fabs.ll
21.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -mtriple=x86_64-unknown -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=SSE
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX256
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=bdver1 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX256
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX256
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skylake-avx512 -mattr=-prefer-256-bit -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX512
; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skylake-avx512 -mattr=+prefer-256-bit -basicaa -slp-vectorizer -S | FileCheck %s --check-prefix=CHECK --check-prefix=AVX --check-prefix=AVX256
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
@src64 = common global [8 x double] zeroinitializer, align 64
@src32 = common global [16 x float] zeroinitializer, align 64
@dst64 = common global [8 x double] zeroinitializer, align 64
@dst32 = common global [16 x float] zeroinitializer, align 64
declare float @llvm.fabs.f32(float)
declare double @llvm.fabs.f64(double)
;
; FABS
;
define void @fabs_2f64() #0 {
; CHECK-LABEL: @fabs_2f64(
; CHECK-NEXT: [[TMP1:%.*]] = load <2 x double>, <2 x double>* bitcast ([8 x double]* @src64 to <2 x double>*), align 8
; CHECK-NEXT: [[TMP2:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP1]])
; CHECK-NEXT: store <2 x double> [[TMP2]], <2 x double>* bitcast ([8 x double]* @dst64 to <2 x double>*), align 8
; CHECK-NEXT: ret void
;
%a0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
%a1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
%fabs0 = call double @llvm.fabs.f64(double %a0)
%fabs1 = call double @llvm.fabs.f64(double %a1)
store double %fabs0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
store double %fabs1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
ret void
}
define void @fabs_4f64() #0 {
; SSE-LABEL: @fabs_4f64(
; SSE-NEXT: [[TMP1:%.*]] = load <2 x double>, <2 x double>* bitcast ([8 x double]* @src64 to <2 x double>*), align 8
; SSE-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 8
; SSE-NEXT: [[TMP3:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP2]])
; SSE-NEXT: store <2 x double> [[TMP3]], <2 x double>* bitcast ([8 x double]* @dst64 to <2 x double>*), align 8
; SSE-NEXT: store <2 x double> [[TMP4]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 8
; SSE-NEXT: ret void
;
; AVX-LABEL: @fabs_4f64(
; AVX-NEXT: [[TMP1:%.*]] = load <4 x double>, <4 x double>* bitcast ([8 x double]* @src64 to <4 x double>*), align 8
; AVX-NEXT: [[TMP2:%.*]] = call <4 x double> @llvm.fabs.v4f64(<4 x double> [[TMP1]])
; AVX-NEXT: store <4 x double> [[TMP2]], <4 x double>* bitcast ([8 x double]* @dst64 to <4 x double>*), align 8
; AVX-NEXT: ret void
;
%a0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 8
%a1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 8
%a2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 8
%a3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 8
%fabs0 = call double @llvm.fabs.f64(double %a0)
%fabs1 = call double @llvm.fabs.f64(double %a1)
%fabs2 = call double @llvm.fabs.f64(double %a2)
%fabs3 = call double @llvm.fabs.f64(double %a3)
store double %fabs0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 8
store double %fabs1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 8
store double %fabs2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 8
store double %fabs3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 8
ret void
}
define void @fabs_8f64() #0 {
; SSE-LABEL: @fabs_8f64(
; SSE-NEXT: [[TMP1:%.*]] = load <2 x double>, <2 x double>* bitcast ([8 x double]* @src64 to <2 x double>*), align 4
; SSE-NEXT: [[TMP2:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2) to <2 x double>*), align 4
; SSE-NEXT: [[TMP3:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <2 x double>*), align 4
; SSE-NEXT: [[TMP4:%.*]] = load <2 x double>, <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6) to <2 x double>*), align 4
; SSE-NEXT: [[TMP5:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP1]])
; SSE-NEXT: [[TMP6:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP2]])
; SSE-NEXT: [[TMP7:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP3]])
