Debian Codegraph @

gcc/testsuite/gcc.target/aarch64/simd/vmla_fpm.c

raw
1 /* { dg-do compile } */
2 /* { dg-additional-options "-O3 -march=armv9-a+fp8fma -moverride=tune=cheap_fpmr_write" } */
3 /* { dg-final { check-function-bodies "**" "" } } */
4
5 #include "arm_neon.h"
6
7 /*
8 ** test_vmlalbq_f16_fpm:
9 ** msr fpmr, x0
10 ** fmlalb v0.8h, v1.16b, v2.16b
11 ** ret
12 */
13 float16x8_t
14 test_vmlalbq_f16_fpm (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
15 {
16 return vmlalbq_f16_mf8_fpm (a, b, c, d);
17 }
18
19 /*
20 ** test_vmlaltq_f16_fpm:
21 ** msr fpmr, x0
22 ** fmlalt v0.8h, v1.16b, v2.16b
23 ** ret
24 */
25 float16x8_t
26 test_vmlaltq_f16_fpm (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
27 {
28 return vmlaltq_f16_mf8_fpm (a, b, c, d);
29 }
30
31 /*
32 ** test_vmlallbbq_f32_fpm:
33 ** msr fpmr, x0
34 ** fmlallbb v0.4s, v1.16b, v2.16b
35 ** ret
36 */
37 float32x4_t
38 test_vmlallbbq_f32_fpm (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
39 {
40 return vmlallbbq_f32_mf8_fpm (a, b, c, d);
41 }
42
43 /*
44 ** test_vmlallbtq_f32_fpm:
45 ** msr fpmr, x0
46 ** fmlallbt v0.4s, v1.16b, v2.16b
47 ** ret
48 */
49 float32x4_t
50 test_vmlallbtq_f32_fpm (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
51 {
52 return vmlallbtq_f32_mf8_fpm (a, b, c, d);
53 }
54
55 /*
56 ** test_vmlalltbq_f32_fpm:
57 ** msr fpmr, x0
58 ** fmlalltb v0.4s, v1.16b, v2.16b
59 ** ret
60 */
61 float32x4_t
62 test_vmlalltbq_f32_fpm (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
63 {
64 return vmlalltbq_f32_mf8_fpm (a, b, c, d);
65 }
66
67 /*
68 ** test_vmlallttq_f32_fpm:
69 ** msr fpmr, x0
70 ** fmlalltt v0.4s, v1.16b, v2.16b
71 ** ret
72 */
73 float32x4_t
74 test_vmlallttq_f32_fpm (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
75 {
76 return vmlallttq_f32_mf8_fpm (a, b, c, d);
77 }
78
79 /*
80 ** test_vmlalbq_lane_f16_fpm_0:
81 ** msr fpmr, x0
82 ** fmlalb v0.8h, v1.16b, v2.b\[0\]
83 ** ret
84 */
85 float16x8_t
86 test_vmlalbq_lane_f16_fpm_0 (float16x8_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
87 {
88 return vmlalbq_lane_f16_mf8_fpm (a, b, c, 0, d);
89 }
90
91 /*
92 ** test_vmlalbq_lane_f16_fpm_7:
93 ** msr fpmr, x0
94 ** fmlalb v0.8h, v1.16b, v2.b\[7\]
95 ** ret
96 */
97 float16x8_t
98 test_vmlalbq_lane_f16_fpm_7 (float16x8_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
99 {
100 return vmlalbq_lane_f16_mf8_fpm (a, b, c, 7, d);
101 }
102
103 /*
104 ** test_vmlalbq_laneq_f16_fpm_0:
105 ** msr fpmr, x0
106 ** fmlalb v0.8h, v1.16b, v2.b\[0\]
107 ** ret
108 */
109 float16x8_t
110 test_vmlalbq_laneq_f16_fpm_0 (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
111 {
112 return vmlalbq_laneq_f16_mf8_fpm (a, b, c, 0, d);
113 }
114
115 /*
116 ** test_vmlalbq_laneq_f16_fpm_15:
117 ** msr fpmr, x0
118 ** fmlalb v0.8h, v1.16b, v2.b\[15\]
119 ** ret
120 */
121 float16x8_t
122 test_vmlalbq_laneq_f16_fpm_15 (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
123 {
124 return vmlalbq_laneq_f16_mf8_fpm (a, b, c, 15, d);
