1 ; RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN %s
3 ; GCN-LABEL: {{^}}load_fma_store
4 ; GCN-NOT: scratch_store
7 define amdgpu_kernel void @load_fma_store(ptr addrspace(3) nocapture readonly %arg, ptr addrspace(1) nocapture %arg1, i1 %cc) {
9 %vgpr0 = call <32 x i32> asm sideeffect "; def $0","=${v[0:31]}"()
10 %vgpr1 = call <32 x i32> asm sideeffect "; def $0","=${v[32:63]}"()
11 %vgpr2 = call <32 x i32> asm sideeffect "; def $0","=${v[64:95]}"()
12 %vgpr3 = call <32 x i32> asm sideeffect "; def $0","=${v[96:127]}"()
13 br i1 %cc, label %bb1, label %bb2
16 %tmp = getelementptr inbounds float, ptr addrspace(3) %arg, i32 1
17 %tmp2 = load float, ptr addrspace(3) %tmp, align 4
18 %tmp3 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 2
19 %tmp4 = load float, ptr addrspace(3) %tmp3, align 4
20 %tmp5 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 3
21 %tmp6 = load float, ptr addrspace(3) %tmp5, align 4
22 %tmp7 = tail call float @llvm.fmuladd.f32(float %tmp2, float %tmp4, float %tmp6)
23 %tmp8 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 5
24 %tmp9 = load float, ptr addrspace(3) %tmp8, align 4
25 %tmp10 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 6
26 %tmp11 = load float, ptr addrspace(3) %tmp10, align 4
27 %tmp12 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 7
28 %tmp13 = load float, ptr addrspace(3) %tmp12, align 4
29 %tmp14 = tail call float @llvm.fmuladd.f32(float %tmp9, float %tmp11, float %tmp13)
30 %tmp15 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 9
31 %tmp16 = load float, ptr addrspace(3) %tmp15, align 4
32 %tmp17 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 10
33 %tmp18 = load float, ptr addrspace(3) %tmp17, align 4
34 %tmp19 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 11
35 %tmp20 = load float, ptr addrspace(3) %tmp19, align 4
36 %tmp21 = tail call float @llvm.fmuladd.f32(float %tmp16, float %tmp18, float %tmp20)
37 %tmp22 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 13
38 %tmp23 = load float, ptr addrspace(3) %tmp22, align 4
39 %tmp24 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 14
40 %tmp25 = load float, ptr addrspace(3) %tmp24, align 4
41 %tmp26 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 15
42 %tmp27 = load float, ptr addrspace(3) %tmp26, align 4
43 %tmp28 = tail call float @llvm.fmuladd.f32(float %tmp23, float %tmp25, float %tmp27)
44 %tmp29 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 17
45 %tmp30 = load float, ptr addrspace(3) %tmp29, align 4
46 %tmp31 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 18
47 %tmp32 = load float, ptr addrspace(3) %tmp31, align 4
48 %tmp33 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 19
49 %tmp34 = load float, ptr addrspace(3) %tmp33, align 4
50 %tmp35 = tail call float @llvm.fmuladd.f32(float %tmp30, float %tmp32, float %tmp34)
51 %tmp36 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 21
52 %tmp37 = load float, ptr addrspace(3) %tmp36, align 4
53 %tmp38 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 22
54 %tmp39 = load float, ptr addrspace(3) %tmp38, align 4
55 %tmp40 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 23
56 %tmp41 = load float, ptr addrspace(3) %tmp40, align 4
57 %tmp42 = tail call float @llvm.fmuladd.f32(float %tmp37, float %tmp39, float %tmp41)
58 %tmp43 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 25
59 %tmp44 = load float, ptr addrspace(3) %tmp43, align 4
60 %tmp45 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 26
61 %tmp46 = load float, ptr addrspace(3) %tmp45, align 4
62 %tmp47 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 27
63 %tmp48 = load float, ptr addrspace(3) %tmp47, align 4
64 %tmp49 = tail call float @llvm.fmuladd.f32(float %tmp44, float %tmp46, float %tmp48)
65 %tmp50 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 29
66 %tmp51 = load float, ptr addrspace(3) %tmp50, align 4
67 %tmp52 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 30
68 %tmp53 = load float, ptr addrspace(3) %tmp52, align 4
69 %tmp54 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 31
70 %tmp55 = load float, ptr addrspace(3) %tmp54, align 4
71 %tmp56 = tail call float @llvm.fmuladd.f32(float %tmp51, float %tmp53, float %tmp55)
72 %tmp57 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 33
73 %tmp58 = load float, ptr addrspace(3) %tmp57, align 4
74 %tmp59 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 34
75 %tmp60 = load float, ptr addrspace(3) %tmp59, align 4
76 %tmp61 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 35
77 %tmp62 = load float, ptr addrspace(3) %tmp61, align 4
78 %tmp63 = tail call float @llvm.fmuladd.f32(float %tmp58, float %tmp60, float %tmp62)
