1 /* { dg-final { check-function-bodies "**" "" "-DCHECK_ASM" } } */
2
3 #include "test_sve_acle.h"
4
5 /*
6 ** div_f32_m_tied1:
7 ** fdiv z0\.s, p0/m, z0\.s, z1\.s
8 ** ret
9 */
10 TEST_UNIFORM_Z (div_f32_m_tied1, svfloat32_t,
11 z0 = svdiv_f32_m (p0, z0, z1),
12 z0 = svdiv_m (p0, z0, z1))
13
14 /*
15 ** div_f32_m_tied2:
16 ** mov (z[0-9]+)\.d, z0\.d
17 ** movprfx z0, z1
18 ** fdiv z0\.s, p0/m, z0\.s, \1\.s
19 ** ret
20 */
21 TEST_UNIFORM_Z (div_f32_m_tied2, svfloat32_t,
22 z0 = svdiv_f32_m (p0, z1, z0),
23 z0 = svdiv_m (p0, z1, z0))
24
25 /*
26 ** div_f32_m_untied:
27 ** movprfx z0, z1
28 ** fdiv z0\.s, p0/m, z0\.s, z2\.s
29 ** ret
30 */
31 TEST_UNIFORM_Z (div_f32_m_untied, svfloat32_t,
32 z0 = svdiv_f32_m (p0, z1, z2),
33 z0 = svdiv_m (p0, z1, z2))
34
35 /*
36 ** div_s4_f32_m_tied1:
37 ** mov (z[0-9]+\.s), s4
38 ** fdiv z0\.s, p0/m, z0\.s, \1
39 ** ret
40 */
41 TEST_UNIFORM_ZD (div_s4_f32_m_tied1, svfloat32_t, float,
42 z0 = svdiv_n_f32_m (p0, z0, d4),
43 z0 = svdiv_m (p0, z0, d4))
44
45 /*
46 ** div_s4_f32_m_untied:
47 ** mov (z[0-9]+\.s), s4
48 ** movprfx z0, z1
49 ** fdiv z0\.s, p0/m, z0\.s, \1
50 ** ret
51 */
52 TEST_UNIFORM_ZD (div_s4_f32_m_untied, svfloat32_t, float,
53 z0 = svdiv_n_f32_m (p0, z1, d4),
54 z0 = svdiv_m (p0, z1, d4))
55
56 /*
57 ** div_1_f32_m_tied1:
58 ** fmov (z[0-9]+\.s), #1\.0(?:e\+0)?
59 ** fdiv z0\.s, p0/m, z0\.s, \1
60 ** ret
61 */
62 TEST_UNIFORM_Z (div_1_f32_m_tied1, svfloat32_t,
63 z0 = svdiv_n_f32_m (p0, z0, 1),
64 z0 = svdiv_m (p0, z0, 1))
65
66 /*
67 ** div_1_f32_m_untied: { xfail *-*-* }
68 ** fmov (z[0-9]+\.s), #1\.0(?:e\+0)?
69 ** movprfx z0, z1
70 ** fdiv z0\.s, p0/m, z0\.s, \1
71 ** ret
72 */
73 TEST_UNIFORM_Z (div_1_f32_m_untied, svfloat32_t,
74 z0 = svdiv_n_f32_m (p0, z1, 1),
75 z0 = svdiv_m (p0, z1, 1))
76
77 /*
78 ** div_f32_z_tied1:
79 ** movprfx z0\.s, p0/z, z0\.s
80 ** fdiv z0\.s, p0/m, z0\.s, z1\.s
81 ** ret
82 */
83 TEST_UNIFORM_Z (div_f32_z_tied1, svfloat32_t,
84 z0 = svdiv_f32_z (p0, z0, z1),
85 z0 = svdiv_z (p0, z0, z1))
86
87 /*
88 ** div_f32_z_tied2:
89 ** movprfx z0\.s, p0/z, z0\.s
90 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
91 ** ret
92 */
93 TEST_UNIFORM_Z (div_f32_z_tied2, svfloat32_t,
94 z0 = svdiv_f32_z (p0, z1, z0),
95 z0 = svdiv_z (p0, z1, z0))
96
97 /*
98 ** div_f32_z_untied:
99 ** (
100 ** movprfx z0\.s, p0/z, z1\.s
101 ** fdiv z0\.s, p0/m, z0\.s, z2\.s
102 ** |
103 ** movprfx z0\.s, p0/z, z2\.s
104 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
105 ** )
106 ** ret
107 */
108 TEST_UNIFORM_Z (div_f32_z_untied, svfloat32_t,
109 z0 = svdiv_f32_z (p0, z1, z2),
110 z0 = svdiv_z (p0, z1, z2))
111
112 /*
113 ** div_s4_f32_z_tied1:
114 ** mov (z[0-9]+\.s), s4
115 ** movprfx z0\.s, p0/z, z0\.s
116 ** fdiv z0\.s, p0/m, z0\.s, \1
117 ** ret
118 */
119 TEST_UNIFORM_ZD (div_s4_f32_z_tied1, svfloat32_t, float,
