1 ; RUN: llc -march=mips -mattr=+msa < %s | FileCheck %s
3 define void @and_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
6 %1 = load <16 x i8>* %a
7 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
8 %2 = load <16 x i8>* %b
9 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
10 %3 = and <16 x i8> %1, %2
11 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
12 store <16 x i8> %3, <16 x i8>* %c
13 ; CHECK-DAG: st.b [[R3]], 0($4)
16 ; CHECK: .size and_v16i8
19 define void @and_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
22 %1 = load <8 x i16>* %a
23 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
24 %2 = load <8 x i16>* %b
25 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
26 %3 = and <8 x i16> %1, %2
27 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
28 store <8 x i16> %3, <8 x i16>* %c
29 ; CHECK-DAG: st.h [[R3]], 0($4)
32 ; CHECK: .size and_v8i16
35 define void @and_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
38 %1 = load <4 x i32>* %a
39 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
40 %2 = load <4 x i32>* %b
41 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
42 %3 = and <4 x i32> %1, %2
43 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
44 store <4 x i32> %3, <4 x i32>* %c
45 ; CHECK-DAG: st.w [[R3]], 0($4)
48 ; CHECK: .size and_v4i32
51 define void @and_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
54 %1 = load <2 x i64>* %a
55 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
56 %2 = load <2 x i64>* %b
57 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
58 %3 = and <2 x i64> %1, %2
59 ; CHECK-DAG: and.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
60 store <2 x i64> %3, <2 x i64>* %c
61 ; CHECK-DAG: st.d [[R3]], 0($4)
64 ; CHECK: .size and_v2i64
67 define void @or_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
70 %1 = load <16 x i8>* %a
71 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
72 %2 = load <16 x i8>* %b
73 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
74 %3 = or <16 x i8> %1, %2
75 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
76 store <16 x i8> %3, <16 x i8>* %c
77 ; CHECK-DAG: st.b [[R3]], 0($4)
80 ; CHECK: .size or_v16i8
83 define void @or_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
86 %1 = load <8 x i16>* %a
87 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
88 %2 = load <8 x i16>* %b
89 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
90 %3 = or <8 x i16> %1, %2
91 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
92 store <8 x i16> %3, <8 x i16>* %c
93 ; CHECK-DAG: st.h [[R3]], 0($4)
96 ; CHECK: .size or_v8i16
99 define void @or_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
102 %1 = load <4 x i32>* %a
103 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
104 %2 = load <4 x i32>* %b
105 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
106 %3 = or <4 x i32> %1, %2
107 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
108 store <4 x i32> %3, <4 x i32>* %c
109 ; CHECK-DAG: st.w [[R3]], 0($4)
112 ; CHECK: .size or_v4i32
115 define void @or_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
118 %1 = load <2 x i64>* %a
119 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
120 %2 = load <2 x i64>* %b
121 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
122 %3 = or <2 x i64> %1, %2
123 ; CHECK-DAG: or.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
124 store <2 x i64> %3, <2 x i64>* %c
125 ; CHECK-DAG: st.d [[R3]], 0($4)
128 ; CHECK: .size or_v2i64
131 define void @nor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
134 %1 = load <16 x i8>* %a
135 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
136 %2 = load <16 x i8>* %b
137 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
138 %3 = or <16 x i8> %1, %2
139 %4 = xor <16 x i8> %3, <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>
140 ; CHECK-DAG: nor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
141 store <16 x i8> %4, <16 x i8>* %c
142 ; CHECK-DAG: st.b [[R3]], 0($4)
145 ; CHECK: .size nor_v16i8
