1; RUN: llc -enable-amdgpu-aa=0 -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck %s
2; RUN: llc -enable-amdgpu-aa=0 -march=amdgcn -mcpu=tonga -misched=gcn-minreg -verify-machineinstrs < %s | FileCheck %s
3; RUN: llc -enable-amdgpu-aa=0 -march=amdgcn -mcpu=tonga -misched=gcn-max-occupancy-experimental -verify-machineinstrs < %s | FileCheck %s
4
5; We expect a two digit VGPR usage here, not a three digit.
6; CHECK: NumVgprs: {{[0-9][0-9]$}}
7
8define amdgpu_kernel void @load_fma_store(float addrspace(3)* nocapture readonly %arg, float addrspace(1)* nocapture %arg1) {
9bb:
10  %tmp = getelementptr inbounds float, float addrspace(3)* %arg, i32 1
11  %tmp2 = load float, float addrspace(3)* %tmp, align 4
12  %tmp3 = getelementptr inbounds float, float addrspace(3)* %arg, i32 2
13  %tmp4 = load float, float addrspace(3)* %tmp3, align 4
14  %tmp5 = getelementptr inbounds float, float addrspace(3)* %arg, i32 3
15  %tmp6 = load float, float addrspace(3)* %tmp5, align 4
16  %tmp7 = tail call float @llvm.fmuladd.f32(float %tmp2, float %tmp4, float %tmp6)
17  %tmp8 = getelementptr inbounds float, float addrspace(3)* %arg, i32 5
18  %tmp9 = load float, float addrspace(3)* %tmp8, align 4
19  %tmp10 = getelementptr inbounds float, float addrspace(3)* %arg, i32 6
20  %tmp11 = load float, float addrspace(3)* %tmp10, align 4
21  %tmp12 = getelementptr inbounds float, float addrspace(3)* %arg, i32 7
22  %tmp13 = load float, float addrspace(3)* %tmp12, align 4
23  %tmp14 = tail call float @llvm.fmuladd.f32(float %tmp9, float %tmp11, float %tmp13)
24  %tmp15 = getelementptr inbounds float, float addrspace(3)* %arg, i32 9
25  %tmp16 = load float, float addrspace(3)* %tmp15, align 4
26  %tmp17 = getelementptr inbounds float, float addrspace(3)* %arg, i32 10
27  %tmp18 = load float, float addrspace(3)* %tmp17, align 4
28  %tmp19 = getelementptr inbounds float, float addrspace(3)* %arg, i32 11
29  %tmp20 = load float, float addrspace(3)* %tmp19, align 4
30  %tmp21 = tail call float @llvm.fmuladd.f32(float %tmp16, float %tmp18, float %tmp20)
31  %tmp22 = getelementptr inbounds float, float addrspace(3)* %arg, i32 13
32  %tmp23 = load float, float addrspace(3)* %tmp22, align 4
33  %tmp24 = getelementptr inbounds float, float addrspace(3)* %arg, i32 14
34  %tmp25 = load float, float addrspace(3)* %tmp24, align 4
35  %tmp26 = getelementptr inbounds float, float addrspace(3)* %arg, i32 15
36  %tmp27 = load float, float addrspace(3)* %tmp26, align 4
37  %tmp28 = tail call float @llvm.fmuladd.f32(float %tmp23, float %tmp25, float %tmp27)
38  %tmp29 = getelementptr inbounds float, float addrspace(3)* %arg, i32 17
39  %tmp30 = load float, float addrspace(3)* %tmp29, align 4
40  %tmp31 = getelementptr inbounds float, float addrspace(3)* %arg, i32 18
41  %tmp32 = load float, float addrspace(3)* %tmp31, align 4
42  %tmp33 = getelementptr inbounds float, float addrspace(3)* %arg, i32 19
43  %tmp34 = load float, float addrspace(3)* %tmp33, align 4
44  %tmp35 = tail call float @llvm.fmuladd.f32(float %tmp30, float %tmp32, float %tmp34)
45  %tmp36 = getelementptr inbounds float, float addrspace(3)* %arg, i32 21
46  %tmp37 = load float, float addrspace(3)* %tmp36, align 4
47  %tmp38 = getelementptr inbounds float, float addrspace(3)* %arg, i32 22
48  %tmp39 = load float, float addrspace(3)* %tmp38, align 4
49  %tmp40 = getelementptr inbounds float, float addrspace(3)* %arg, i32 23
50  %tmp41 = load float, float addrspace(3)* %tmp40, align 4
51  %tmp42 = tail call float @llvm.fmuladd.f32(float %tmp37, float %tmp39, float %tmp41)
52  %tmp43 = getelementptr inbounds float, float addrspace(3)* %arg, i32 25
53  %tmp44 = load float, float addrspace(3)* %tmp43, align 4
54  %tmp45 = getelementptr inbounds float, float addrspace(3)* %arg, i32 26
55  %tmp46 = load float, float addrspace(3)* %tmp45, align 4
56  %tmp47 = getelementptr inbounds float, float addrspace(3)* %arg, i32 27
57  %tmp48 = load float, float addrspace(3)* %tmp47, align 4
58  %tmp49 = tail call float @llvm.fmuladd.f32(float %tmp44, float %tmp46, float %tmp48)
59  %tmp50 = getelementptr inbounds float, float addrspace(3)* %arg, i32 29
60  %tmp51 = load float, float addrspace(3)* %tmp50, align 4
61  %tmp52 = getelementptr inbounds float, float addrspace(3)* %arg, i32 30
62  %tmp53 = load float, float addrspace(3)* %tmp52, align 4
63  %tmp54 = getelementptr inbounds float, float addrspace(3)* %arg, i32 31
64  %tmp55 = load float, float addrspace(3)* %tmp54, align 4
65  %tmp56 = tail call float @llvm.fmuladd.f32(float %tmp51, float %tmp53, float %tmp55)
66  %tmp57 = getelementptr inbounds float, float addrspace(3)* %arg, i32 33
67  %tmp58 = load float, float addrspace(3)* %tmp57, align 4
68  %tmp59 = getelementptr inbounds float, float addrspace(3)* %arg, i32 34
69  %tmp60 = load float, float addrspace(3)* %tmp59, align 4
70  %tmp61 = getelementptr inbounds float, float addrspace(3)* %arg, i32 35
71  %tmp62 = load float, float addrspace(3)* %tmp61, align 4
72  %tmp63 = tail call float @llvm.fmuladd.f32(float %tmp58, float %tmp60, float %tmp62)
73  %tmp64 = getelementptr inbounds float, float addrspace(3)* %arg, i32 37
74  %tmp65 = load float, float addrspace(3)* %tmp64, align 4
75  %tmp66 = getelementptr inbounds float, float addrspace(3)* %arg, i32 38