120 z0 = svdiv_n_f32_z (p0, z0, d4),
121 z0 = svdiv_z (p0, z0, d4))
122
123 /*
124 ** div_s4_f32_z_untied:
125 ** mov (z[0-9]+\.s), s4
126 ** (
127 ** movprfx z0\.s, p0/z, z1\.s
128 ** fdiv z0\.s, p0/m, z0\.s, \1
129 ** |
130 ** movprfx z0\.s, p0/z, \1
131 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
132 ** )
133 ** ret
134 */
135 TEST_UNIFORM_ZD (div_s4_f32_z_untied, svfloat32_t, float,
136 z0 = svdiv_n_f32_z (p0, z1, d4),
137 z0 = svdiv_z (p0, z1, d4))
138
139 /*
140 ** div_1_f32_z_tied1:
141 ** fmov (z[0-9]+\.s), #1\.0(?:e\+0)?
142 ** movprfx z0\.s, p0/z, z0\.s
143 ** fdiv z0\.s, p0/m, z0\.s, \1
144 ** ret
145 */
146 TEST_UNIFORM_Z (div_1_f32_z_tied1, svfloat32_t,
147 z0 = svdiv_n_f32_z (p0, z0, 1),
148 z0 = svdiv_z (p0, z0, 1))
149
150 /*
151 ** div_1_f32_z_untied:
152 ** fmov (z[0-9]+\.s), #1\.0(?:e\+0)?
153 ** (
154 ** movprfx z0\.s, p0/z, z1\.s
155 ** fdiv z0\.s, p0/m, z0\.s, \1
156 ** |
157 ** movprfx z0\.s, p0/z, \1
158 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
159 ** )
160 ** ret
161 */
162 TEST_UNIFORM_Z (div_1_f32_z_untied, svfloat32_t,
163 z0 = svdiv_n_f32_z (p0, z1, 1),
164 z0 = svdiv_z (p0, z1, 1))
165
166 /*
167 ** div_0p5_f32_z:
168 ** fmov (z[0-9]+\.s), #(?:0\.5|5\.0e-1)
169 ** movprfx z0\.s, p0/z, z0\.s
170 ** fdiv z0\.s, p0/m, z0\.s, \1
171 ** ret
172 */
173 TEST_UNIFORM_Z (div_0p5_f32_z, svfloat32_t,
174 z0 = svdiv_n_f32_z (p0, z0, 0.5),
175 z0 = svdiv_z (p0, z0, 0.5))
176
177 /*
178 ** div_f32_x_tied1:
179 ** fdiv z0\.s, p0/m, z0\.s, z1\.s
180 ** ret
181 */
182 TEST_UNIFORM_Z (div_f32_x_tied1, svfloat32_t,
183 z0 = svdiv_f32_x (p0, z0, z1),
184 z0 = svdiv_x (p0, z0, z1))
185
186 /*
187 ** div_f32_x_tied2:
188 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
189 ** ret
190 */
191 TEST_UNIFORM_Z (div_f32_x_tied2, svfloat32_t,
192 z0 = svdiv_f32_x (p0, z1, z0),
193 z0 = svdiv_x (p0, z1, z0))
194
195 /*
196 ** div_f32_x_untied:
197 ** (
198 ** movprfx z0, z1
199 ** fdiv z0\.s, p0/m, z0\.s, z2\.s
200 ** |
201 ** movprfx z0, z2
202 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
203 ** )
204 ** ret
205 */
206 TEST_UNIFORM_Z (div_f32_x_untied, svfloat32_t,
207 z0 = svdiv_f32_x (p0, z1, z2),
208 z0 = svdiv_x (p0, z1, z2))
209
210 /*
211 ** div_s4_f32_x_tied1:
212 ** mov (z[0-9]+\.s), s4
213 ** fdiv z0\.s, p0/m, z0\.s, \1
214 ** ret
215 */
216 TEST_UNIFORM_ZD (div_s4_f32_x_tied1, svfloat32_t, float,
217 z0 = svdiv_n_f32_x (p0, z0, d4),
218 z0 = svdiv_x (p0, z0, d4))
219
220 /*
221 ** div_s4_f32_x_untied:
222 ** mov z0\.s, s4
223 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
224 ** ret
225 */
226 TEST_UNIFORM_ZD (div_s4_f32_x_untied, svfloat32_t, float,
227 z0 = svdiv_n_f32_x (p0, z1, d4),
228 z0 = svdiv_x (p0, z1, d4))
229
230 /*
231 ** div_1_f32_x_tied1:
232 ** fmov (z[0-9]+\.s), #1\.0(?:e\+0)?