125 }
126
127 /*
128 ** test_vmlaltq_lane_f16_fpm_0:
129 ** msr fpmr, x0
130 ** fmlalt v0.8h, v1.16b, v2.b\[0\]
131 ** ret
132 */
133 float16x8_t
134 test_vmlaltq_lane_f16_fpm_0 (float16x8_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
135 {
136 return vmlaltq_lane_f16_mf8_fpm (a, b, c, 0, d);
137 }
138
139 /*
140 ** test_vmlaltq_lane_f16_fpm_7:
141 ** msr fpmr, x0
142 ** fmlalt v0.8h, v1.16b, v2.b\[7\]
143 ** ret
144 */
145 float16x8_t
146 test_vmlaltq_lane_f16_fpm_7 (float16x8_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
147 {
148 return vmlaltq_lane_f16_mf8_fpm (a, b, c, 7, d);
149 }
150
151 /*
152 ** test_vmlaltq_laneq_f16_fpm_0:
153 ** msr fpmr, x0
154 ** fmlalt v0.8h, v1.16b, v2.b\[0\]
155 ** ret
156 */
157 float16x8_t
158 test_vmlaltq_laneq_f16_fpm_0 (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
159 {
160 return vmlaltq_laneq_f16_mf8_fpm (a, b, c, 0, d);
161 }
162
163 /*
164 ** test_vmlaltq_laneq_f16_fpm_15:
165 ** msr fpmr, x0
166 ** fmlalt v0.8h, v1.16b, v2.b\[15\]
167 ** ret
168 */
169 float16x8_t
170 test_vmlaltq_laneq_f16_fpm_15 (float16x8_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
171 {
172 return vmlaltq_laneq_f16_mf8_fpm (a, b, c, 15, d);
173 }
174
175 /*
176 ** test_vmlallbbq_lane_f32_fpm_0:
177 ** msr fpmr, x0
178 ** fmlallbb v0.4s, v1.16b, v2.b\[0\]
179 ** ret
180 */
181 float32x4_t
182 test_vmlallbbq_lane_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
183 {
184 return vmlallbbq_lane_f32_mf8_fpm (a, b, c, 0, d);
185 }
186
187 /*
188 ** test_vmlallbbq_lane_f32_fpm_7:
189 ** msr fpmr, x0
190 ** fmlallbb v0.4s, v1.16b, v2.b\[7\]
191 ** ret
192 */
193 float32x4_t
194 test_vmlallbbq_lane_f32_fpm_7 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
195 {
196 return vmlallbbq_lane_f32_mf8_fpm (a, b, c, 7, d);
197 }
198
199 /*
200 ** test_vmlallbbq_laneq_f32_fpm_0:
201 ** msr fpmr, x0
202 ** fmlallbb v0.4s, v1.16b, v2.b\[0\]
203 ** ret
204 */
205 float32x4_t
206 test_vmlallbbq_laneq_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
207 {
208 return vmlallbbq_laneq_f32_mf8_fpm (a, b, c, 0, d);
209 }
210
211 /*
212 ** test_vmlallbbq_laneq_f32_fpm_15:
213 ** msr fpmr, x0
214 ** fmlallbb v0.4s, v1.16b, v2.b\[15\]
215 ** ret
216 */
217 float32x4_t
218 test_vmlallbbq_laneq_f32_fpm_15 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
219 {
220 return vmlallbbq_laneq_f32_mf8_fpm (a, b, c, 15, d);
221 }
222
223 /*
224 ** test_vmlallbtq_lane_f32_fpm_0:
225 ** msr fpmr, x0
226 ** fmlallbt v0.4s, v1.16b, v2.b\[0\]
227 ** ret
228 */
229 float32x4_t
230 test_vmlallbtq_lane_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
231 {
232 return vmlallbtq_lane_f32_mf8_fpm (a, b, c, 0, d);
233 }
234
235 /*
236 ** test_vmlallbtq_lane_f32_fpm_7:
237 ** msr fpmr, x0
238 ** fmlallbt v0.4s, v1.16b, v2.b\[7\]
239 ** ret
240 */
241 float32x4_t
242 test_vmlallbtq_lane_f32_fpm_7 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
243 {
244 return vmlallbtq_lane_f32_mf8_fpm (a, b, c, 7, d);
