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 30), align 1
817 store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 31), align 1
818 store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1
819 store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 33), align 1
820 store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 34), align 1
821 store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 35), align 1
822 store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 36), align 1
823 store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 37), align 1
824 store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 38), align 1
825 store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 39), align 1
826 store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 40), align 1
827 store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 41), align 1
828 store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 42), align 1
829 store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 43), align 1
830 store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 44), align 1
831 store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 45), align 1
832 store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 46), align 1
833 store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 47), align 1
834 store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1
835 store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 49), align 1
836 store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 50), align 1
837 store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 51), align 1
838 store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 52), align 1
839 store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 53), align 1
840 store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 54), align 1
841 store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 55), align 1
842 store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 56), align 1
843 store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 57), align 1
844 store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 58), align 1
845 store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 59), align 1
846 store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 60), align 1
847 store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 61), align 1
848 store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 62), align 1
849 store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 63), align 1