76  %tmp67 = load float, float addrspace(3)* %tmp66, align 4
77  %tmp68 = getelementptr inbounds float, float addrspace(3)* %arg, i32 39
78  %tmp69 = load float, float addrspace(3)* %tmp68, align 4
79  %tmp70 = tail call float @llvm.fmuladd.f32(float %tmp65, float %tmp67, float %tmp69)
80  %tmp71 = getelementptr inbounds float, float addrspace(3)* %arg, i32 41
81  %tmp72 = load float, float addrspace(3)* %tmp71, align 4
82  %tmp73 = getelementptr inbounds float, float addrspace(3)* %arg, i32 42
83  %tmp74 = load float, float addrspace(3)* %tmp73, align 4
84  %tmp75 = getelementptr inbounds float, float addrspace(3)* %arg, i32 43
85  %tmp76 = load float, float addrspace(3)* %tmp75, align 4
86  %tmp77 = tail call float @llvm.fmuladd.f32(float %tmp72, float %tmp74, float %tmp76)
87  %tmp78 = getelementptr inbounds float, float addrspace(3)* %arg, i32 45
88  %tmp79 = load float, float addrspace(3)* %tmp78, align 4
89  %tmp80 = getelementptr inbounds float, float addrspace(3)* %arg, i32 46
90  %tmp81 = load float, float addrspace(3)* %tmp80, align 4
91  %tmp82 = getelementptr inbounds float, float addrspace(3)* %arg, i32 47
92  %tmp83 = load float, float addrspace(3)* %tmp82, align 4
93  %tmp84 = tail call float @llvm.fmuladd.f32(float %tmp79, float %tmp81, float %tmp83)
94  %tmp85 = getelementptr inbounds float, float addrspace(3)* %arg, i32 49
95  %tmp86 = load float, float addrspace(3)* %tmp85, align 4
96  %tmp87 = getelementptr inbounds float, float addrspace(3)* %arg, i32 50
97  %tmp88 = load float, float addrspace(3)* %tmp87, align 4
98  %tmp89 = getelementptr inbounds float, float addrspace(3)* %arg, i32 51
99  %tmp90 = load float, float addrspace(3)* %tmp89, align 4
100  %tmp91 = tail call float @llvm.fmuladd.f32(float %tmp86, float %tmp88, float %tmp90)
101  %tmp92 = getelementptr inbounds float, float addrspace(3)* %arg, i32 53
102  %tmp93 = load float, float addrspace(3)* %tmp92, align 4
103  %tmp94 = getelementptr inbounds float, float addrspace(3)* %arg, i32 54
104  %tmp95 = load float, float addrspace(3)* %tmp94, align 4
105  %tmp96 = getelementptr inbounds float, float addrspace(3)* %arg, i32 55
106  %tmp97 = load float, float addrspace(3)* %tmp96, align 4
107  %tmp98 = tail call float @llvm.fmuladd.f32(float %tmp93, float %tmp95, float %tmp97)
108  %tmp99 = getelementptr inbounds float, float addrspace(3)* %arg, i32 57
109  %tmp100 = load float, float addrspace(3)* %tmp99, align 4
110  %tmp101 = getelementptr inbounds float, float addrspace(3)* %arg, i32 58
111  %tmp102 = load float, float addrspace(3)* %tmp101, align 4
112  %tmp103 = getelementptr inbounds float, float addrspace(3)* %arg, i32 59
113  %tmp104 = load float, float addrspace(3)* %tmp103, align 4
114  %tmp105 = tail call float @llvm.fmuladd.f32(float %tmp100, float %tmp102, float %tmp104)
115  %tmp106 = getelementptr inbounds float, float addrspace(3)* %arg, i32 61
116  %tmp107 = load float, float addrspace(3)* %tmp106, align 4
117  %tmp108 = getelementptr inbounds float, float addrspace(3)* %arg, i32 62
118  %tmp109 = load float, float addrspace(3)* %tmp108, align 4
119  %tmp110 = getelementptr inbounds float, float addrspace(3)* %arg, i32 63
120  %tmp111 = load float, float addrspace(3)* %tmp110, align 4
121  %tmp112 = tail call float @llvm.fmuladd.f32(float %tmp107, float %tmp109, float %tmp111)
122  %tmp113 = getelementptr inbounds float, float addrspace(3)* %arg, i32 65
123  %tmp114 = load float, float addrspace(3)* %tmp113, align 4
124  %tmp115 = getelementptr inbounds float, float addrspace(3)* %arg, i32 66
125  %tmp116 = load float, float addrspace(3)* %tmp115, align 4
126  %tmp117 = getelementptr inbounds float, float addrspace(3)* %arg, i32 67
127  %tmp118 = load float, float addrspace(3)* %tmp117, align 4
128  %tmp119 = tail call float @llvm.fmuladd.f32(float %tmp114, float %tmp116, float %tmp118)
129  %tmp120 = getelementptr inbounds float, float addrspace(3)* %arg, i32 69
130  %tmp121 = load float, float addrspace(3)* %tmp120, align 4
131  %tmp122 = getelementptr inbounds float, float addrspace(3)* %arg, i32 70
132  %tmp123 = load float, float addrspace(3)* %tmp122, align 4
133  %tmp124 = getelementptr inbounds float, float addrspace(3)* %arg, i32 71
134  %tmp125 = load float, float addrspace(3)* %tmp124, align 4
135  %tmp126 = tail call float @llvm.fmuladd.f32(float %tmp121, float %tmp123, float %tmp125)
136  %tmp127 = getelementptr inbounds float, float addrspace(3)* %arg, i32 73
137  %tmp128 = load float, float addrspace(3)* %tmp127, align 4
138  %tmp129 = getelementptr inbounds float, float addrspace(3)* %arg, i32 74
139  %tmp130 = load float, float addrspace(3)* %tmp129, align 4
140  %tmp131 = getelementptr inbounds float, float addrspace(3)* %arg, i32 75
141  %tmp132 = load float, float addrspace(3)* %tmp131, align 4
142  %tmp133 = tail call float @llvm.fmuladd.f32(float %tmp128, float %tmp130, float %tmp132)
143  %tmp134 = getelementptr inbounds float, float addrspace(3)* %arg, i32 77
144  %tmp135 = load float, float addrspace(3)* %tmp134, align 4
145  %tmp136 = getelementptr inbounds float, float addrspace(3)* %arg, i32 78
146  %tmp137 = load float, float addrspace(3)* %tmp136, align 4
147  %tmp138 = getelementptr inbounds float, float addrspace(3)* %arg, i32 79
148  %tmp139 = load float, float addrspace(3)* %tmp138, align 4
