1 /* { dg-do compile } */
2 /* { dg-require-effective-target powerpc_p8vector_ok } */
3 /* { dg-options "-mdejagnu-cpu=power8 -O0 -mno-fold-gimple -dp" } */
4 /* { dg-prune-output "gimple folding of rs6000 builtins has been disabled." } */
5
6 #include <altivec.h>
7
8 vector double y = { 2.0, 4.0 };
9 vector double z;
10
main()11 int main ()
12 {
13 vector float fa = {1.0, 2.0, 3.0, -4.0};
14 vector float fb = {-2.0, -3.0, -4.0, -5.0};
15 vector float fd = vec_and (fa, fb);
16 vector float fc = vec_cpsgn (fa, fb);
17 vector float fe = vec_mergeh (fa, fb);
18 vector float ff = vec_mergel (fa, fb);
19
20 vector double da = {1.0, 2.0};
21 vector double db = {-2.0, -3.0};
22 vector double dz = vec_and (da, db);
23
24 vector signed int si_a = {1, 2, 3, 4};
25 vector unsigned int ui_a = {1, 2, 3, 4};
26
27 vector long long la = {5L, 14L};
28 vector long long lb = {3L, 86L};
29 vector long long lc = vec_and (la, lb);
30 vector bool long long ld = {0, -1};
31 vector long long le = vec_and (la, ld);
32 vector long long lf = vec_and (ld, lb);
33
34 vector unsigned long long ua = {5L, 14L};
35 vector unsigned long long ub = {3L, 86L};
36 vector unsigned long long uc = vec_and (ua, ub);
37 vector bool long long ud = {0, -1};
38 vector unsigned long long ue = vec_and (ua, ud);
39 vector unsigned long long uf = vec_and (ud, ub);
40
41 vector long long lg = vec_andc (la, lb);
42 vector long long lh = vec_andc (la, ld);
43 vector long long li = vec_andc (ld, lb);
44
45 vector unsigned long long ug = vec_andc (ua, ub);
46 vector unsigned long long uh = vec_andc (ua, ud);
47 vector unsigned long long ui = vec_andc (ud, ub);
48
49 vector double de = {1.0, -4.0};
50 vector double df = {-2.0, 5.0};
51 vector double dg = vec_cpsgn (de, df);
52 vector double dzw = vec_mergeh (de, df);
53 vector double dze = vec_mergel (de, df);
54
55 vector long long lj = vec_mergeh (la, lb);
56 vector long long lk = vec_mergeh (la, ld);
57 vector long long ll = vec_mergeh (ld, la);
58
59 vector unsigned long long uj = vec_mergeh (ua, ub);
60 vector unsigned long long uk = vec_mergeh (ua, ud);
61 vector unsigned long long ul = vec_mergeh (ud, ua);
62
63 vector pixel pa = {9, 16, 25, 36, 1, 2, 3, 4};
64 vector pixel pb = {25, 36, 1, 2, 45, 3, 4, 99};
65 vector pixel pc = vec_mergeh (pa, pb);
66 vector pixel pd = vec_mergel (pa, pb);
67
68 vector long long lm = vec_mergel (la, lb);
69 vector long long ln = vec_mergel (la, ld);
70 vector long long lo = vec_mergel (ld, la);
71
72 vector unsigned long long um = vec_mergel (ua, ub);
73 vector unsigned long long un = vec_mergel (ua, ud);
74 vector unsigned long long uo = vec_mergel (ud, ua);
75
76 vector long long lp = vec_nor (la, lb);
77 vector long long lq = vec_nor (la, ld);
78 vector long long lr = vec_nor (ld, la);
79
80 vector unsigned long long up = vec_nor (ua, ub);