233 ** fdiv z0\.s, p0/m, z0\.s, \1
234 ** ret
235 */
236 TEST_UNIFORM_Z (div_1_f32_x_tied1, svfloat32_t,
237 z0 = svdiv_n_f32_x (p0, z0, 1),
238 z0 = svdiv_x (p0, z0, 1))
239
240 /*
241 ** div_1_f32_x_untied:
242 ** fmov z0\.s, #1\.0(?:e\+0)?
243 ** fdivr z0\.s, p0/m, z0\.s, z1\.s
244 ** ret
245 */
246 TEST_UNIFORM_Z (div_1_f32_x_untied, svfloat32_t,
247 z0 = svdiv_n_f32_x (p0, z1, 1),
248 z0 = svdiv_x (p0, z1, 1))
249
250 /*
251 ** ptrue_div_f32_x_tied1:
252 ** ...
253 ** ptrue p[0-9]+\.b[^\n]*
254 ** ...
255 ** ret
256 */
257 TEST_UNIFORM_Z (ptrue_div_f32_x_tied1, svfloat32_t,
258 z0 = svdiv_f32_x (svptrue_b32 (), z0, z1),
259 z0 = svdiv_x (svptrue_b32 (), z0, z1))
260
261 /*
262 ** ptrue_div_f32_x_tied2:
263 ** ...
264 ** ptrue p[0-9]+\.b[^\n]*
265 ** ...
266 ** ret
267 */
268 TEST_UNIFORM_Z (ptrue_div_f32_x_tied2, svfloat32_t,
269 z0 = svdiv_f32_x (svptrue_b32 (), z1, z0),
270 z0 = svdiv_x (svptrue_b32 (), z1, z0))
271
272 /*
273 ** ptrue_div_f32_x_untied:
274 ** ...
275 ** ptrue p[0-9]+\.b[^\n]*
276 ** ...
277 ** ret
278 */
279 TEST_UNIFORM_Z (ptrue_div_f32_x_untied, svfloat32_t,
280 z0 = svdiv_f32_x (svptrue_b32 (), z1, z2),
281 z0 = svdiv_x (svptrue_b32 (), z1, z2))
282
283 /*
284 ** ptrue_div_1_f32_x_tied1:
285 ** ...
286 ** ptrue p[0-9]+\.b[^\n]*
287 ** ...
288 ** ret
289 */
290 TEST_UNIFORM_Z (ptrue_div_1_f32_x_tied1, svfloat32_t,
291 z0 = svdiv_n_f32_x (svptrue_b32 (), z0, 1),
292 z0 = svdiv_x (svptrue_b32 (), z0, 1))
293
294 /*
295 ** ptrue_div_1_f32_x_untied:
296 ** ...
297 ** ptrue p[0-9]+\.b[^\n]*
298 ** ...
299 ** ret
300 */
301 TEST_UNIFORM_Z (ptrue_div_1_f32_x_untied, svfloat32_t,
302 z0 = svdiv_n_f32_x (svptrue_b32 (), z1, 1),
303 z0 = svdiv_x (svptrue_b32 (), z1, 1))