79 %tmp64 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 37
80 %tmp65 = load float, ptr addrspace(3) %tmp64, align 4
81 %tmp66 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 38
82 %tmp67 = load float, ptr addrspace(3) %tmp66, align 4
83 %tmp68 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 39
84 %tmp69 = load float, ptr addrspace(3) %tmp68, align 4
85 %tmp70 = tail call float @llvm.fmuladd.f32(float %tmp65, float %tmp67, float %tmp69)
86 %tmp71 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 41
87 %tmp72 = load float, ptr addrspace(3) %tmp71, align 4
88 %tmp73 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 42
89 %tmp74 = load float, ptr addrspace(3) %tmp73, align 4
90 %tmp75 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 43
91 %tmp76 = load float, ptr addrspace(3) %tmp75, align 4
92 %tmp77 = tail call float @llvm.fmuladd.f32(float %tmp72, float %tmp74, float %tmp76)
93 %tmp78 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 45
94 %tmp79 = load float, ptr addrspace(3) %tmp78, align 4
95 %tmp80 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 46
96 %tmp81 = load float, ptr addrspace(3) %tmp80, align 4
97 %tmp82 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 47
98 %tmp83 = load float, ptr addrspace(3) %tmp82, align 4
99 %tmp84 = tail call float @llvm.fmuladd.f32(float %tmp79, float %tmp81, float %tmp83)
100 %tmp85 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 49
101 %tmp86 = load float, ptr addrspace(3) %tmp85, align 4
102 %tmp87 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 50
103 %tmp88 = load float, ptr addrspace(3) %tmp87, align 4
104 %tmp89 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 51
105 %tmp90 = load float, ptr addrspace(3) %tmp89, align 4
106 %tmp91 = tail call float @llvm.fmuladd.f32(float %tmp86, float %tmp88, float %tmp90)
107 %tmp92 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 53
108 %tmp93 = load float, ptr addrspace(3) %tmp92, align 4
109 %tmp94 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 54
110 %tmp95 = load float, ptr addrspace(3) %tmp94, align 4
111 %tmp96 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 55
112 %tmp97 = load float, ptr addrspace(3) %tmp96, align 4
113 %tmp98 = tail call float @llvm.fmuladd.f32(float %tmp93, float %tmp95, float %tmp97)
114 %tmp99 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 57
115 %tmp100 = load float, ptr addrspace(3) %tmp99, align 4
116 %tmp101 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 58
117 %tmp102 = load float, ptr addrspace(3) %tmp101, align 4
118 %tmp103 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 59
119 %tmp104 = load float, ptr addrspace(3) %tmp103, align 4
120 %tmp105 = tail call float @llvm.fmuladd.f32(float %tmp100, float %tmp102, float %tmp104)
121 %tmp106 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 61
122 %tmp107 = load float, ptr addrspace(3) %tmp106, align 4
123 %tmp108 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 62
124 %tmp109 = load float, ptr addrspace(3) %tmp108, align 4
125 %tmp110 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 63
126 %tmp111 = load float, ptr addrspace(3) %tmp110, align 4
127 %tmp112 = tail call float @llvm.fmuladd.f32(float %tmp107, float %tmp109, float %tmp111)
128 %tmp113 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 65
129 %tmp114 = load float, ptr addrspace(3) %tmp113, align 4
130 %tmp115 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 66
131 %tmp116 = load float, ptr addrspace(3) %tmp115, align 4
132 %tmp117 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 67
133 %tmp118 = load float, ptr addrspace(3) %tmp117, align 4
134 %tmp119 = tail call float @llvm.fmuladd.f32(float %tmp114, float %tmp116, float %tmp118)
135 %tmp120 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 69
136 %tmp121 = load float, ptr addrspace(3) %tmp120, align 4
137 %tmp122 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 70
138 %tmp123 = load float, ptr addrspace(3) %tmp122, align 4
139 %tmp124 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 71
140 %tmp125 = load float, ptr addrspace(3) %tmp124, align 4
141 %tmp126 = tail call float @llvm.fmuladd.f32(float %tmp121, float %tmp123, float %tmp125)
142 %tmp127 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 73
143 %tmp128 = load float, ptr addrspace(3) %tmp127, align 4
144 %tmp129 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 74
145 %tmp130 = load float, ptr addrspace(3) %tmp129, align 4
146 %tmp131 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 75
147 %tmp132 = load float, ptr addrspace(3) %tmp131, align 4
148 %tmp133 = tail call float @llvm.fmuladd.f32(float %tmp128, float %tmp130, float %tmp132)