81 vector unsigned long long uq = vec_nor (ua, ud);
82 vector unsigned long long ur = vec_nor (ud, ua);
83
84 vector unsigned char ca = {0,4,8,1,5,9,2,6,10,3,7,11,15,12,14,13};
85 vector unsigned char cbb = {5,4,8,3,1,9,2,6,10,3,7,11,15,12,14,13};
86
87 vector unsigned char ucba = {5,4,8,3,1,9,2,6,10,3,7,11,15,12,14,13};
88 vector unsigned char ucbb = {5,4,8,3,1,9,2,6,10,3,7,11,15,12,14,13};
89 vector unsigned char ucbc = {5,4,8,3,1,9,2,6,10,3,7,11,15,12,14,13};
90
91 vector long long lv = vec_perm (la, lb, ca);
92
93 vector unsigned char ucm = vec_and (ca, cbb);
94 vector unsigned char ucn = vec_andc (ca, cbb);
95 vector unsigned char uco = vec_mergel (ca, cbb);
96
97 vector unsigned long long uv = vec_perm (ua, ub, ca);
98
99 vector long long lx = vec_sel (la, lb, uc);
100 vector long long ly = vec_sel (la, lb, ld);
101
102 vector unsigned long long uw = vec_sel (ua, ub, lc);
103 vector unsigned long long ux = vec_sel (ua, ub, uc);
104 vector unsigned long long uy = vec_sel (ua, ub, ld);
105
106 vector long long lz = vec_xor (la, lb);
107 vector long long l0 = vec_xor (la, ld);
108 vector long long l1 = vec_xor (ld, la);
109
110 vector unsigned long long uz = vec_xor (ua, ub);
111 vector unsigned long long u0 = vec_xor (ua, ud);
112 vector unsigned long long u1 = vec_xor (ud, ua);
113
114 int ia = vec_all_eq (ua, ub);
115 int ib = vec_all_ge (ua, ub);
116 int ic = vec_all_gt (ua, ub);
117 int id = vec_all_le (ua, ub);
118 int ie = vec_all_lt (ua, ub);
119 int ig = vec_all_ne (ua, ub);
120
121 int ih = vec_any_eq (ua, ub);
122 int ii = vec_any_ge (ua, ub);
123 int ij = vec_any_gt (ua, ub);
124 int ik = vec_any_le (ua, ub);
125 int il = vec_any_lt (ua, ub);
126 int im = vec_any_ne (ua, ub);
127
128 vector short ssa = {9, 16, 25, 36, 1, 2, 3, 4};
129 vector short ssb = {-8, -27, -64, -125, 2, 3, 5, 3};
130 vector short sscc = vec_and (ssa, ssb);
131 vector short sscd = vec_mergeh (ssa, ssb);
132 vector short ssce = vec_mergel (ssa, ssb);
133
134 vector int sia = {9, 16, 25, 36};
135 vector int sib = {-8, -27, -64, -125};
136 vector int sicc = vec_and (sia, sib);
137 vector int sicd = vec_andc (sia, sib);
138 vector int sig = vec_mergel (sia, sib);
139
140 vector unsigned int uia = {9, 16, 25, 36};
141 vector unsigned int uib = {8, 27, 64, 125};
142 vector unsigned int uicc = vec_and (uia, uib);
143 vector unsigned int uidd = vec_andc (uia, uib);
144 vector unsigned int uig = vec_mergel (uia, uib);
145
146 vector bool char bca = {0, 1, 4, 7};
147 vector bool char bcb = {-8, 9, 2, 9};
148 vector bool char bcc = vec_and (bca, bcb);
149 vector bool char bcd = vec_andc (bca, bcb);
150 vector bool char bce = vec_mergel (bca, bcb);
151
152 vector bool short bsa = {0, -1, -1, 0, 3, 4, 6, 7};
153 vector bool short bsb = {-1, -1, 0, -1, 0, 0, 0, 0};
154 vector bool short bscc = vec_and (bsa, bsb);
155 vector bool short bscd = vec_andc (bsa, bsb);