149  %tmp140 = tail call float @llvm.fmuladd.f32(float %tmp135, float %tmp137, float %tmp139)
150  %tmp141 = getelementptr inbounds float, float addrspace(3)* %arg, i32 81
151  %tmp142 = load float, float addrspace(3)* %tmp141, align 4
152  %tmp143 = getelementptr inbounds float, float addrspace(3)* %arg, i32 82
153  %tmp144 = load float, float addrspace(3)* %tmp143, align 4
154  %tmp145 = getelementptr inbounds float, float addrspace(3)* %arg, i32 83
155  %tmp146 = load float, float addrspace(3)* %tmp145, align 4
156  %tmp147 = tail call float @llvm.fmuladd.f32(float %tmp142, float %tmp144, float %tmp146)
157  %tmp148 = getelementptr inbounds float, float addrspace(3)* %arg, i32 85
158  %tmp149 = load float, float addrspace(3)* %tmp148, align 4
159  %tmp150 = getelementptr inbounds float, float addrspace(3)* %arg, i32 86
160  %tmp151 = load float, float addrspace(3)* %tmp150, align 4
161  %tmp152 = getelementptr inbounds float, float addrspace(3)* %arg, i32 87
162  %tmp153 = load float, float addrspace(3)* %tmp152, align 4
163  %tmp154 = tail call float @llvm.fmuladd.f32(float %tmp149, float %tmp151, float %tmp153)
164  %tmp155 = getelementptr inbounds float, float addrspace(3)* %arg, i32 89
165  %tmp156 = load float, float addrspace(3)* %tmp155, align 4
166  %tmp157 = getelementptr inbounds float, float addrspace(3)* %arg, i32 90
167  %tmp158 = load float, float addrspace(3)* %tmp157, align 4
168  %tmp159 = getelementptr inbounds float, float addrspace(3)* %arg, i32 91
169  %tmp160 = load float, float addrspace(3)* %tmp159, align 4
170  %tmp161 = tail call float @llvm.fmuladd.f32(float %tmp156, float %tmp158, float %tmp160)
171  %tmp162 = getelementptr inbounds float, float addrspace(3)* %arg, i32 93
172  %tmp163 = load float, float addrspace(3)* %tmp162, align 4
173  %tmp164 = getelementptr inbounds float, float addrspace(3)* %arg, i32 94
174  %tmp165 = load float, float addrspace(3)* %tmp164, align 4
175  %tmp166 = getelementptr inbounds float, float addrspace(3)* %arg, i32 95
176  %tmp167 = load float, float addrspace(3)* %tmp166, align 4
177  %tmp168 = tail call float @llvm.fmuladd.f32(float %tmp163, float %tmp165, float %tmp167)
178  %tmp169 = getelementptr inbounds float, float addrspace(3)* %arg, i32 97
179  %tmp170 = load float, float addrspace(3)* %tmp169, align 4
180  %tmp171 = getelementptr inbounds float, float addrspace(3)* %arg, i32 98
181  %tmp172 = load float, float addrspace(3)* %tmp171, align 4
182  %tmp173 = getelementptr inbounds float, float addrspace(3)* %arg, i32 99
183  %tmp174 = load float, float addrspace(3)* %tmp173, align 4
184  %tmp175 = tail call float @llvm.fmuladd.f32(float %tmp170, float %tmp172, float %tmp174)
185  %tmp176 = getelementptr inbounds float, float addrspace(3)* %arg, i32 101
186  %tmp177 = load float, float addrspace(3)* %tmp176, align 4
187  %tmp178 = getelementptr inbounds float, float addrspace(3)* %arg, i32 102
188  %tmp179 = load float, float addrspace(3)* %tmp178, align 4
189  %tmp180 = getelementptr inbounds float, float addrspace(3)* %arg, i32 103
190  %tmp181 = load float, float addrspace(3)* %tmp180, align 4
191  %tmp182 = tail call float @llvm.fmuladd.f32(float %tmp177, float %tmp179, float %tmp181)
192  %tmp183 = getelementptr inbounds float, float addrspace(3)* %arg, i32 105
193  %tmp184 = load float, float addrspace(3)* %tmp183, align 4
194  %tmp185 = getelementptr inbounds float, float addrspace(3)* %arg, i32 106
195  %tmp186 = load float, float addrspace(3)* %tmp185, align 4
196  %tmp187 = getelementptr inbounds float, float addrspace(3)* %arg, i32 107
197  %tmp188 = load float, float addrspace(3)* %tmp187, align 4
198  %tmp189 = tail call float @llvm.fmuladd.f32(float %tmp184, float %tmp186, float %tmp188)
199  %tmp190 = getelementptr inbounds float, float addrspace(3)* %arg, i32 109
200  %tmp191 = load float, float addrspace(3)* %tmp190, align 4
201  %tmp192 = getelementptr inbounds float, float addrspace(3)* %arg, i32 110
202  %tmp193 = load float, float addrspace(3)* %tmp192, align 4
203  %tmp194 = getelementptr inbounds float, float addrspace(3)* %arg, i32 111
204  %tmp195 = load float, float addrspace(3)* %tmp194, align 4
205  %tmp196 = tail call float @llvm.fmuladd.f32(float %tmp191, float %tmp193, float %tmp195)
206  %tmp197 = getelementptr inbounds float, float addrspace(3)* %arg, i32 113
207  %tmp198 = load float, float addrspace(3)* %tmp197, align 4
208  %tmp199 = getelementptr inbounds float, float addrspace(3)* %arg, i32 114
209  %tmp200 = load float, float addrspace(3)* %tmp199, align 4
210  %tmp201 = getelementptr inbounds float, float addrspace(3)* %arg, i32 115
211  %tmp202 = load float, float addrspace(3)* %tmp201, align 4
212  %tmp203 = tail call float @llvm.fmuladd.f32(float %tmp198, float %tmp200, float %tmp202)
213  %tmp204 = getelementptr inbounds float, float addrspace(3)* %arg, i32 117
214  %tmp205 = load float, float addrspace(3)* %tmp204, align 4
215  %tmp206 = getelementptr inbounds float, float addrspace(3)* %arg, i32 118
216  %tmp207 = load float, float addrspace(3)* %tmp206, align 4
217  %tmp208 = getelementptr inbounds float, float addrspace(3)* %arg, i32 119
218  %tmp209 = load float, float addrspace(3)* %tmp208, align 4
219  %tmp210 = tail call float @llvm.fmuladd.f32(float %tmp205, float %tmp207, float %tmp209)
