Debian Codegraph @

gcc/testsuite/gcc.target/aarch64/sve2/acle/asm/amax_f32.c

raw
1 /* { dg-do compile } */
2 /* { dg-final { check-function-bodies "**" "" "-DCHECK_ASM" } } */
3
4 #include "test_sve_acle.h"
5
6 #pragma GCC target "+sve2+faminmax"
7 #if STREAMING_COMPATIBLE
8 #pragma GCC target "+sme2"
9 #endif
10
11 /*
12 ** amax_f32_m_tied1:
13 ** famax z0\.s, p0/m, z0\.s, z1\.s
14 ** ret
15 */
16 TEST_UNIFORM_Z (amax_f32_m_tied1, svfloat32_t,
17 z0 = svamax_f32_m (p0, z0, z1),
18 z0 = svamax_m (p0, z0, z1))
19
20 /*
21 ** amax_f32_m_tied2:
22 ** mov (z[0-9]+)\.d, z0\.d
23 ** ...
24 ** famax z0\.s, p0/m, z0\.s, \1\.s
25 ** ret
26 */
27 TEST_UNIFORM_Z (amax_f32_m_tied2, svfloat32_t,
28 z0 = svamax_f32_m (p0, z1, z0),
29 z0 = svamax_m (p0, z1, z0))
30
31 /*
32 ** amax_f32_m_untied:
33 ** ...
34 ** famax z0\.s, p0/m, z0\.s, z2\.s
35 ** ret
36 */
37 TEST_UNIFORM_Z (amax_f32_m_untied, svfloat32_t,
38 z0 = svamax_f32_m (p0, z1, z2),
39 z0 = svamax_m (p0, z1, z2))
40
41 /*
42 ** amax_s4_f32_m_tied1:
43 ** mov (z[0-9]+\.s), s4
44 ** famax z0\.s, p0/m, z0\.s, \1
45 ** ret
46 */
47 TEST_UNIFORM_ZD (amax_s4_f32_m_tied1, svfloat32_t, float,
48 z0 = svamax_n_f32_m (p0, z0, d4),
49 z0 = svamax_m (p0, z0, d4))
50
51 /*
52 ** amax_s4_f32_m_untied:
53 ** mov (z[0-9]+\.s), s4
54 ** ...
55 ** famax z0\.s, p0/m, z0\.s, \1
56 ** ret
57 */
58 TEST_UNIFORM_ZD (amax_s4_f32_m_untied, svfloat32_t, float,
59 z0 = svamax_n_f32_m (p0, z1, d4),
60 z0 = svamax_m (p0, z1, d4))
61
62 /*
63 ** amax_0_f32_m_tied1:
64 ** ...
65 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
66 ** ret
67 */
68 TEST_UNIFORM_Z (amax_0_f32_m_tied1, svfloat32_t,
69 z0 = svamax_n_f32_m (p0, z0, 0),
70 z0 = svamax_m (p0, z0, 0))
71
72 /*
73 ** amax_0_f32_m_untied:
74 ** ...
75 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
76 ** ret
77 */
78 TEST_UNIFORM_Z (amax_0_f32_m_untied, svfloat32_t,
79 z0 = svamax_n_f32_m (p0, z1, 0),
80 z0 = svamax_m (p0, z1, 0))
81
82 /*
83 ** amax_1_f32_m_tied1:
84 ** ...
85 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
86 ** ret
87 */
88 TEST_UNIFORM_Z (amax_1_f32_m_tied1, svfloat32_t,
89 z0 = svamax_n_f32_m (p0, z0, 1),
90 z0 = svamax_m (p0, z0, 1))
91
92 /*
93 ** amax_1_f32_m_untied:
94 ** ...
95 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
96 ** ret
97 */
98 TEST_UNIFORM_Z (amax_1_f32_m_untied, svfloat32_t,
99 z0 = svamax_n_f32_m (p0, z1, 1),
100 z0 = svamax_m (p0, z1, 1))
101
102 /*
103 ** amax_2_f32_m:
104 ** ...
105 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
106 ** ret
107 */
108 TEST_UNIFORM_Z (amax_2_f32_m, svfloat32_t,
109 z0 = svamax_n_f32_m (p0, z0, 2),
110 z0 = svamax_m (p0, z0, 2))
111
112 /*
113 ** amax_f32_z_tied1:
114 ** ...
