1 /* { dg-do compile } */
2 /* { dg-options "-march=rv32gcv -mabi=ilp32d -O3 -fno-schedule-insns -fno-schedule-insns2" } */
3
4 #include "riscv_vector.h"
5
6 void f0 (uint16_t *base,uint8_t *out,size_t vl, size_t shift)
7 {
8 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
9 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
10 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
11 __riscv_vse8_v_u8m1 (out,v,vl);
12 }
13
14 void f1 (uint16_t *base,uint8_t *out,size_t vl, size_t shift)
15 {
16 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
17 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
18 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
19 __riscv_vse8_v_u8m1 (out,v,vl);
20 }
21
22 void f2 (void *base,void *out,size_t vl, size_t shift)
23 {
24 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
25 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
26 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
27 __riscv_vse8_v_u8m1 (out,v,vl);
28 __riscv_vse16_v_u16m2 (out+100,src,vl);
29 }
30
31 void f3 (void *base,void *out,size_t vl, size_t shift)
32 {
33 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
34 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
35 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
36 __riscv_vse8_v_u8m1 (out,v,vl);
37 __riscv_vse16_v_u16m2 (out+100,src,vl);
38 }
39
40 void f4 (void *base,void *out,size_t vl, size_t shift)
41 {
42 vbool8_t m = __riscv_vlm_v_b8 (base + 500, vl);
43 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
44 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
45 v = __riscv_vnclipu_wv_u8m1_tumu(m,v,src,v,vl);
46 __riscv_vse8_v_u8m1 (out,v,vl);
47 __riscv_vse16_v_u16m2 (out+100,src,vl);
48 }
49
50 void f5 (void *base,void *out,size_t vl, size_t shift)
51 {
52 vbool8_t m = __riscv_vlm_v_b8 (base + 500, vl);
53 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
54 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
55 v = __riscv_vnclipu_wv_u8m1_m(m,src,v,vl);
56 __riscv_vse8_v_u8m1 (out,v,vl);
57 __riscv_vse16_v_u16m2 (out+100,src,vl);
58 }
59
60 void f6 (void *base,void *out,size_t vl, size_t shift)
61 {
62 vbool8_t m = __riscv_vlm_v_b8 (base + 500, vl);
63 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
64 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
65 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1_m(m,src,v,vl);
66 __riscv_vse8_v_u8m1 (out,v2,vl);
67 __riscv_vse8_v_u8m1 (out+100,v,vl);
68 }
69
70 void f7 (void *base,void *out,size_t vl, size_t shift)
71 {
72 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
73 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
74 vuint8m1_t v2 = __riscv_vnclipu_wx_u8m1(src,shift,vl);
75 v2 = __riscv_vnclipu_wv_u8m1 (src,v,vl);
76 __riscv_vse8_v_u8m1 (out,v2,vl);
77 __riscv_vse8_v_u8m1 (out+100,v,vl);
78 }
79
80 void f8 (void *base,void *out,size_t vl, size_t shift)
81 {
82 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
83 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
84 vuint8m1_t v2 = __riscv_vnclipu_wx_u8m1(src,shift,vl);
85 v2 = __riscv_vnclipu_wv_u8m1 (src,v,vl);
86 __riscv_vse8_v_u8m1 (out,v2,vl);
87 __riscv_vse8_v_u8m1 (out+100,v,vl);
88 __riscv_vse16_v_u16m2 (out+200,src,vl);
89 }
90
91 void f9 (void *base,void *out,size_t vl, size_t shift)
92 {
93 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
94 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
95 vuint8m1_t v2 = __riscv_vnclipu_wx_u8m1(src,shift,vl);
96 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
97 __riscv_vse8_v_u8m1 (out,v2,vl);
98 __riscv_vse8_v_u8m1 (out+100,v,vl);
99 __riscv_vse16_v_u16m2 (out+200,src,vl);
100 }
101
102 void f10 (void *base,void *out,size_t vl, size_t shift)
103 {
104 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
105 vuint8m1_t v = __riscv_vnclipu_wx_u8m1(src,shift,vl);
106 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
107 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
108 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