149 %tmp134 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 77
150 %tmp135 = load float, ptr addrspace(3) %tmp134, align 4
151 %tmp136 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 78
152 %tmp137 = load float, ptr addrspace(3) %tmp136, align 4
153 %tmp138 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 79
154 %tmp139 = load float, ptr addrspace(3) %tmp138, align 4
155 %tmp140 = tail call float @llvm.fmuladd.f32(float %tmp135, float %tmp137, float %tmp139)
156 %tmp141 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 81
157 %tmp142 = load float, ptr addrspace(3) %tmp141, align 4
158 %tmp143 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 82
159 %tmp144 = load float, ptr addrspace(3) %tmp143, align 4
160 %tmp145 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 83
161 %tmp146 = load float, ptr addrspace(3) %tmp145, align 4
162 %tmp147 = tail call float @llvm.fmuladd.f32(float %tmp142, float %tmp144, float %tmp146)
163 %tmp148 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 85
164 %tmp149 = load float, ptr addrspace(3) %tmp148, align 4
165 %tmp150 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 86
166 %tmp151 = load float, ptr addrspace(3) %tmp150, align 4
167 %tmp152 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 87
168 %tmp153 = load float, ptr addrspace(3) %tmp152, align 4
169 %tmp154 = tail call float @llvm.fmuladd.f32(float %tmp149, float %tmp151, float %tmp153)
170 %tmp155 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 89
171 %tmp156 = load float, ptr addrspace(3) %tmp155, align 4
172 %tmp157 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 90
173 %tmp158 = load float, ptr addrspace(3) %tmp157, align 4
174 %tmp159 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 91
175 %tmp160 = load float, ptr addrspace(3) %tmp159, align 4
176 %tmp161 = tail call float @llvm.fmuladd.f32(float %tmp156, float %tmp158, float %tmp160)
177 %tmp162 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 93
178 %tmp163 = load float, ptr addrspace(3) %tmp162, align 4
179 %tmp164 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 94
180 %tmp165 = load float, ptr addrspace(3) %tmp164, align 4
181 %tmp166 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 95
182 %tmp167 = load float, ptr addrspace(3) %tmp166, align 4
183 %tmp168 = tail call float @llvm.fmuladd.f32(float %tmp163, float %tmp165, float %tmp167)
184 %tmp169 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 97
185 %tmp170 = load float, ptr addrspace(3) %tmp169, align 4
186 %tmp171 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 98
187 %tmp172 = load float, ptr addrspace(3) %tmp171, align 4
188 %tmp173 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 99
189 %tmp174 = load float, ptr addrspace(3) %tmp173, align 4
190 %tmp175 = tail call float @llvm.fmuladd.f32(float %tmp170, float %tmp172, float %tmp174)
191 %tmp176 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 101
192 %tmp177 = load float, ptr addrspace(3) %tmp176, align 4
193 %tmp178 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 102
194 %tmp179 = load float, ptr addrspace(3) %tmp178, align 4
195 %tmp180 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 103
196 %tmp181 = load float, ptr addrspace(3) %tmp180, align 4
197 %tmp182 = tail call float @llvm.fmuladd.f32(float %tmp177, float %tmp179, float %tmp181)
198 %tmp183 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 105
199 %tmp184 = load float, ptr addrspace(3) %tmp183, align 4
200 %tmp185 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 106
201 %tmp186 = load float, ptr addrspace(3) %tmp185, align 4
202 %tmp187 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 107
203 %tmp188 = load float, ptr addrspace(3) %tmp187, align 4
204 %tmp189 = tail call float @llvm.fmuladd.f32(float %tmp184, float %tmp186, float %tmp188)
205 %tmp190 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 109
206 %tmp191 = load float, ptr addrspace(3) %tmp190, align 4
207 %tmp192 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 110
208 %tmp193 = load float, ptr addrspace(3) %tmp192, align 4
209 %tmp194 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 111
210 %tmp195 = load float, ptr addrspace(3) %tmp194, align 4
211 %tmp196 = tail call float @llvm.fmuladd.f32(float %tmp191, float %tmp193, float %tmp195)
212 %tmp197 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 113
213 %tmp198 = load float, ptr addrspace(3) %tmp197, align 4
214 %tmp199 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 114
215 %tmp200 = load float, ptr addrspace(3) %tmp199, align 4
216 %tmp201 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 115
217 %tmp202 = load float, ptr addrspace(3) %tmp201, align 4