220  %tmp211 = getelementptr inbounds float, float addrspace(3)* %arg, i32 121
221  %tmp212 = load float, float addrspace(3)* %tmp211, align 4
222  %tmp213 = getelementptr inbounds float, float addrspace(3)* %arg, i32 122
223  %tmp214 = load float, float addrspace(3)* %tmp213, align 4
224  %tmp215 = getelementptr inbounds float, float addrspace(3)* %arg, i32 123
225  %tmp216 = load float, float addrspace(3)* %tmp215, align 4
226  %tmp217 = tail call float @llvm.fmuladd.f32(float %tmp212, float %tmp214, float %tmp216)
227  %tmp218 = getelementptr inbounds float, float addrspace(3)* %arg, i32 125
228  %tmp219 = load float, float addrspace(3)* %tmp218, align 4
229  %tmp220 = getelementptr inbounds float, float addrspace(3)* %arg, i32 126
230  %tmp221 = load float, float addrspace(3)* %tmp220, align 4
231  %tmp222 = getelementptr inbounds float, float addrspace(3)* %arg, i32 127
232  %tmp223 = load float, float addrspace(3)* %tmp222, align 4
233  %tmp224 = tail call float @llvm.fmuladd.f32(float %tmp219, float %tmp221, float %tmp223)
234  %tmp225 = getelementptr inbounds float, float addrspace(3)* %arg, i32 129
235  %tmp226 = load float, float addrspace(3)* %tmp225, align 4
236  %tmp227 = getelementptr inbounds float, float addrspace(3)* %arg, i32 130
237  %tmp228 = load float, float addrspace(3)* %tmp227, align 4
238  %tmp229 = getelementptr inbounds float, float addrspace(3)* %arg, i32 131
239  %tmp230 = load float, float addrspace(3)* %tmp229, align 4
240  %tmp231 = tail call float @llvm.fmuladd.f32(float %tmp226, float %tmp228, float %tmp230)
241  %tmp232 = getelementptr inbounds float, float addrspace(3)* %arg, i32 133
242  %tmp233 = load float, float addrspace(3)* %tmp232, align 4
243  %tmp234 = getelementptr inbounds float, float addrspace(3)* %arg, i32 134
244  %tmp235 = load float, float addrspace(3)* %tmp234, align 4
245  %tmp236 = getelementptr inbounds float, float addrspace(3)* %arg, i32 135
246  %tmp237 = load float, float addrspace(3)* %tmp236, align 4
247  %tmp238 = tail call float @llvm.fmuladd.f32(float %tmp233, float %tmp235, float %tmp237)
248  %tmp239 = getelementptr inbounds float, float addrspace(3)* %arg, i32 137
249  %tmp240 = load float, float addrspace(3)* %tmp239, align 4
250  %tmp241 = getelementptr inbounds float, float addrspace(3)* %arg, i32 138
251  %tmp242 = load float, float addrspace(3)* %tmp241, align 4
252  %tmp243 = getelementptr inbounds float, float addrspace(3)* %arg, i32 139
253  %tmp244 = load float, float addrspace(3)* %tmp243, align 4
254  %tmp245 = tail call float @llvm.fmuladd.f32(float %tmp240, float %tmp242, float %tmp244)
255  %tmp246 = getelementptr inbounds float, float addrspace(3)* %arg, i32 141
256  %tmp247 = load float, float addrspace(3)* %tmp246, align 4
257  %tmp248 = getelementptr inbounds float, float addrspace(3)* %arg, i32 142
258  %tmp249 = load float, float addrspace(3)* %tmp248, align 4
259  %tmp250 = getelementptr inbounds float, float addrspace(3)* %arg, i32 143
260  %tmp251 = load float, float addrspace(3)* %tmp250, align 4
261  %tmp252 = tail call float @llvm.fmuladd.f32(float %tmp247, float %tmp249, float %tmp251)
262  %tmp253 = getelementptr inbounds float, float addrspace(3)* %arg, i32 145
263  %tmp254 = load float, float addrspace(3)* %tmp253, align 4
264  %tmp255 = getelementptr inbounds float, float addrspace(3)* %arg, i32 146
265  %tmp256 = load float, float addrspace(3)* %tmp255, align 4
266  %tmp257 = getelementptr inbounds float, float addrspace(3)* %arg, i32 147
267  %tmp258 = load float, float addrspace(3)* %tmp257, align 4
268  %tmp259 = tail call float @llvm.fmuladd.f32(float %tmp254, float %tmp256, float %tmp258)
269  %tmp260 = getelementptr inbounds float, float addrspace(3)* %arg, i32 149
270  %tmp261 = load float, float addrspace(3)* %tmp260, align 4
271  %tmp262 = getelementptr inbounds float, float addrspace(3)* %arg, i32 150
272  %tmp263 = load float, float addrspace(3)* %tmp262, align 4
273  %tmp264 = getelementptr inbounds float, float addrspace(3)* %arg, i32 151
274  %tmp265 = load float, float addrspace(3)* %tmp264, align 4
275  %tmp266 = tail call float @llvm.fmuladd.f32(float %tmp261, float %tmp263, float %tmp265)
276  %tmp267 = getelementptr inbounds float, float addrspace(3)* %arg, i32 153
277  %tmp268 = load float, float addrspace(3)* %tmp267, align 4
278  %tmp269 = getelementptr inbounds float, float addrspace(3)* %arg, i32 154
279  %tmp270 = load float, float addrspace(3)* %tmp269, align 4
280  %tmp271 = getelementptr inbounds float, float addrspace(3)* %arg, i32 155
281  %tmp272 = load float, float addrspace(3)* %tmp271, align 4
282  %tmp273 = tail call float @llvm.fmuladd.f32(float %tmp268, float %tmp270, float %tmp272)
283  %tmp274 = getelementptr inbounds float, float addrspace(3)* %arg, i32 157
284  %tmp275 = load float, float addrspace(3)* %tmp274, align 4
285  %tmp276 = getelementptr inbounds float, float addrspace(3)* %arg, i32 158
286  %tmp277 = load float, float addrspace(3)* %tmp276, align 4
287  %tmp278 = getelementptr inbounds float, float addrspace(3)* %arg, i32 159
288  %tmp279 = load float, float addrspace(3)* %tmp278, align 4
289  %tmp280 = tail call float @llvm.fmuladd.f32(float %tmp275, float %tmp277, float %tmp279)