; SSE-NEXT: [[TMP8:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[TMP4]])
; SSE-NEXT: store <2 x double> [[TMP5]], <2 x double>* bitcast ([8 x double]* @dst64 to <2 x double>*), align 4
; SSE-NEXT: store <2 x double> [[TMP6]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2) to <2 x double>*), align 4
; SSE-NEXT: store <2 x double> [[TMP7]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <2 x double>*), align 4
; SSE-NEXT: store <2 x double> [[TMP8]], <2 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6) to <2 x double>*), align 4
; SSE-NEXT: ret void
;
; AVX256-LABEL: @fabs_8f64(
; AVX256-NEXT: [[TMP1:%.*]] = load <4 x double>, <4 x double>* bitcast ([8 x double]* @src64 to <4 x double>*), align 4
; AVX256-NEXT: [[TMP2:%.*]] = load <4 x double>, <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4) to <4 x double>*), align 4
; AVX256-NEXT: [[TMP3:%.*]] = call <4 x double> @llvm.fabs.v4f64(<4 x double> [[TMP1]])
; AVX256-NEXT: [[TMP4:%.*]] = call <4 x double> @llvm.fabs.v4f64(<4 x double> [[TMP2]])
; AVX256-NEXT: store <4 x double> [[TMP3]], <4 x double>* bitcast ([8 x double]* @dst64 to <4 x double>*), align 4
; AVX256-NEXT: store <4 x double> [[TMP4]], <4 x double>* bitcast (double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4) to <4 x double>*), align 4
; AVX256-NEXT: ret void
;
; AVX512-LABEL: @fabs_8f64(
; AVX512-NEXT: [[TMP1:%.*]] = load <8 x double>, <8 x double>* bitcast ([8 x double]* @src64 to <8 x double>*), align 4
; AVX512-NEXT: [[TMP2:%.*]] = call <8 x double> @llvm.fabs.v8f64(<8 x double> [[TMP1]])
; AVX512-NEXT: store <8 x double> [[TMP2]], <8 x double>* bitcast ([8 x double]* @dst64 to <8 x double>*), align 4
; AVX512-NEXT: ret void
;
%a0 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 0), align 4
%a1 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 1), align 4
%a2 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 2), align 4
%a3 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 3), align 4
%a4 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 4), align 4
%a5 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 5), align 4
%a6 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 6), align 4
%a7 = load double, double* getelementptr inbounds ([8 x double], [8 x double]* @src64, i32 0, i64 7), align 4
%fabs0 = call double @llvm.fabs.f64(double %a0)
%fabs1 = call double @llvm.fabs.f64(double %a1)
%fabs2 = call double @llvm.fabs.f64(double %a2)
%fabs3 = call double @llvm.fabs.f64(double %a3)
%fabs4 = call double @llvm.fabs.f64(double %a4)
%fabs5 = call double @llvm.fabs.f64(double %a5)
%fabs6 = call double @llvm.fabs.f64(double %a6)
%fabs7 = call double @llvm.fabs.f64(double %a7)
store double %fabs0, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 0), align 4
store double %fabs1, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 1), align 4
store double %fabs2, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 2), align 4
store double %fabs3, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 3), align 4
store double %fabs4, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 4), align 4
store double %fabs5, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 5), align 4
store double %fabs6, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 6), align 4
store double %fabs7, double* getelementptr inbounds ([8 x double], [8 x double]* @dst64, i32 0, i64 7), align 4
ret void
}
define void @fabs_4f32() #0 {
; CHECK-LABEL: @fabs_4f32(
; CHECK-NEXT: [[TMP1:%.*]] = load <4 x float>, <4 x float>* bitcast ([16 x float]* @src32 to <4 x float>*), align 4
; CHECK-NEXT: [[TMP2:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP1]])
; CHECK-NEXT: store <4 x float> [[TMP2]], <4 x float>* bitcast ([16 x float]* @dst32 to <4 x float>*), align 4
; CHECK-NEXT: ret void
;
%a0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
%a1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
%a2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
%a3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
%fabs0 = call float @llvm.fabs.f32(float %a0)
%fabs1 = call float @llvm.fabs.f32(float %a1)
%fabs2 = call float @llvm.fabs.f32(float %a2)
%fabs3 = call float @llvm.fabs.f32(float %a3)