109 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
110 __riscv_vse8_v_u8m1 (out,v,vl);
111 __riscv_vse16_v_u16m2 (out+100,src,vl);
112 }
113
114 void f11 (void *base,void *out,size_t vl, size_t shift)
115 {
116 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
117 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
118 vuint8m1_t v2 = __riscv_vnclipu_wx_u8m1(src,shift,vl);
119 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
120 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
121 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
122 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
123 v2 = __riscv_vnclipu_wv_u8m1_tu (v2,src,v,vl);
124 __riscv_vse8_v_u8m1 (out,v2,vl);
125 __riscv_vse8_v_u8m1 (out+100,v,vl);
126 __riscv_vse16_v_u16m2 (out+200,src,vl);
127 }
128
129 void f12 (void *base,void *out,size_t vl, size_t shift)
130 {
131 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
132 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
133 vuint8m1_t v2 = __riscv_vnclipu_wx_u8m1(src,shift,vl);
134 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
135 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
136 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
137 v2 = __riscv_vnclipu_wv_u8m1 (src,v2,vl);
138 __riscv_vse8_v_u8m1 (out,v2,vl);
139 __riscv_vse8_v_u8m1 (out+100,v,vl);
140 }
141
142 void f13 (void *base,void *base2,void *out,size_t vl, int n)
143 {
144 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100, vl);
145 for (int i = 0; i < n; i++){
146 vbool8_t m = __riscv_vlm_v_b8 (base + i, vl);
147 vuint8m1_t v = __riscv_vnclipu_wx_u8m1_m(m,src,vl,vl);
148 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
149 v = __riscv_vle8_v_u8m1_tu (v, base2, vl);
150 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
151 }
152 }
153
154 void f14 (void *base,void *base2,void *out,size_t vl, int n)
155 {
156 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100, vl);
157 for (int i = 0; i < n; i++){
158 vbool8_t m = __riscv_vlm_v_b8 (base + i, vl);
159 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
160 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
161 v = __riscv_vle8_v_u8m1_tu (v, base2, vl);
162 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
163 __riscv_vse8_v_u8m1 (out + 100*i,v2,vl);
164 }
165 }
166
167 void f15 (void *base,void *base2,void *out,size_t vl, int n)
168 {
169 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100, vl);
170 for (int i = 0; i < n; i++){
171 vbool8_t m = __riscv_vlm_v_b8 (base + i, vl);
172 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 600, vl);
173 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
174 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
175 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
176 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
177 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
178 v = __riscv_vle8_v_u8m1_tu (v, base2, vl);
179 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
180 __riscv_vse8_v_u8m1 (out + 100*i,v2,vl);
181 }
182 }
183
184 void f16 (uint16_t *base,uint8_t *out,size_t vl, size_t shift)
185 {
186 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
187 vuint8m1_t v = __riscv_vncvt_x_x_w_u8m1(src,vl);
188 vuint8m1_t v3 = __riscv_vnclipu_wv_u8m1(src,v,vl);
189 __riscv_vse8_v_u8m1 (out,v,vl);
190 __riscv_vse8_v_u8m1 (out + 100,v3,vl);
191 }
192
193 void f17 (void *base,void *out,size_t vl, int n)
194 {
195 for (int i = 0; i < n; i++){
196 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100*i, vl);
197 vuint8m1_t src2 = __riscv_vle8_v_u8m1 (base + 200*i, vl);
198 vuint8m1_t v = __riscv_vnclipu_wv_u8m1(src,src2,vl);
199 vuint16m2_t v2 = __riscv_vadd_vv_u16m2 (src, src,vl);
200 asm volatile ("":::"memory");
201 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
202 __riscv_vse16_v_u16m2 ((out + 200*i),src,vl);
203 __riscv_vse8_v_u8m1 ((out + 300*i),src2,vl);
204 }
205 }
206
207 void f18 (void *base,void *out,size_t vl, int n)
208 {