290  %tmp281 = getelementptr inbounds float, float addrspace(3)* %arg, i32 161
291  %tmp282 = load float, float addrspace(3)* %tmp281, align 4
292  %tmp283 = getelementptr inbounds float, float addrspace(3)* %arg, i32 162
293  %tmp284 = load float, float addrspace(3)* %tmp283, align 4
294  %tmp285 = getelementptr inbounds float, float addrspace(3)* %arg, i32 163
295  %tmp286 = load float, float addrspace(3)* %tmp285, align 4
296  %tmp287 = tail call float @llvm.fmuladd.f32(float %tmp282, float %tmp284, float %tmp286)
297  %tmp288 = getelementptr inbounds float, float addrspace(3)* %arg, i32 165
298  %tmp289 = load float, float addrspace(3)* %tmp288, align 4
299  %tmp290 = getelementptr inbounds float, float addrspace(3)* %arg, i32 166
300  %tmp291 = load float, float addrspace(3)* %tmp290, align 4
301  %tmp292 = getelementptr inbounds float, float addrspace(3)* %arg, i32 167
302  %tmp293 = load float, float addrspace(3)* %tmp292, align 4
303  %tmp294 = tail call float @llvm.fmuladd.f32(float %tmp289, float %tmp291, float %tmp293)
304  %tmp295 = getelementptr inbounds float, float addrspace(3)* %arg, i32 169
305  %tmp296 = load float, float addrspace(3)* %tmp295, align 4
306  %tmp297 = getelementptr inbounds float, float addrspace(3)* %arg, i32 170
307  %tmp298 = load float, float addrspace(3)* %tmp297, align 4
308  %tmp299 = getelementptr inbounds float, float addrspace(3)* %arg, i32 171
309  %tmp300 = load float, float addrspace(3)* %tmp299, align 4
310  %tmp301 = tail call float @llvm.fmuladd.f32(float %tmp296, float %tmp298, float %tmp300)
311  %tmp302 = getelementptr inbounds float, float addrspace(3)* %arg, i32 173
312  %tmp303 = load float, float addrspace(3)* %tmp302, align 4
313  %tmp304 = getelementptr inbounds float, float addrspace(3)* %arg, i32 174
314  %tmp305 = load float, float addrspace(3)* %tmp304, align 4
315  %tmp306 = getelementptr inbounds float, float addrspace(3)* %arg, i32 175
316  %tmp307 = load float, float addrspace(3)* %tmp306, align 4
317  %tmp308 = tail call float @llvm.fmuladd.f32(float %tmp303, float %tmp305, float %tmp307)
318  %tmp309 = getelementptr inbounds float, float addrspace(3)* %arg, i32 177
319  %tmp310 = load float, float addrspace(3)* %tmp309, align 4
320  %tmp311 = getelementptr inbounds float, float addrspace(3)* %arg, i32 178
321  %tmp312 = load float, float addrspace(3)* %tmp311, align 4
322  %tmp313 = getelementptr inbounds float, float addrspace(3)* %arg, i32 179
323  %tmp314 = load float, float addrspace(3)* %tmp313, align 4
324  %tmp315 = tail call float @llvm.fmuladd.f32(float %tmp310, float %tmp312, float %tmp314)
325  %tmp316 = getelementptr inbounds float, float addrspace(3)* %arg, i32 181
326  %tmp317 = load float, float addrspace(3)* %tmp316, align 4
327  %tmp318 = getelementptr inbounds float, float addrspace(3)* %arg, i32 182
328  %tmp319 = load float, float addrspace(3)* %tmp318, align 4
329  %tmp320 = getelementptr inbounds float, float addrspace(3)* %arg, i32 183
330  %tmp321 = load float, float addrspace(3)* %tmp320, align 4
331  %tmp322 = tail call float @llvm.fmuladd.f32(float %tmp317, float %tmp319, float %tmp321)
332  %tmp323 = getelementptr inbounds float, float addrspace(3)* %arg, i32 185
333  %tmp324 = load float, float addrspace(3)* %tmp323, align 4
334  %tmp325 = getelementptr inbounds float, float addrspace(3)* %arg, i32 186
335  %tmp326 = load float, float addrspace(3)* %tmp325, align 4
336  %tmp327 = getelementptr inbounds float, float addrspace(3)* %arg, i32 187
337  %tmp328 = load float, float addrspace(3)* %tmp327, align 4
338  %tmp329 = tail call float @llvm.fmuladd.f32(float %tmp324, float %tmp326, float %tmp328)
339  %tmp330 = getelementptr inbounds float, float addrspace(3)* %arg, i32 189
340  %tmp331 = load float, float addrspace(3)* %tmp330, align 4
341  %tmp332 = getelementptr inbounds float, float addrspace(3)* %arg, i32 190
342  %tmp333 = load float, float addrspace(3)* %tmp332, align 4
343  %tmp334 = getelementptr inbounds float, float addrspace(3)* %arg, i32 191
344  %tmp335 = load float, float addrspace(3)* %tmp334, align 4
345  %tmp336 = tail call float @llvm.fmuladd.f32(float %tmp331, float %tmp333, float %tmp335)
346  %tmp337 = getelementptr inbounds float, float addrspace(3)* %arg, i32 193
347  %tmp338 = load float, float addrspace(3)* %tmp337, align 4
348  %tmp339 = getelementptr inbounds float, float addrspace(3)* %arg, i32 194
349  %tmp340 = load float, float addrspace(3)* %tmp339, align 4
350  %tmp341 = getelementptr inbounds float, float addrspace(3)* %arg, i32 195
351  %tmp342 = load float, float addrspace(3)* %tmp341, align 4
352  %tmp343 = tail call float @llvm.fmuladd.f32(float %tmp338, float %tmp340, float %tmp342)
353  %tmp344 = getelementptr inbounds float, float addrspace(3)* %arg, i32 197
354  %tmp345 = load float, float addrspace(3)* %tmp344, align 4
355  %tmp346 = getelementptr inbounds float, float addrspace(3)* %arg, i32 198
356  %tmp347 = load float, float addrspace(3)* %tmp346, align 4
357  %tmp348 = getelementptr inbounds float, float addrspace(3)* %arg, i32 199
358  %tmp349 = load float, float addrspace(3)* %tmp348, align 4
359  %tmp350 = tail call float @llvm.fmuladd.f32(float %tmp345, float %tmp347, float %tmp349)
360  %tmp351 = getelementptr inbounds float, float addrspace(3)* %arg, i32 201