148 define void @nor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
151 %1 = load <8 x i16>* %a
152 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
153 %2 = load <8 x i16>* %b
154 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
155 %3 = or <8 x i16> %1, %2
156 %4 = xor <8 x i16> %3, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
157 ; CHECK-DAG: nor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
158 store <8 x i16> %4, <8 x i16>* %c
159 ; CHECK-DAG: st.h [[R3]], 0($4)
162 ; CHECK: .size nor_v8i16
165 define void @nor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
168 %1 = load <4 x i32>* %a
169 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
170 %2 = load <4 x i32>* %b
171 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
172 %3 = or <4 x i32> %1, %2
173 %4 = xor <4 x i32> %3, <i32 -1, i32 -1, i32 -1, i32 -1>
174 ; CHECK-DAG: nor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
175 store <4 x i32> %4, <4 x i32>* %c
176 ; CHECK-DAG: st.w [[R3]], 0($4)
179 ; CHECK: .size nor_v4i32
182 define void @nor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
185 %1 = load <2 x i64>* %a
186 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
187 %2 = load <2 x i64>* %b
188 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
189 %3 = or <2 x i64> %1, %2
190 %4 = xor <2 x i64> %3, <i64 -1, i64 -1>
191 ; CHECK-DAG: nor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
192 store <2 x i64> %4, <2 x i64>* %c
193 ; CHECK-DAG: st.d [[R3]], 0($4)
196 ; CHECK: .size nor_v2i64
199 define void @xor_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
202 %1 = load <16 x i8>* %a
203 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
204 %2 = load <16 x i8>* %b
205 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
206 %3 = xor <16 x i8> %1, %2
207 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
208 store <16 x i8> %3, <16 x i8>* %c
209 ; CHECK-DAG: st.b [[R3]], 0($4)
212 ; CHECK: .size xor_v16i8
215 define void @xor_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
218 %1 = load <8 x i16>* %a
219 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
220 %2 = load <8 x i16>* %b
221 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
222 %3 = xor <8 x i16> %1, %2
223 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
224 store <8 x i16> %3, <8 x i16>* %c
225 ; CHECK-DAG: st.h [[R3]], 0($4)
228 ; CHECK: .size xor_v8i16
231 define void @xor_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
234 %1 = load <4 x i32>* %a
235 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
236 %2 = load <4 x i32>* %b
237 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
238 %3 = xor <4 x i32> %1, %2
239 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
240 store <4 x i32> %3, <4 x i32>* %c
241 ; CHECK-DAG: st.w [[R3]], 0($4)
244 ; CHECK: .size xor_v4i32
247 define void @xor_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
250 %1 = load <2 x i64>* %a
251 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
252 %2 = load <2 x i64>* %b
253 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
254 %3 = xor <2 x i64> %1, %2
255 ; CHECK-DAG: xor.v [[R3:\$w[0-9]+]], [[R1]], [[R2]]
256 store <2 x i64> %3, <2 x i64>* %c
257 ; CHECK-DAG: st.d [[R3]], 0($4)
260 ; CHECK: .size xor_v2i64
263 define void @sll_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
266 %1 = load <16 x i8>* %a
267 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
268 %2 = load <16 x i8>* %b
269 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
270 %3 = shl <16 x i8> %1, %2
271 ; CHECK-DAG: sll.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
272 store <16 x i8> %3, <16 x i8>* %c
273 ; CHECK-DAG: st.b [[R3]], 0($4)
276 ; CHECK: .size sll_v16i8
279 define void @sll_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
282 %1 = load <8 x i16>* %a
283 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
284 %2 = load <8 x i16>* %b
285 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
286 %3 = shl <8 x i16> %1, %2
287 ; CHECK-DAG: sll.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