209 vuint8m1_t v = __riscv_vle8_v_u8m1 ((base + 1000), vl);
210 for (int i = 0; i < n; i++){
211 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100*i, vl);
212 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
213 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
214 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
215 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
216 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
217 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
218 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
219 }
220 }
221
222 void f19 (void *base,void *out,size_t vl, int n)
223 {
224 vuint8m1_t v = __riscv_vle8_v_u8m1 ((base + 1000), vl);
225 for (int i = 0; i < n; i++){
226 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100*i, vl);
227 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
228 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
229 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
230 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
231 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
232 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
233 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
234 __riscv_vse8_v_u8m1 (out + 200*i,v2,vl);
235 }
236 }
237
238 void f20 (void *base,void *out,size_t vl, int n)
239 {
240 vuint8m1_t v = __riscv_vle8_v_u8m1 ((base + 1000), vl);
241 for (int i = 0; i < n; i++){
242 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100*i, vl);
243 v = __riscv_vnclipu_wv_u8m1(src,v,vl);
244 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
245 v2 = __riscv_vnclipu_wv_u8m1(src,v2,vl);
246 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
247 __riscv_vse8_v_u8m1 (out + 200*i,v2,vl);
248 }
249 }
250
251 void f21 (void *base,void *out,size_t vl, int n)
252 {
253 for (int i = 0; i < n; i++){
254 vuint8m1_t v = __riscv_vle8_v_u8m1 ((base + 1000 * i), vl);
255 vuint16m2_t src = __riscv_vle16_v_u16m2 (base + 100*i, vl);
256 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
257 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
258 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
259 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
260 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
261 v = __riscv_vnclipu_wv_u8m1_tu(v,src,v,vl);
262 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
263 __riscv_vse16_v_u16m2 (out + 200*i,src,vl);
264 }
265 }
266
267 void f22 (uint16_t *base,uint8_t *out,size_t vl, int n)
268 {
269 for (int i = 0; i < n; i++){
270 vuint8m1_t v = __riscv_vle8_v_u8m1 ((uint8_t *)(base + 1000 * i), vl);
271 vuint16m2_t src1 = __riscv_vle16_v_u16m2 (base + 100*i, vl);
272 vuint16m2_t src2 = __riscv_vle16_v_u16m2 (base + 200*i, vl);
273 vuint16m2_t src3 = __riscv_vle16_v_u16m2 (base + 300*i, vl);
274 vuint16m2_t src4 = __riscv_vle16_v_u16m2 (base + 400*i, vl);
275 vuint16m2_t src5 = __riscv_vle16_v_u16m2 (base + 500*i, vl);
276 vuint16m2_t src6 = __riscv_vle16_v_u16m2 (base + 600*i, vl);
277 v = __riscv_vnclipu_wv_u8m1_tu(v,src1,v,vl);
278 v = __riscv_vnclipu_wv_u8m1_tu(v,src2,v,vl);
279 v = __riscv_vnclipu_wv_u8m1_tu(v,src3,v,vl);
280 v = __riscv_vnclipu_wv_u8m1_tu(v,src4,v,vl);
281 v = __riscv_vnclipu_wv_u8m1_tu(v,src5,v,vl);
282 v = __riscv_vnclipu_wv_u8m1_tu(v,src6,v,vl);
283 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
284 }
285 }
286
287 void f23 (uint16_t *base,uint8_t *out,size_t vl, int n)
288 {
289 for (int i = 0; i < n; i++){
290 vuint8m1_t v = __riscv_vle8_v_u8m1 ((uint8_t *)(base + 1000 * i), vl);
291 vuint8m1_t v2 = __riscv_vle8_v_u8m1 ((uint8_t *)(base + 2000 * i), vl);
292 vuint16m2_t src1 = __riscv_vle16_v_u16m2 (base + 100*i, vl);
293 vuint16m2_t src2 = __riscv_vle16_v_u16m2 (base + 200*i, vl);
294 vuint16m2_t src3 = __riscv_vle16_v_u16m2 (base + 300*i, vl);
295 vuint16m2_t src4 = __riscv_vle16_v_u16m2 (base + 400*i, vl);
296 vuint16m2_t src5 = __riscv_vle16_v_u16m2 (base + 500*i, vl);
297 vuint16m2_t src6 = __riscv_vle16_v_u16m2 (base + 600*i, vl);
298 v = __riscv_vnclipu_wv_u8m1_tu(v,src1,v2,vl);