361  %tmp352 = load float, float addrspace(3)* %tmp351, align 4
362  %tmp353 = getelementptr inbounds float, float addrspace(3)* %arg, i32 202
363  %tmp354 = load float, float addrspace(3)* %tmp353, align 4
364  %tmp355 = getelementptr inbounds float, float addrspace(3)* %arg, i32 203
365  %tmp356 = load float, float addrspace(3)* %tmp355, align 4
366  %tmp357 = tail call float @llvm.fmuladd.f32(float %tmp352, float %tmp354, float %tmp356)
367  %tmp358 = getelementptr inbounds float, float addrspace(3)* %arg, i32 205
368  %tmp359 = load float, float addrspace(3)* %tmp358, align 4
369  %tmp360 = getelementptr inbounds float, float addrspace(3)* %arg, i32 206
370  %tmp361 = load float, float addrspace(3)* %tmp360, align 4
371  %tmp362 = getelementptr inbounds float, float addrspace(3)* %arg, i32 207
372  %tmp363 = load float, float addrspace(3)* %tmp362, align 4
373  %tmp364 = tail call float @llvm.fmuladd.f32(float %tmp359, float %tmp361, float %tmp363)
374  %tmp365 = getelementptr inbounds float, float addrspace(3)* %arg, i32 209
375  %tmp366 = load float, float addrspace(3)* %tmp365, align 4
376  %tmp367 = getelementptr inbounds float, float addrspace(3)* %arg, i32 210
377  %tmp368 = load float, float addrspace(3)* %tmp367, align 4
378  %tmp369 = getelementptr inbounds float, float addrspace(3)* %arg, i32 211
379  %tmp370 = load float, float addrspace(3)* %tmp369, align 4
380  %tmp371 = tail call float @llvm.fmuladd.f32(float %tmp366, float %tmp368, float %tmp370)
381  %tmp372 = getelementptr inbounds float, float addrspace(3)* %arg, i32 213
382  %tmp373 = load float, float addrspace(3)* %tmp372, align 4
383  %tmp374 = getelementptr inbounds float, float addrspace(3)* %arg, i32 214
384  %tmp375 = load float, float addrspace(3)* %tmp374, align 4
385  %tmp376 = getelementptr inbounds float, float addrspace(3)* %arg, i32 215
386  %tmp377 = load float, float addrspace(3)* %tmp376, align 4
387  %tmp378 = tail call float @llvm.fmuladd.f32(float %tmp373, float %tmp375, float %tmp377)
388  %tmp379 = getelementptr inbounds float, float addrspace(3)* %arg, i32 217
389  %tmp380 = load float, float addrspace(3)* %tmp379, align 4
390  %tmp381 = getelementptr inbounds float, float addrspace(3)* %arg, i32 218
391  %tmp382 = load float, float addrspace(3)* %tmp381, align 4
392  %tmp383 = getelementptr inbounds float, float addrspace(3)* %arg, i32 219
393  %tmp384 = load float, float addrspace(3)* %tmp383, align 4
394  %tmp385 = tail call float @llvm.fmuladd.f32(float %tmp380, float %tmp382, float %tmp384)
395  %tmp386 = getelementptr inbounds float, float addrspace(3)* %arg, i32 221
396  %tmp387 = load float, float addrspace(3)* %tmp386, align 4
397  %tmp388 = getelementptr inbounds float, float addrspace(3)* %arg, i32 222
398  %tmp389 = load float, float addrspace(3)* %tmp388, align 4
399  %tmp390 = getelementptr inbounds float, float addrspace(3)* %arg, i32 223
400  %tmp391 = load float, float addrspace(3)* %tmp390, align 4
401  %tmp392 = tail call float @llvm.fmuladd.f32(float %tmp387, float %tmp389, float %tmp391)
402  %tmp393 = getelementptr inbounds float, float addrspace(3)* %arg, i32 225
403  %tmp394 = load float, float addrspace(3)* %tmp393, align 4
404  %tmp395 = getelementptr inbounds float, float addrspace(3)* %arg, i32 226
405  %tmp396 = load float, float addrspace(3)* %tmp395, align 4
406  %tmp397 = getelementptr inbounds float, float addrspace(3)* %arg, i32 227
407  %tmp398 = load float, float addrspace(3)* %tmp397, align 4
408  %tmp399 = tail call float @llvm.fmuladd.f32(float %tmp394, float %tmp396, float %tmp398)
409  %tmp400 = getelementptr inbounds float, float addrspace(3)* %arg, i32 229
410  %tmp401 = load float, float addrspace(3)* %tmp400, align 4
411  %tmp402 = getelementptr inbounds float, float addrspace(3)* %arg, i32 230
412  %tmp403 = load float, float addrspace(3)* %tmp402, align 4
413  %tmp404 = getelementptr inbounds float, float addrspace(3)* %arg, i32 231
414  %tmp405 = load float, float addrspace(3)* %tmp404, align 4
415  %tmp406 = tail call float @llvm.fmuladd.f32(float %tmp401, float %tmp403, float %tmp405)
416  %tmp407 = getelementptr inbounds float, float addrspace(3)* %arg, i32 233
417  %tmp408 = load float, float addrspace(3)* %tmp407, align 4
418  %tmp409 = getelementptr inbounds float, float addrspace(3)* %arg, i32 234
419  %tmp410 = load float, float addrspace(3)* %tmp409, align 4
420  %tmp411 = getelementptr inbounds float, float addrspace(3)* %arg, i32 235
421  %tmp412 = load float, float addrspace(3)* %tmp411, align 4
422  %tmp413 = tail call float @llvm.fmuladd.f32(float %tmp408, float %tmp410, float %tmp412)
423  %tmp414 = getelementptr inbounds float, float addrspace(3)* %arg, i32 237
424  %tmp415 = load float, float addrspace(3)* %tmp414, align 4
425  %tmp416 = getelementptr inbounds float, float addrspace(3)* %arg, i32 238
426  %tmp417 = load float, float addrspace(3)* %tmp416, align 4
427  %tmp418 = getelementptr inbounds float, float addrspace(3)* %arg, i32 239
428  %tmp419 = load float, float addrspace(3)* %tmp418, align 4
429  %tmp420 = tail call float @llvm.fmuladd.f32(float %tmp415, float %tmp417, float %tmp419)
430  %tmp421 = getelementptr inbounds float, float addrspace(3)* %arg, i32 241
431  %tmp422 = load float, float addrspace(3)* %tmp421, align 4