115 ** famax z0\.s, p0/m, z0\.s, z1\.s
116 ** ret
117 */
118 TEST_UNIFORM_Z (amax_f32_z_tied1, svfloat32_t,
119 z0 = svamax_f32_z (p0, z0, z1),
120 z0 = svamax_z (p0, z0, z1))
121
122 /*
123 ** amax_f32_z_tied2:
124 ** ...
125 ** famax z0\.s, p0/m, z0\.s, z1\.s
126 ** ret
127 */
128 TEST_UNIFORM_Z (amax_f32_z_tied2, svfloat32_t,
129 z0 = svamax_f32_z (p0, z1, z0),
130 z0 = svamax_z (p0, z1, z0))
131
132 /*
133 ** amax_f32_z_untied:
134 ** (
135 ** ...
136 ** famax z0\.s, p0/m, z0\.s, z2\.s
137 ** |
138 ** ...
139 ** famax z0\.s, p0/m, z0\.s, z1\.s
140 ** )
141 ** ret
142 */
143 TEST_UNIFORM_Z (amax_f32_z_untied, svfloat32_t,
144 z0 = svamax_f32_z (p0, z1, z2),
145 z0 = svamax_z (p0, z1, z2))
146
147 /*
148 ** amax_s4_f32_z_tied1:
149 ** mov (z[0-9]+\.s), s4
150 ** ...
151 ** famax z0\.s, p0/m, z0\.s, \1
152 ** ret
153 */
154 TEST_UNIFORM_ZD (amax_s4_f32_z_tied1, svfloat32_t, float,
155 z0 = svamax_n_f32_z (p0, z0, d4),
156 z0 = svamax_z (p0, z0, d4))
157
158 /*
159 ** amax_s4_f32_z_untied:
160 ** mov (z[0-9]+\.s), s4
161 ** (
162 ** ...
163 ** famax z0\.s, p0/m, z0\.s, \1
164 ** |
165 ** ...
166 ** famax z0\.s, p0/m, z0\.s, z1\.s
167 ** )
168 ** ret
169 */
170 TEST_UNIFORM_ZD (amax_s4_f32_z_untied, svfloat32_t, float,
171 z0 = svamax_n_f32_z (p0, z1, d4),
172 z0 = svamax_z (p0, z1, d4))
173
174 /*
175 ** amax_0_f32_z_tied1:
176 ** ...
177 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
178 ** ret
179 */
180 TEST_UNIFORM_Z (amax_0_f32_z_tied1, svfloat32_t,
181 z0 = svamax_n_f32_z (p0, z0, 0),
182 z0 = svamax_z (p0, z0, 0))
183
184 /*
185 ** amax_0_f32_z_untied:
186 ** ...
187 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
188 ** ret
189 */
190 TEST_UNIFORM_Z (amax_0_f32_z_untied, svfloat32_t,
191 z0 = svamax_n_f32_z (p0, z1, 0),
192 z0 = svamax_z (p0, z1, 0))
193
194 /*
195 ** amax_1_f32_z_tied1:
196 ** ...
197 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
198 ** ret
199 */
200 TEST_UNIFORM_Z (amax_1_f32_z_tied1, svfloat32_t,
201 z0 = svamax_n_f32_z (p0, z0, 1),
202 z0 = svamax_z (p0, z0, 1))
203
204 /*
205 ** amax_1_f32_z_untied:
206 ** ...
207 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
208 ** ret
209 */
210 TEST_UNIFORM_Z (amax_1_f32_z_untied, svfloat32_t,
211 z0 = svamax_n_f32_z (p0, z1, 1),
212 z0 = svamax_z (p0, z1, 1))
213
214 /*
215 ** amax_2_f32_z:
216 ** ...