299 v = __riscv_vnclipu_wv_u8m1_tu(v,src2,v2,vl);
300 v = __riscv_vnclipu_wv_u8m1_tu(v,src3,v2,vl);
301 v = __riscv_vnclipu_wv_u8m1_tu(v,src4,v2,vl);
302 v = __riscv_vnclipu_wv_u8m1_tu(v,src5,v2,vl);
303 v = __riscv_vnclipu_wv_u8m1_tu(v,src6,v2,vl);
304 __riscv_vse8_v_u8m1 (out + 100*i,v,vl);
305 }
306 }
307
308 void f24 (void *base,void *base2,void *out,size_t vl, int n)
309 {
310 vuint32m4_t src = __riscv_vle32_v_u32m4 (base + 100, vl);
311 vuint16m2_t src2 = __riscv_vle16_v_u16m2 (base + 200, vl);
312 vuint8m1_t src3 = __riscv_vle8_v_u8m1 (base + 300, vl);
313 for (int i = 0; i < n; i++){
314 vbool8_t m = __riscv_vlm_v_b8 (base + i, vl);
315 vuint16m2_t v = __riscv_vnclipu_wv_u16m2_m(m,src,src2,vl);
316 vuint16m2_t v2 = __riscv_vle16_v_u16m2_tu (v, base2 + i, vl);
317 vuint8m1_t v3 = __riscv_vnclipu_wv_u8m1_m(m,v2,src3,vl);
318 __riscv_vse8_v_u8m1 (out + 100*i,v3,vl);
319 }
320 }
321
322 void f25 (void *base,void *out,size_t vl, size_t shift)
323 {
324 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
325 asm volatile("#" ::
326 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
327 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
328 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
329 "v26", "v27", "v28", "v29");
330 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 100, vl);
331 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
332 asm volatile("#" ::
333 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
334 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
335 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
336 "v26", "v27", "v28", "v29", "v31");
337 __riscv_vse8_v_u8m1 (out,v2,vl);
338 asm volatile("#" ::
339 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
340 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
341 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
342 "v26", "v27", "v28", "v29", "v31");
343 }
344
345 void f26 (void *base,void *out,size_t vl, size_t shift)
346 {
347 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 100, vl);
348 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
349 asm volatile("#" ::
350 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
351 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
352 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
353 "v26", "v27", "v28");
354 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
355 asm volatile("#" ::
356 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
357 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
358 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
359 "v26", "v27", "v28", "v30", "v31");
360 __riscv_vse8_v_u8m1 (out,v2,vl);
361 asm volatile("#" ::
362 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
363 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
364 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
365 "v26", "v27", "v28", "v30", "v31");
366 }
367
368 void f27 (void *base,void *out,size_t vl, size_t shift)
369 {
370 vuint8m1_t v = __riscv_vle8_v_u8m1 (base + 100, vl);
371 vuint16m2_t src = __riscv_vle16_v_u16m2 (base, vl);
372 asm volatile("#" ::
373 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
374 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
375 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
376 "v26", "v27", "v28");
377 vuint8m1_t v2 = __riscv_vnclipu_wv_u8m1(src,v,vl);
378 asm volatile("#" ::
379 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
380 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
381 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
382 "v26", "v29", "v28", "v30", "v31");
383 __riscv_vse8_v_u8m1 (out,v2,vl);
384 asm volatile("#" ::
385 : "v0", "v1", "v2", "v3", "v4", "v5", "v6", "v7", "v8", "v9",
386 "v10", "v11", "v12", "v13", "v14", "v15", "v16", "v17",
387 "v18", "v19", "v20", "v21", "v22", "v23", "v24", "v25",
388 "v26", "v29", "v28", "v30", "v31");
389 }
390
391 /* { dg-final { scan-assembler-not {vmv} } } */
392 /* { dg-final { scan-assembler-not {csrr} } } */