store float %fabs0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
store float %fabs1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
store float %fabs2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
store float %fabs3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
ret void
}
define void @fabs_8f32() #0 {
; SSE-LABEL: @fabs_8f32(
; SSE-NEXT: [[TMP1:%.*]] = load <4 x float>, <4 x float>* bitcast ([16 x float]* @src32 to <4 x float>*), align 4
; SSE-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
; SSE-NEXT: [[TMP3:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP1]])
; SSE-NEXT: [[TMP4:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP2]])
; SSE-NEXT: store <4 x float> [[TMP3]], <4 x float>* bitcast ([16 x float]* @dst32 to <4 x float>*), align 4
; SSE-NEXT: store <4 x float> [[TMP4]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
; SSE-NEXT: ret void
;
; AVX-LABEL: @fabs_8f32(
; AVX-NEXT: [[TMP1:%.*]] = load <8 x float>, <8 x float>* bitcast ([16 x float]* @src32 to <8 x float>*), align 4
; AVX-NEXT: [[TMP2:%.*]] = call <8 x float> @llvm.fabs.v8f32(<8 x float> [[TMP1]])
; AVX-NEXT: store <8 x float> [[TMP2]], <8 x float>* bitcast ([16 x float]* @dst32 to <8 x float>*), align 4
; AVX-NEXT: ret void
;
%a0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
%a1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
%a2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
%a3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
%a4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
%a5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
%a6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
%a7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
%fabs0 = call float @llvm.fabs.f32(float %a0)
%fabs1 = call float @llvm.fabs.f32(float %a1)
%fabs2 = call float @llvm.fabs.f32(float %a2)
%fabs3 = call float @llvm.fabs.f32(float %a3)
%fabs4 = call float @llvm.fabs.f32(float %a4)
%fabs5 = call float @llvm.fabs.f32(float %a5)
%fabs6 = call float @llvm.fabs.f32(float %a6)
%fabs7 = call float @llvm.fabs.f32(float %a7)
store float %fabs0, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
store float %fabs1, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
store float %fabs2, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
store float %fabs3, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
store float %fabs4, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
store float %fabs5, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
store float %fabs6, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
store float %fabs7, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
ret void
}
define void @fabs_16f32() #0 {
; SSE-LABEL: @fabs_16f32(
; SSE-NEXT: [[TMP1:%.*]] = load <4 x float>, <4 x float>* bitcast ([16 x float]* @src32 to <4 x float>*), align 4
; SSE-NEXT: [[TMP2:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4) to <4 x float>*), align 4
; SSE-NEXT: [[TMP3:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <4 x float>*), align 4
; SSE-NEXT: [[TMP4:%.*]] = load <4 x float>, <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12) to <4 x float>*), align 4
; SSE-NEXT: [[TMP5:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP1]])
; SSE-NEXT: [[TMP6:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP2]])
; SSE-NEXT: [[TMP7:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP3]])
; SSE-NEXT: [[TMP8:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[TMP4]])
; SSE-NEXT: store <4 x float> [[TMP5]], <4 x float>* bitcast ([16 x float]* @dst32 to <4 x float>*), align 4
; SSE-NEXT: store <4 x float> [[TMP6]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4) to <4 x float>*), align 4
; SSE-NEXT: store <4 x float> [[TMP7]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <4 x float>*), align 4
; SSE-NEXT: store <4 x float> [[TMP8]], <4 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12) to <4 x float>*), align 4
; SSE-NEXT: ret void
;
; AVX256-LABEL: @fabs_16f32(
; AVX256-NEXT: [[TMP1:%.*]] = load <8 x float>, <8 x float>* bitcast ([16 x float]* @src32 to <8 x float>*), align 4
; AVX256-NEXT: [[TMP2:%.*]] = load <8 x float>, <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8) to <8 x float>*), align 4