217 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
218 ** ret
219 */
220 TEST_UNIFORM_Z (amax_2_f32_z, svfloat32_t,
221 z0 = svamax_n_f32_z (p0, z0, 2),
222 z0 = svamax_z (p0, z0, 2))
223
224 /*
225 ** amax_f32_x_tied1:
226 ** famax z0\.s, p0/m, z0\.s, z1\.s
227 ** ret
228 */
229 TEST_UNIFORM_Z (amax_f32_x_tied1, svfloat32_t,
230 z0 = svamax_f32_x (p0, z0, z1),
231 z0 = svamax_x (p0, z0, z1))
232
233 /*
234 ** amax_f32_x_tied2:
235 ** famax z0\.s, p0/m, z0\.s, z1\.s
236 ** ret
237 */
238 TEST_UNIFORM_Z (amax_f32_x_tied2, svfloat32_t,
239 z0 = svamax_f32_x (p0, z1, z0),
240 z0 = svamax_x (p0, z1, z0))
241
242 /*
243 ** amax_f32_x_untied:
244 ** (
245 ** ...
246 ** famax z0\.s, p0/m, z0\.s, z2\.s
247 ** |
248 ** ...
249 ** famax z0\.s, p0/m, z0\.s, z1\.s
250 ** )
251 ** ret
252 */
253 TEST_UNIFORM_Z (amax_f32_x_untied, svfloat32_t,
254 z0 = svamax_f32_x (p0, z1, z2),
255 z0 = svamax_x (p0, z1, z2))
256
257 /*
258 ** amax_s4_f32_x_tied1:
259 ** mov (z[0-9]+\.s), s4
260 ** famax z0\.s, p0/m, z0\.s, \1
261 ** ret
262 */
263 TEST_UNIFORM_ZD (amax_s4_f32_x_tied1, svfloat32_t, float,
264 z0 = svamax_n_f32_x (p0, z0, d4),
265 z0 = svamax_x (p0, z0, d4))
266
267 /*
268 ** amax_s4_f32_x_untied:
269 ** mov z0\.s, s4
270 ** famax z0\.s, p0/m, z0\.s, z1\.s
271 ** ret
272 */
273 TEST_UNIFORM_ZD (amax_s4_f32_x_untied, svfloat32_t, float,
274 z0 = svamax_n_f32_x (p0, z1, d4),
275 z0 = svamax_x (p0, z1, d4))
276
277 /*
278 ** amax_0_f32_x_tied1:
279 ** ...
280 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
281 ** ret
282 */
283 TEST_UNIFORM_Z (amax_0_f32_x_tied1, svfloat32_t,
284 z0 = svamax_n_f32_x (p0, z0, 0),
285 z0 = svamax_x (p0, z0, 0))
286
287 /*
288 ** amax_0_f32_x_untied:
289 ** ...
290 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
291 ** ret
292 */
293 TEST_UNIFORM_Z (amax_0_f32_x_untied, svfloat32_t,
294 z0 = svamax_n_f32_x (p0, z1, 0),
295 z0 = svamax_x (p0, z1, 0))
296
297 /*
298 ** amax_1_f32_x_tied1:
299 ** ...
300 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
301 ** ret
302 */
303 TEST_UNIFORM_Z (amax_1_f32_x_tied1, svfloat32_t,
304 z0 = svamax_n_f32_x (p0, z0, 1),
305 z0 = svamax_x (p0, z0, 1))
306
307 /*
308 ** amax_1_f32_x_untied:
309 ** ...
310 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
311 ** ret
312 */
313 TEST_UNIFORM_Z (amax_1_f32_x_untied, svfloat32_t,
314 z0 = svamax_n_f32_x (p0, z1, 1),
315 z0 = svamax_x (p0, z1, 1))
316
317 /*
318 ** amax_2_f32_x_tied1:
319 ** ...
320 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
321 ** ret
322 */
323 TEST_UNIFORM_Z (amax_2_f32_x_tied1, svfloat32_t,
324 z0 = svamax_n_f32_x (p0, z0, 2),
325 z0 = svamax_x (p0, z0, 2))
326
327 /*
328 ** amax_2_f32_x_untied:
329 ** ...
330 ** famax z0\.s, p0/m, z0\.s, z[0-9]+\.s
331 ** ret
332 */
333 TEST_UNIFORM_Z (amax_2_f32_x_untied, svfloat32_t,
334 z0 = svamax_n_f32_x (p0, z1, 2),
335 z0 = svamax_x (p0, z1, 2))
336
337 /*
338 ** ptrue_amax_f32_x_tied1:
339 ** ...