288 store <8 x i16> %3, <8 x i16>* %c
289 ; CHECK-DAG: st.h [[R3]], 0($4)
292 ; CHECK: .size sll_v8i16
295 define void @sll_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
298 %1 = load <4 x i32>* %a
299 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
300 %2 = load <4 x i32>* %b
301 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
302 %3 = shl <4 x i32> %1, %2
303 ; CHECK-DAG: sll.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
304 store <4 x i32> %3, <4 x i32>* %c
305 ; CHECK-DAG: st.w [[R3]], 0($4)
308 ; CHECK: .size sll_v4i32
311 define void @sll_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
314 %1 = load <2 x i64>* %a
315 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
316 %2 = load <2 x i64>* %b
317 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
318 %3 = shl <2 x i64> %1, %2
319 ; CHECK-DAG: sll.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
320 store <2 x i64> %3, <2 x i64>* %c
321 ; CHECK-DAG: st.d [[R3]], 0($4)
324 ; CHECK: .size sll_v2i64
327 define void @sra_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
330 %1 = load <16 x i8>* %a
331 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
332 %2 = load <16 x i8>* %b
333 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
334 %3 = ashr <16 x i8> %1, %2
335 ; CHECK-DAG: sra.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
336 store <16 x i8> %3, <16 x i8>* %c
337 ; CHECK-DAG: st.b [[R3]], 0($4)
340 ; CHECK: .size sra_v16i8
343 define void @sra_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
346 %1 = load <8 x i16>* %a
347 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
348 %2 = load <8 x i16>* %b
349 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
350 %3 = ashr <8 x i16> %1, %2
351 ; CHECK-DAG: sra.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
352 store <8 x i16> %3, <8 x i16>* %c
353 ; CHECK-DAG: st.h [[R3]], 0($4)
356 ; CHECK: .size sra_v8i16
359 define void @sra_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
362 %1 = load <4 x i32>* %a
363 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
364 %2 = load <4 x i32>* %b
365 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
366 %3 = ashr <4 x i32> %1, %2
367 ; CHECK-DAG: sra.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
368 store <4 x i32> %3, <4 x i32>* %c
369 ; CHECK-DAG: st.w [[R3]], 0($4)
372 ; CHECK: .size sra_v4i32
375 define void @sra_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
378 %1 = load <2 x i64>* %a
379 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
380 %2 = load <2 x i64>* %b
381 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
382 %3 = ashr <2 x i64> %1, %2
383 ; CHECK-DAG: sra.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
384 store <2 x i64> %3, <2 x i64>* %c
385 ; CHECK-DAG: st.d [[R3]], 0($4)
388 ; CHECK: .size sra_v2i64
391 define void @srl_v16i8(<16 x i8>* %c, <16 x i8>* %a, <16 x i8>* %b) nounwind {
394 %1 = load <16 x i8>* %a
395 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
396 %2 = load <16 x i8>* %b
397 ; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
398 %3 = lshr <16 x i8> %1, %2
399 ; CHECK-DAG: srl.b [[R3:\$w[0-9]+]], [[R1]], [[R2]]
400 store <16 x i8> %3, <16 x i8>* %c
401 ; CHECK-DAG: st.b [[R3]], 0($4)
404 ; CHECK: .size srl_v16i8
407 define void @srl_v8i16(<8 x i16>* %c, <8 x i16>* %a, <8 x i16>* %b) nounwind {
410 %1 = load <8 x i16>* %a
411 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
412 %2 = load <8 x i16>* %b
413 ; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
414 %3 = lshr <8 x i16> %1, %2
415 ; CHECK-DAG: srl.h [[R3:\$w[0-9]+]], [[R1]], [[R2]]
416 store <8 x i16> %3, <8 x i16>* %c
417 ; CHECK-DAG: st.h [[R3]], 0($4)
420 ; CHECK: .size srl_v8i16
423 define void @srl_v4i32(<4 x i32>* %c, <4 x i32>* %a, <4 x i32>* %b) nounwind {
426 %1 = load <4 x i32>* %a
427 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
428 %2 = load <4 x i32>* %b
429 ; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
430 %3 = lshr <4 x i32> %1, %2
431 ; CHECK-DAG: srl.w [[R3:\$w[0-9]+]], [[R1]], [[R2]]