156 vector bool short bsce = vec_mergel (bsa, bsb);
157
158 vector bool int bia = {0, -1, -1, 0};
159 vector bool int bib = {-1, -1, 0, -1};
160 vector bool int bicc = vec_and (bia, bib);
161 vector bool int bicd = vec_andc (bia, bib);
162 vector bool int bide = vec_mergel (bia, bib);
163
164 vector unsigned int uie = vec_packsu (ua, ub);
165
166 vector long long l2 = vec_cntlz (la);
167 vector unsigned long long u2 = vec_cntlz (ua);
168 vector int sie = vec_cntlz (sia);
169 vector unsigned int uif = vec_cntlz (uia);
170 vector short sszz = vec_cntlz (ssa);
171
172 vector unsigned short usa = {81, 72, 63, 54, 45, 36, 27, 18};
173 vector unsigned short usb = {81, 72, 63, 54, 45, 36, 27, 18};
174 vector unsigned short usd = vec_and (usa, usb);
175 vector unsigned short use = vec_andc (usa, usb);
176 vector unsigned short usc = vec_cntlz (usa);
177 vector unsigned short uscd = vec_mergeh (usa, usb);
178 vector unsigned short usce = vec_mergel (usa, usb);
179
180 vector signed char sca = {-4, 3, -9, 15, -31, 31, 0, 0,
181 1, 117, -36, 99, 98, 97, 96, 95};
182 vector signed char scb = vec_cntlz (sca);
183 vector signed char scc = vec_mergel (sca, scb);
184
185 vector unsigned char uca = {4, 3, 9, 15, 30, 31, 0, 0,
186 1, 117, 36, 99, 98, 97, 96, 95};
187 vector unsigned char cb = vec_cntlz (ca);
188
189 vector double dd = vec_xl (0, &y);
190 vec_xst (dd, 0, &z);
191
192 vector double dzz = vec_round (dd);
193 vector double dzz1 = vec_rsqrt (dd);
194 vector double dzz2 = vec_rsqrte (dd);
195
196 vector float ff1 = vec_round (fa);
197 vector float ff2 = vec_rsqrt (fa);
198 vector float ff3 = vec_rsqrte (fa);
199
200 vector double dff = vec_splat (de, 0);
201 vector double dgg = vec_splat (de, 1);
202 vector long long l3 = vec_splat (l2, 0);
203 vector long long l4 = vec_splat (l2, 1);
204 vector unsigned long long u3 = vec_splat (u2, 0);
205 vector unsigned long long u4 = vec_splat (u2, 1);
206 vector bool long long l5 = vec_splat (ld, 0);
207 vector bool long long l6 = vec_splat (ld, 1);
208 vector bool long long l10 = vec_mergee (ld, ld);
209 vector bool long long l11 = vec_mergeo (ld, ld);
210 vector bool long long l15 = vec_and (ld, ld);
211
212 vector long long l7 = vec_div (l3, l4);
213 vector unsigned long long u5 = vec_div (u3, u4);
214 vector long long l12 = vec_mergee (la, lb);
215 vector long long l13 = vec_mergeo (la, lb);
216 vector unsigned long long u8 = vec_mergee (u3, u4);
217 vector unsigned long long u9 = vec_mergeo (u3, u4);
218
219 vector long long l8 = vec_mul (l3, l4);
220 vector unsigned long long u6 = vec_mul (u3, u4);
221
222 vector int sz = vec_cts (fa, 0x1F);
223 vector unsigned int usz = vec_ctu (fa, 0x1F);
224
225 vector float f1 = vec_mergee (fa, fb);
226 vector float f2 = vec_mergeo (fa, fb);
227
228 vector double d1 = vec_mergee (da, db);
229 vector double d2 = vec_mergeo (da, db);