340 ** ptrue p[0-9]+\.b[^\n]*
341 ** ...
342 ** ret
343 */
344 TEST_UNIFORM_Z (ptrue_amax_f32_x_tied1, svfloat32_t,
345 z0 = svamax_f32_x (svptrue_b32 (), z0, z1),
346 z0 = svamax_x (svptrue_b32 (), z0, z1))
347
348 /*
349 ** ptrue_amax_f32_x_tied2:
350 ** ...
351 ** ptrue p[0-9]+\.b[^\n]*
352 ** ...
353 ** ret
354 */
355 TEST_UNIFORM_Z (ptrue_amax_f32_x_tied2, svfloat32_t,
356 z0 = svamax_f32_x (svptrue_b32 (), z1, z0),
357 z0 = svamax_x (svptrue_b32 (), z1, z0))
358
359 /*
360 ** ptrue_amax_f32_x_untied:
361 ** ...
362 ** ptrue p[0-9]+\.b[^\n]*
363 ** ...
364 ** ret
365 */
366 TEST_UNIFORM_Z (ptrue_amax_f32_x_untied, svfloat32_t,
367 z0 = svamax_f32_x (svptrue_b32 (), z1, z2),
368 z0 = svamax_x (svptrue_b32 (), z1, z2))
369
370 /*
371 ** ptrue_amax_0_f32_x_tied1:
372 ** ...
373 ** ptrue p[0-9]+\.b[^\n]*
374 ** ...
375 ** ret
376 */
377 TEST_UNIFORM_Z (ptrue_amax_0_f32_x_tied1, svfloat32_t,
378 z0 = svamax_n_f32_x (svptrue_b32 (), z0, 0),
379 z0 = svamax_x (svptrue_b32 (), z0, 0))
380
381 /*
382 ** ptrue_amax_0_f32_x_untied:
383 ** ...
384 ** ptrue p[0-9]+\.b[^\n]*
385 ** ...
386 ** ret
387 */
388 TEST_UNIFORM_Z (ptrue_amax_0_f32_x_untied, svfloat32_t,
389 z0 = svamax_n_f32_x (svptrue_b32 (), z1, 0),
390 z0 = svamax_x (svptrue_b32 (), z1, 0))
391
392 /*
393 ** ptrue_amax_1_f32_x_tied1:
394 ** ...
395 ** ptrue p[0-9]+\.b[^\n]*
396 ** ...
397 ** ret
398 */
399 TEST_UNIFORM_Z (ptrue_amax_1_f32_x_tied1, svfloat32_t,
400 z0 = svamax_n_f32_x (svptrue_b32 (), z0, 1),
401 z0 = svamax_x (svptrue_b32 (), z0, 1))
402
403 /*
404 ** ptrue_amax_1_f32_x_untied:
405 ** ...
406 ** ptrue p[0-9]+\.b[^\n]*
407 ** ...
408 ** ret
409 */
410 TEST_UNIFORM_Z (ptrue_amax_1_f32_x_untied, svfloat32_t,
411 z0 = svamax_n_f32_x (svptrue_b32 (), z1, 1),
412 z0 = svamax_x (svptrue_b32 (), z1, 1))
413
414 /*
415 ** ptrue_amax_2_f32_x_tied1:
416 ** ...
417 ** ptrue p[0-9]+\.b[^\n]*
418 ** ...
419 ** ret
420 */
421 TEST_UNIFORM_Z (ptrue_amax_2_f32_x_tied1, svfloat32_t,
422 z0 = svamax_n_f32_x (svptrue_b32 (), z0, 2),
423 z0 = svamax_x (svptrue_b32 (), z0, 2))
424
425 /*
426 ** ptrue_amax_2_f32_x_untied:
427 ** ...
428 ** ptrue p[0-9]+\.b[^\n]*
429 ** ...
430 ** ret
431 */
432 TEST_UNIFORM_Z (ptrue_amax_2_f32_x_untied, svfloat32_t,
433 z0 = svamax_n_f32_x (svptrue_b32 (), z1, 2),
434 z0 = svamax_x (svptrue_b32 (), z1, 2))