432  %tmp423 = getelementptr inbounds float, float addrspace(3)* %arg, i32 242
433  %tmp424 = load float, float addrspace(3)* %tmp423, align 4
434  %tmp425 = getelementptr inbounds float, float addrspace(3)* %arg, i32 243
435  %tmp426 = load float, float addrspace(3)* %tmp425, align 4
436  %tmp427 = tail call float @llvm.fmuladd.f32(float %tmp422, float %tmp424, float %tmp426)
437  %tmp428 = getelementptr inbounds float, float addrspace(3)* %arg, i32 245
438  %tmp429 = load float, float addrspace(3)* %tmp428, align 4
439  %tmp430 = getelementptr inbounds float, float addrspace(3)* %arg, i32 246
440  %tmp431 = load float, float addrspace(3)* %tmp430, align 4
441  %tmp432 = getelementptr inbounds float, float addrspace(3)* %arg, i32 247
442  %tmp433 = load float, float addrspace(3)* %tmp432, align 4
443  %tmp434 = tail call float @llvm.fmuladd.f32(float %tmp429, float %tmp431, float %tmp433)
444  %tmp435 = getelementptr inbounds float, float addrspace(3)* %arg, i32 249
445  %tmp436 = load float, float addrspace(3)* %tmp435, align 4
446  %tmp437 = getelementptr inbounds float, float addrspace(3)* %arg, i32 250
447  %tmp438 = load float, float addrspace(3)* %tmp437, align 4
448  %tmp439 = getelementptr inbounds float, float addrspace(3)* %arg, i32 251
449  %tmp440 = load float, float addrspace(3)* %tmp439, align 4
450  %tmp441 = tail call float @llvm.fmuladd.f32(float %tmp436, float %tmp438, float %tmp440)
451  %tmp442 = getelementptr inbounds float, float addrspace(3)* %arg, i32 253
452  %tmp443 = load float, float addrspace(3)* %tmp442, align 4
453  %tmp444 = getelementptr inbounds float, float addrspace(3)* %arg, i32 254
454  %tmp445 = load float, float addrspace(3)* %tmp444, align 4
455  %tmp446 = getelementptr inbounds float, float addrspace(3)* %arg, i32 255
456  %tmp447 = load float, float addrspace(3)* %tmp446, align 4
457  %tmp448 = tail call float @llvm.fmuladd.f32(float %tmp443, float %tmp445, float %tmp447)
458  store float %tmp7, float addrspace(1)* %arg1, align 4
459  %tmp449 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 1
460  store float %tmp14, float addrspace(1)* %tmp449, align 4
461  %tmp450 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 2
462  store float %tmp21, float addrspace(1)* %tmp450, align 4
463  %tmp451 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 3
464  store float %tmp28, float addrspace(1)* %tmp451, align 4
465  %tmp452 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 4
466  store float %tmp35, float addrspace(1)* %tmp452, align 4
467  %tmp453 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 5
468  store float %tmp42, float addrspace(1)* %tmp453, align 4
469  %tmp454 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 6
470  store float %tmp49, float addrspace(1)* %tmp454, align 4
471  %tmp455 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 7
472  store float %tmp56, float addrspace(1)* %tmp455, align 4
473  %tmp456 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 8
474  store float %tmp63, float addrspace(1)* %tmp456, align 4
475  %tmp457 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 9
476  store float %tmp70, float addrspace(1)* %tmp457, align 4
477  %tmp458 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 10
478  store float %tmp77, float addrspace(1)* %tmp458, align 4
479  %tmp459 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 11
480  store float %tmp84, float addrspace(1)* %tmp459, align 4
481  %tmp460 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 12
482  store float %tmp91, float addrspace(1)* %tmp460, align 4
483  %tmp461 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 13
484  store float %tmp98, float addrspace(1)* %tmp461, align 4
485  %tmp462 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 14
486  store float %tmp105, float addrspace(1)* %tmp462, align 4
487  %tmp463 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 15
488  store float %tmp112, float addrspace(1)* %tmp463, align 4
489  %tmp464 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 16
490  store float %tmp119, float addrspace(1)* %tmp464, align 4
491  %tmp465 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 17
492  store float %tmp126, float addrspace(1)* %tmp465, align 4
493  %tmp466 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 18
494  store float %tmp133, float addrspace(1)* %tmp466, align 4
495  %tmp467 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 19
496  store float %tmp140, float addrspace(1)* %tmp467, align 4
497  %tmp468 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 20
498  store float %tmp147, float addrspace(1)* %tmp468, align 4
499  %tmp469 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 21
500  store float %tmp154, float addrspace(1)* %tmp469, align 4
501  %tmp470 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 22
502  store float %tmp161, float addrspace(1)* %tmp470, align 4
503  %tmp471 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 23
504  store float %tmp168, float addrspace(1)* %tmp471, align 4
505  %tmp472 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 24
506  store float %tmp175, float addrspace(1)* %tmp472, align 4
507  %tmp473 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 25