218 %tmp203 = tail call float @llvm.fmuladd.f32(float %tmp198, float %tmp200, float %tmp202)
219 %tmp204 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 117
220 %tmp205 = load float, ptr addrspace(3) %tmp204, align 4
221 %tmp206 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 118
222 %tmp207 = load float, ptr addrspace(3) %tmp206, align 4
223 %tmp208 = getelementptr inbounds float, ptr addrspace(3) %arg, i32 119
224 %tmp209 = load float, ptr addrspace(3) %tmp208, align 4
225 %tmp210 = tail call float @llvm.fmuladd.f32(float %tmp205, float %tmp207, float %tmp209)
226 store float %tmp7, ptr addrspace(1) %arg1, align 4
227 %tmp449 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 1
228 store float %tmp14, ptr addrspace(1) %tmp449, align 4
229 %tmp450 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 2
230 store float %tmp21, ptr addrspace(1) %tmp450, align 4
231 %tmp451 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 3
232 store float %tmp28, ptr addrspace(1) %tmp451, align 4
233 %tmp452 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 4
234 store float %tmp35, ptr addrspace(1) %tmp452, align 4
235 %tmp453 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 5
236 store float %tmp42, ptr addrspace(1) %tmp453, align 4
237 %tmp454 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 6
238 store float %tmp49, ptr addrspace(1) %tmp454, align 4
239 %tmp455 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 7
240 store float %tmp56, ptr addrspace(1) %tmp455, align 4
241 %tmp456 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 8
242 store float %tmp63, ptr addrspace(1) %tmp456, align 4
243 %tmp457 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 9
244 store float %tmp70, ptr addrspace(1) %tmp457, align 4
245 %tmp458 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 10
246 store float %tmp77, ptr addrspace(1) %tmp458, align 4
247 %tmp459 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 11
248 store float %tmp84, ptr addrspace(1) %tmp459, align 4
249 %tmp460 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 12
250 store float %tmp91, ptr addrspace(1) %tmp460, align 4
251 %tmp461 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 13
252 store float %tmp98, ptr addrspace(1) %tmp461, align 4
253 %tmp462 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 14
254 store float %tmp105, ptr addrspace(1) %tmp462, align 4
255 %tmp463 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 15
256 store float %tmp112, ptr addrspace(1) %tmp463, align 4
257 %tmp464 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 16
258 store float %tmp119, ptr addrspace(1) %tmp464, align 4
259 %tmp465 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 17
260 store float %tmp126, ptr addrspace(1) %tmp465, align 4
261 %tmp466 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 18
262 store float %tmp133, ptr addrspace(1) %tmp466, align 4
263 %tmp467 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 19
264 store float %tmp140, ptr addrspace(1) %tmp467, align 4
265 %tmp468 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 20
266 store float %tmp147, ptr addrspace(1) %tmp468, align 4
267 %tmp469 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 21
268 store float %tmp154, ptr addrspace(1) %tmp469, align 4
269 %tmp470 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 22
270 store float %tmp161, ptr addrspace(1) %tmp470, align 4
271 %tmp471 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 23
272 store float %tmp168, ptr addrspace(1) %tmp471, align 4
273 %tmp472 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 24
274 store float %tmp175, ptr addrspace(1) %tmp472, align 4
275 %tmp473 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 25
276 store float %tmp182, ptr addrspace(1) %tmp473, align 4
277 %tmp474 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 26
278 store float %tmp189, ptr addrspace(1) %tmp474, align 4
279 %tmp475 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 27
280 store float %tmp196, ptr addrspace(1) %tmp475, align 4
281 %tmp476 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 28
282 store float %tmp203, ptr addrspace(1) %tmp476, align 4
283 %tmp477 = getelementptr inbounds float, ptr addrspace(1) %arg1, i64 29
284 store float %tmp210, ptr addrspace(1) %tmp477, align 4
285 br i1 %cc, label %bb1, label %bb2
288 call void asm sideeffect "; use $0","{v[0:31]}"(<32 x i32> %vgpr0)
289 call void asm sideeffect "; use $0","{v[0:31]}"(<32 x i32> %vgpr1)
290 call void asm sideeffect "; use $0","{v[0:31]}"(<32 x i32> %vgpr2)
291 call void asm sideeffect "; use $0","{v[0:31]}"(<32 x i32> %vgpr3)
295 declare float @llvm.fmuladd.f32(float, float, float)