230
231 vector float f3 = vec_ctf (si_a, 1);
232 vector float f4 = vec_ctf (ui_a, 2);
233
234 vector bool char z_vbc2 = vec_splat (bca, 0);
235 vector signed char z_vsc1 = vec_splat (sca, 1);
236 vector unsigned char z_vuc1 = vec_splat (ucbc, 2);
237
238 vector bool int z_vbi1 = vec_splat (bia, 3);
239 vector signed int z_vsi1 = vec_splat (sia, 1);
240 vector unsigned int z_vui1 = vec_splat (uia, 2);
241
242 vector bool int z_bi2 = vec_mergee (bia, bib);
243 vector signed int z_si2 = vec_mergee (sia, sib);
244 vector unsigned int z_ui2 = vec_mergee (uia, uib);
245
246 vector bool char z_bc2 = vec_mergeh (bca, bcb);
247 vector signed char z_sc2 = vec_mergeh (sca, scb);
248 vector bool int z_bi3 = vec_mergeh (bia, bib);
249 vector signed int z_si3 = vec_mergeh (sia, sib);
250 vector unsigned int z_ui3 = vec_mergeh (uia, uib);
251 vector bool short z_bs1 = vec_mergeh (bsa, bsb);
252
253 vector bool int z_bi4 = vec_mergeo (bia, bib);
254 vector signed int z_si4 = vec_mergeo (sia, sib);
255 vector unsigned int z_ui4 = vec_mergeo (uia, uib);
256
257 vector pixel int z_vp1 = vec_splat (pa, 1);
258 vector bool short z_bs2 = vec_splat (bsa, 0);
259 vector short signed int z_vss1 = vec_splat (ssa, 2);
260 vector unsigned short int z_vuss1 = vec_splat (usa, 1);
261
262 return 0;
263 }
264
265 /* Expected results:
266 vec_all_eq vcmpequd.
267 vec_all_ge vcmpgtud.
268 vec_all_ne vcmpequd.
269 vec_any_eq vcmpequd.
270 vec_any_ne vcmpequd.
271 vec_all_gt vcmpgtud.
272 vec_all_le vcmpgtud.
273 vec_all_lt vcmpgtud.
274 vec_any_ge vcmpgtud.
275 vec_any_gt vcmpgtud.
276 vec_any_lt vcmpgtud.
277 vec_any_le vcmpgtud.
278 vec_and xxland
279 vec_andc xxlandc
280 vec_cntlz vclzd, vclzb, vclzw, vclzh
281 vec_cpsgn xvcpsgnsp
282 vec_ctf xvmuldp
283 vec_cts vctsxs
284 vec_ctu vctuxs
285 vec_div divd, divdu | __divdi3(), __udivdi3()
286 vec_mergel vmrghb, vmrghh, xxmrghw
287 vec_mergeh xxmrglw, vmrglh, vmrglb
288 vec_mul mulld | mullw, mulhwu
289 vec_nor xxlnor
290 vec_packsu vpkudus
291 vec_perm vperm
292 vec_round xvrdpi
293 vec_sel xxsel
294 vec_xor xxlxor
295 vec_rsqrt xvrsqrtesp
296 vec_rsqrte xvrsqrtesp
297 vec_xl lxvd2x
298 vec_xst stxvd2x
299 vec_splat xxspltb, xxspltw, vsplth
300 vec_mergee xxmrgld, vmrgow
301 vec_mergeo xxmrghd, vmrgew */
302
303 /* { dg-final { scan-assembler-times "vcmpequd" 8 } } */
304 /* { dg-final { scan-assembler-times "vcmpgtud" 16 } } */
305 /* { dg-final { scan-assembler-times "xxland" 30 } } */
306 /* { dg-final { scan-assembler-times "xxlandc" 13 } } */
307 /* { dg-final { scan-assembler-times "vclzb" 2 } } */
308 /* { dg-final { scan-assembler-times "vclzd" 2 } } */
309 /* { dg-final { scan-assembler-times "vclzw" 2 } } */
310 /* { dg-final { scan-assembler-times "vclzh" 2 } } */
311 /* { dg-final { scan-assembler-times "xvcpsgnsp" 1 } } */
312 /* { dg-final { scan-assembler-times "xvcpsgndp" 1 } } */