432 store <4 x i32> %3, <4 x i32>* %c
433 ; CHECK-DAG: st.w [[R3]], 0($4)
436 ; CHECK: .size srl_v4i32
439 define void @srl_v2i64(<2 x i64>* %c, <2 x i64>* %a, <2 x i64>* %b) nounwind {
442 %1 = load <2 x i64>* %a
443 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
444 %2 = load <2 x i64>* %b
445 ; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
446 %3 = lshr <2 x i64> %1, %2
447 ; CHECK-DAG: srl.d [[R3:\$w[0-9]+]], [[R1]], [[R2]]
448 store <2 x i64> %3, <2 x i64>* %c
449 ; CHECK-DAG: st.d [[R3]], 0($4)
452 ; CHECK: .size srl_v2i64
455 define void @ctpop_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
456 ; CHECK: ctpop_v16i8:
458 %1 = load <16 x i8>* %a
459 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
460 %2 = tail call <16 x i8> @llvm.ctpop.v16i8 (<16 x i8> %1)
461 ; CHECK-DAG: pcnt.b [[R3:\$w[0-9]+]], [[R1]]
462 store <16 x i8> %2, <16 x i8>* %c
463 ; CHECK-DAG: st.b [[R3]], 0($4)
466 ; CHECK: .size ctpop_v16i8
469 define void @ctpop_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
470 ; CHECK: ctpop_v8i16:
472 %1 = load <8 x i16>* %a
473 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
474 %2 = tail call <8 x i16> @llvm.ctpop.v8i16 (<8 x i16> %1)
475 ; CHECK-DAG: pcnt.h [[R3:\$w[0-9]+]], [[R1]]
476 store <8 x i16> %2, <8 x i16>* %c
477 ; CHECK-DAG: st.h [[R3]], 0($4)
480 ; CHECK: .size ctpop_v8i16
483 define void @ctpop_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
484 ; CHECK: ctpop_v4i32:
486 %1 = load <4 x i32>* %a
487 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
488 %2 = tail call <4 x i32> @llvm.ctpop.v4i32 (<4 x i32> %1)
489 ; CHECK-DAG: pcnt.w [[R3:\$w[0-9]+]], [[R1]]
490 store <4 x i32> %2, <4 x i32>* %c
491 ; CHECK-DAG: st.w [[R3]], 0($4)
494 ; CHECK: .size ctpop_v4i32
497 define void @ctpop_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
498 ; CHECK: ctpop_v2i64:
500 %1 = load <2 x i64>* %a
501 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
502 %2 = tail call <2 x i64> @llvm.ctpop.v2i64 (<2 x i64> %1)
503 ; CHECK-DAG: pcnt.d [[R3:\$w[0-9]+]], [[R1]]
504 store <2 x i64> %2, <2 x i64>* %c
505 ; CHECK-DAG: st.d [[R3]], 0($4)
508 ; CHECK: .size ctpop_v2i64
511 define void @ctlz_v16i8(<16 x i8>* %c, <16 x i8>* %a) nounwind {
514 %1 = load <16 x i8>* %a
515 ; CHECK-DAG: ld.b [[R1:\$w[0-9]+]], 0($5)
516 %2 = tail call <16 x i8> @llvm.ctlz.v16i8 (<16 x i8> %1)
517 ; CHECK-DAG: nlzc.b [[R3:\$w[0-9]+]], [[R1]]
518 store <16 x i8> %2, <16 x i8>* %c
519 ; CHECK-DAG: st.b [[R3]], 0($4)
522 ; CHECK: .size ctlz_v16i8
525 define void @ctlz_v8i16(<8 x i16>* %c, <8 x i16>* %a) nounwind {
528 %1 = load <8 x i16>* %a
529 ; CHECK-DAG: ld.h [[R1:\$w[0-9]+]], 0($5)
530 %2 = tail call <8 x i16> @llvm.ctlz.v8i16 (<8 x i16> %1)
531 ; CHECK-DAG: nlzc.h [[R3:\$w[0-9]+]], [[R1]]
532 store <8 x i16> %2, <8 x i16>* %c
533 ; CHECK-DAG: st.h [[R3]], 0($4)
536 ; CHECK: .size ctlz_v8i16
539 define void @ctlz_v4i32(<4 x i32>* %c, <4 x i32>* %a) nounwind {
542 %1 = load <4 x i32>* %a
543 ; CHECK-DAG: ld.w [[R1:\$w[0-9]+]], 0($5)
544 %2 = tail call <4 x i32> @llvm.ctlz.v4i32 (<4 x i32> %1)
545 ; CHECK-DAG: nlzc.w [[R3:\$w[0-9]+]], [[R1]]
546 store <4 x i32> %2, <4 x i32>* %c
547 ; CHECK-DAG: st.w [[R3]], 0($4)
550 ; CHECK: .size ctlz_v4i32
553 define void @ctlz_v2i64(<2 x i64>* %c, <2 x i64>* %a) nounwind {
556 %1 = load <2 x i64>* %a
557 ; CHECK-DAG: ld.d [[R1:\$w[0-9]+]], 0($5)
558 %2 = tail call <2 x i64> @llvm.ctlz.v2i64 (<2 x i64> %1)
559 ; CHECK-DAG: nlzc.d [[R3:\$w[0-9]+]], [[R1]]
560 store <2 x i64> %2, <2 x i64>* %c
561 ; CHECK-DAG: st.d [[R3]], 0($4)
564 ; CHECK: .size ctlz_v2i64
567 declare <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %val)
568 declare <8 x i16> @llvm.ctpop.v8i16(<8 x i16> %val)
569 declare <4 x i32> @llvm.ctpop.v4i32(<4 x i32> %val)
570 declare <2 x i64> @llvm.ctpop.v2i64(<2 x i64> %val)
571 declare <16 x i8> @llvm.ctlz.v16i8(<16 x i8> %val)
572 declare <8 x i16> @llvm.ctlz.v8i16(<8 x i16> %val)
573 declare <4 x i32> @llvm.ctlz.v4i32(<4 x i32> %val)
574 declare <2 x i64> @llvm.ctlz.v2i64(<2 x i64> %val)