245 }
246
247 /*
248 ** test_vmlallbtq_laneq_f32_fpm_0:
249 ** msr fpmr, x0
250 ** fmlallbt v0.4s, v1.16b, v2.b\[0\]
251 ** ret
252 */
253 float32x4_t
254 test_vmlallbtq_laneq_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
255 {
256 return vmlallbtq_laneq_f32_mf8_fpm (a, b, c, 0, d);
257 }
258
259 /*
260 ** test_vmlallbtq_laneq_f32_fpm_15:
261 ** msr fpmr, x0
262 ** fmlallbt v0.4s, v1.16b, v2.b\[15\]
263 ** ret
264 */
265 float32x4_t
266 test_vmlallbtq_laneq_f32_fpm_15 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
267 {
268 return vmlallbtq_laneq_f32_mf8_fpm (a, b, c, 15, d);
269 }
270
271 /*
272 ** test_vmlalltbq_lane_f32_fpm_0:
273 ** msr fpmr, x0
274 ** fmlalltb v0.4s, v1.16b, v2.b\[0\]
275 ** ret
276 */
277 float32x4_t
278 test_vmlalltbq_lane_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
279 {
280 return vmlalltbq_lane_f32_mf8_fpm (a, b, c, 0, d);
281 }
282
283 /*
284 ** test_vmlalltbq_lane_f32_fpm_7:
285 ** msr fpmr, x0
286 ** fmlalltb v0.4s, v1.16b, v2.b\[7\]
287 ** ret
288 */
289 float32x4_t
290 test_vmlalltbq_lane_f32_fpm_7 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
291 {
292 return vmlalltbq_lane_f32_mf8_fpm (a, b, c, 7, d);
293 }
294
295 /*
296 ** test_vmlalltbq_laneq_f32_fpm_0:
297 ** msr fpmr, x0
298 ** fmlalltb v0.4s, v1.16b, v2.b\[0\]
299 ** ret
300 */
301 float32x4_t
302 test_vmlalltbq_laneq_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
303 {
304 return vmlalltbq_laneq_f32_mf8_fpm (a, b, c, 0, d);
305 }
306
307 /*
308 ** test_vmlalltbq_laneq_f32_fpm_15:
309 ** msr fpmr, x0
310 ** fmlalltb v0.4s, v1.16b, v2.b\[15\]
311 ** ret
312 */
313 float32x4_t
314 test_vmlalltbq_laneq_f32_fpm_15 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
315 {
316 return vmlalltbq_laneq_f32_mf8_fpm (a, b, c, 15, d);
317 }
318
319 /*
320 ** test_vmlallttq_lane_f32_fpm_0:
321 ** msr fpmr, x0
322 ** fmlalltt v0.4s, v1.16b, v2.b\[0\]
323 ** ret
324 */
325 float32x4_t
326 test_vmlallttq_lane_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
327 {
328 return vmlallttq_lane_f32_mf8_fpm (a, b, c, 0, d);
329 }
330
331 /*
332 ** test_vmlallttq_lane_f32_fpm_7:
333 ** msr fpmr, x0
334 ** fmlalltt v0.4s, v1.16b, v2.b\[7\]
335 ** ret
336 */
337 float32x4_t
338 test_vmlallttq_lane_f32_fpm_7 (float32x4_t a, mfloat8x16_t b, mfloat8x8_t c, fpm_t d)
339 {
340 return vmlallttq_lane_f32_mf8_fpm (a, b, c, 7, d);
341 }
342
343 /*
344 ** test_vmlallttq_laneq_f32_fpm_0:
345 ** msr fpmr, x0
346 ** fmlalltt v0.4s, v1.16b, v2.b\[0\]
347 ** ret
348 */
349 float32x4_t
350 test_vmlallttq_laneq_f32_fpm_0 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
351 {
352 return vmlallttq_laneq_f32_mf8_fpm (a, b, c, 0, d);
353 }
354
355 /*
356 ** test_vmlallttq_laneq_f32_fpm_15:
357 ** msr fpmr, x0
358 ** fmlalltt v0.4s, v1.16b, v2.b\[15\]
359 ** ret
360 */
361 float32x4_t
362 test_vmlallttq_laneq_f32_fpm_15 (float32x4_t a, mfloat8x16_t b, mfloat8x16_t c, fpm_t d)
363 {
364 return vmlallttq_laneq_f32_mf8_fpm (a, b, c, 15, d);
365 }