313 /* { dg-final { scan-assembler-times "xvmuldp" 2 } } */
314 /* { dg-final { scan-assembler-times "xvcvdpsxds" 0 } } */
315 /* { dg-final { scan-assembler-times "vctsxs" 2 } } */
316 /* { dg-final { scan-assembler-times "xvcvdpuxds" 0 } } */
317 /* { dg-final { scan-assembler-times "vctuxs" 2 } } */
318
319 /* { dg-final { scan-assembler-times "vmrghb" 4 { target be } } } */
320 /* { dg-final { scan-assembler-times "vmrghb" 5 { target le } } } */
321 /* { dg-final { scan-assembler-times "vmrghh" 8 } } */
322 /* { dg-final { scan-assembler-times "xxmrghw" 8 } } */
323 /* { dg-final { scan-assembler-times "xxmrglw" 8 } } */
324 /* { dg-final { scan-assembler-times "vmrglh" 8 } } */
325 /* { dg-final { scan-assembler-times "xxlnor" 6 } } */
326 /* { dg-final { scan-assembler-times {\mvpkudus\M} 1 } } */
327 /* { dg-final { scan-assembler-times "vperm" 4 } } */
328 /* { dg-final { scan-assembler-times "xvrdpi" 2 } } */
329 /* { dg-final { scan-assembler-times "xxsel" 10 } } */
330 /* { dg-final { scan-assembler-times "xxlxor" 6 } } */
331 /* { dg-final { scan-assembler-times "divd" 8 { target lp64 } } } */
332 /* { dg-final { scan-assembler-times "divdu" 2 { target lp64 } } } */
333 /* { dg-final { scan-assembler-times "mulld" 4 { target lp64 } } } */
334 /* check for .__divdi3 (AIX), __divdi3 (Linux) and ___divdi3 (Darwin) */
335 /* { dg-final { scan-assembler-times {\mbl \.?_?__divdi3\M} 2 { target { ilp32 } } } } */
336 /* check for both .__udivdi3 (AIX), __udivdi3 (Linux) and ___udivdi3 (Darwin) */
337 /* { dg-final { scan-assembler-times {\mbl \.?_?__udivdi3\M} 2 { target { ilp32 } } } } */
338 /* { dg-final { scan-assembler-times "mullw" 12 { target ilp32 } } } */
339 /* { dg-final { scan-assembler-times "mulhwu" 4 { target ilp32 } } } */
340 /* { dg-final { scan-assembler-times "xxmrgld" 0 } } */
341 /* { dg-final { scan-assembler-times "xxmrghd" 0 } } */
342 /* { dg-final { scan-assembler-times "xvrsqrtesp" 2 } } */
343 /* { dg-final { scan-assembler-times "xvrsqrtedp" 2 } } */
344 /* { dg-final { scan-assembler-times "xxspltd" 8 } } */
345 /* { dg-final { scan-assembler-times "vcfsx" 2 } } */
346 /* { dg-final { scan-assembler-times "vcfux" 2 } } */
347 /* { dg-final { scan-assembler-times "vspltb" 6 } } */
348 /* { dg-final { scan-assembler-times "vspltw" 0 } } */
349 /* { dg-final { scan-assembler-times "vmrgow" 8 } } */
350 /* { dg-final { scan-assembler-times "vmrglb" 5 { target le } } } */
351 /* { dg-final { scan-assembler-times "vmrglb" 6 { target be } } } */
352 /* { dg-final { scan-assembler-times "vmrgew" 8 } } */
353 /* { dg-final { scan-assembler-times "vsplth" 8 } } */
354 /* { dg-final { scan-assembler-times "vcmpequd." 8 } } */
355 /* { dg-final { scan-assembler-times "vcmpgtud." 16 } } */
356 /* { dg-final { scan-assembler-times "vrfin" 2 } } */
357
358