; AVX256-NEXT: [[TMP3:%.*]] = call <8 x float> @llvm.fabs.v8f32(<8 x float> [[TMP1]])
; AVX256-NEXT: [[TMP4:%.*]] = call <8 x float> @llvm.fabs.v8f32(<8 x float> [[TMP2]])
; AVX256-NEXT: store <8 x float> [[TMP3]], <8 x float>* bitcast ([16 x float]* @dst32 to <8 x float>*), align 4
; AVX256-NEXT: store <8 x float> [[TMP4]], <8 x float>* bitcast (float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8) to <8 x float>*), align 4
; AVX256-NEXT: ret void
;
; AVX512-LABEL: @fabs_16f32(
; AVX512-NEXT: [[TMP1:%.*]] = load <16 x float>, <16 x float>* bitcast ([16 x float]* @src32 to <16 x float>*), align 4
; AVX512-NEXT: [[TMP2:%.*]] = call <16 x float> @llvm.fabs.v16f32(<16 x float> [[TMP1]])
; AVX512-NEXT: store <16 x float> [[TMP2]], <16 x float>* bitcast ([16 x float]* @dst32 to <16 x float>*), align 4
; AVX512-NEXT: ret void
;
%a0 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 0), align 4
%a1 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 1), align 4
%a2 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 2), align 4
%a3 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 3), align 4
%a4 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 4), align 4
%a5 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 5), align 4
%a6 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 6), align 4
%a7 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 7), align 4
%a8 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 8), align 4
%a9 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 9), align 4
%a10 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 10), align 4
%a11 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 11), align 4
%a12 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 12), align 4
%a13 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 13), align 4
%a14 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 14), align 4
%a15 = load float, float* getelementptr inbounds ([16 x float], [16 x float]* @src32, i32 0, i64 15), align 4
%fabs0 = call float @llvm.fabs.f32(float %a0 )
%fabs1 = call float @llvm.fabs.f32(float %a1 )
%fabs2 = call float @llvm.fabs.f32(float %a2 )
%fabs3 = call float @llvm.fabs.f32(float %a3 )
%fabs4 = call float @llvm.fabs.f32(float %a4 )
%fabs5 = call float @llvm.fabs.f32(float %a5 )
%fabs6 = call float @llvm.fabs.f32(float %a6 )
%fabs7 = call float @llvm.fabs.f32(float %a7 )
%fabs8 = call float @llvm.fabs.f32(float %a8 )
%fabs9 = call float @llvm.fabs.f32(float %a9 )
%fabs10 = call float @llvm.fabs.f32(float %a10)
%fabs11 = call float @llvm.fabs.f32(float %a11)
%fabs12 = call float @llvm.fabs.f32(float %a12)
%fabs13 = call float @llvm.fabs.f32(float %a13)
%fabs14 = call float @llvm.fabs.f32(float %a14)
%fabs15 = call float @llvm.fabs.f32(float %a15)
store float %fabs0 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 0), align 4
store float %fabs1 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 1), align 4
store float %fabs2 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 2), align 4
store float %fabs3 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 3), align 4
store float %fabs4 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 4), align 4
store float %fabs5 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 5), align 4
store float %fabs6 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 6), align 4
store float %fabs7 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 7), align 4
store float %fabs8 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 8), align 4
store float %fabs9 , float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 9), align 4
store float %fabs10, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 10), align 4
store float %fabs11, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 11), align 4
store float %fabs12, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 12), align 4
store float %fabs13, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 13), align 4
store float %fabs14, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 14), align 4
store float %fabs15, float* getelementptr inbounds ([16 x float], [16 x float]* @dst32, i32 0, i64 15), align 4
ret void
}
attributes #0 = { nounwind }