508  store float %tmp182, float addrspace(1)* %tmp473, align 4
509  %tmp474 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 26
510  store float %tmp189, float addrspace(1)* %tmp474, align 4
511  %tmp475 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 27
512  store float %tmp196, float addrspace(1)* %tmp475, align 4
513  %tmp476 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 28
514  store float %tmp203, float addrspace(1)* %tmp476, align 4
515  %tmp477 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 29
516  store float %tmp210, float addrspace(1)* %tmp477, align 4
517  %tmp478 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 30
518  store float %tmp217, float addrspace(1)* %tmp478, align 4
519  %tmp479 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 31
520  store float %tmp224, float addrspace(1)* %tmp479, align 4
521  %tmp480 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 32
522  store float %tmp231, float addrspace(1)* %tmp480, align 4
523  %tmp481 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 33
524  store float %tmp238, float addrspace(1)* %tmp481, align 4
525  %tmp482 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 34
526  store float %tmp245, float addrspace(1)* %tmp482, align 4
527  %tmp483 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 35
528  store float %tmp252, float addrspace(1)* %tmp483, align 4
529  %tmp484 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 36
530  store float %tmp259, float addrspace(1)* %tmp484, align 4
531  %tmp485 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 37
532  store float %tmp266, float addrspace(1)* %tmp485, align 4
533  %tmp486 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 38
534  store float %tmp273, float addrspace(1)* %tmp486, align 4
535  %tmp487 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 39
536  store float %tmp280, float addrspace(1)* %tmp487, align 4
537  %tmp488 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 40
538  store float %tmp287, float addrspace(1)* %tmp488, align 4
539  %tmp489 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 41
540  store float %tmp294, float addrspace(1)* %tmp489, align 4
541  %tmp490 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 42
542  store float %tmp301, float addrspace(1)* %tmp490, align 4
543  %tmp491 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 43
544  store float %tmp308, float addrspace(1)* %tmp491, align 4
545  %tmp492 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 44
546  store float %tmp315, float addrspace(1)* %tmp492, align 4
547  %tmp493 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 45
548  store float %tmp322, float addrspace(1)* %tmp493, align 4
549  %tmp494 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 46
550  store float %tmp329, float addrspace(1)* %tmp494, align 4
551  %tmp495 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 47
552  store float %tmp336, float addrspace(1)* %tmp495, align 4
553  %tmp496 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 48
554  store float %tmp343, float addrspace(1)* %tmp496, align 4
555  %tmp497 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 49
556  store float %tmp350, float addrspace(1)* %tmp497, align 4
557  %tmp498 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 50
558  store float %tmp357, float addrspace(1)* %tmp498, align 4
559  %tmp499 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 51
560  store float %tmp364, float addrspace(1)* %tmp499, align 4
561  %tmp500 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 52
562  store float %tmp371, float addrspace(1)* %tmp500, align 4
563  %tmp501 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 53
564  store float %tmp378, float addrspace(1)* %tmp501, align 4
565  %tmp502 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 54
566  store float %tmp385, float addrspace(1)* %tmp502, align 4
567  %tmp503 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 55
568  store float %tmp392, float addrspace(1)* %tmp503, align 4
569  %tmp504 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 56
570  store float %tmp399, float addrspace(1)* %tmp504, align 4
571  %tmp505 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 57
572  store float %tmp406, float addrspace(1)* %tmp505, align 4
573  %tmp506 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 58
574  store float %tmp413, float addrspace(1)* %tmp506, align 4
575  %tmp507 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 59
576  store float %tmp420, float addrspace(1)* %tmp507, align 4
577  %tmp508 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 60
578  store float %tmp427, float addrspace(1)* %tmp508, align 4
579  %tmp509 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 61
580  store float %tmp434, float addrspace(1)* %tmp509, align 4
581  %tmp510 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 62
582  store float %tmp441, float addrspace(1)* %tmp510, align 4
583  %tmp511 = getelementptr inbounds float, float addrspace(1)* %arg1, i64 63
584  store float %tmp448, float addrspace(1)* %tmp511, align 4
585  ret void
586}
587
588; Function Attrs: nounwind readnone
589declare float @llvm.fmuladd.f32(float, float, float) #0
590
591attributes #0 = { nounwind readnone }
592