1 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx -x86-experimental-vector-shuffle-lowering | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
2 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 -x86-experimental-vector-shuffle-lowering | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX2
4 target triple = "x86_64-unknown-unknown"
6 define <4 x double> @shuffle_v4f64_0000(<4 x double> %a, <4 x double> %b) {
7 ; AVX1-LABEL: shuffle_v4f64_0000:
9 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
10 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
13 ; AVX2-LABEL: shuffle_v4f64_0000:
15 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
17 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
18 ret <4 x double> %shuffle
21 define <4 x double> @shuffle_v4f64_0001(<4 x double> %a, <4 x double> %b) {
22 ; AVX1-LABEL: shuffle_v4f64_0001:
24 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]
25 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
28 ; AVX2-LABEL: shuffle_v4f64_0001:
30 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
32 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
33 ret <4 x double> %shuffle
36 define <4 x double> @shuffle_v4f64_0020(<4 x double> %a, <4 x double> %b) {
37 ; AVX1-LABEL: shuffle_v4f64_0020:
39 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
40 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
41 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
42 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
45 ; AVX2-LABEL: shuffle_v4f64_0020:
47 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,2,0]
49 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 0>
50 ret <4 x double> %shuffle
53 define <4 x double> @shuffle_v4f64_0300(<4 x double> %a, <4 x double> %b) {
54 ; AVX1-LABEL: shuffle_v4f64_0300:
56 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
57 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,1,2,2]
58 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
61 ; AVX2-LABEL: shuffle_v4f64_0300:
63 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,3,0,0]
65 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
66 ret <4 x double> %shuffle
69 define <4 x double> @shuffle_v4f64_1000(<4 x double> %a, <4 x double> %b) {
70 ; AVX1-LABEL: shuffle_v4f64_1000:
72 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
73 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
74 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
77 ; AVX2-LABEL: shuffle_v4f64_1000:
79 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[1,0,0,0]
81 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 0, i32 0>
82 ret <4 x double> %shuffle
85 define <4 x double> @shuffle_v4f64_2200(<4 x double> %a, <4 x double> %b) {
86 ; AVX1-LABEL: shuffle_v4f64_2200:
88 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
89 ; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
92 ; AVX2-LABEL: shuffle_v4f64_2200:
94 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]
96 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 2, i32 2, i32 0, i32 0>
97 ret <4 x double> %shuffle
100 define <4 x double> @shuffle_v4f64_3330(<4 x double> %a, <4 x double> %b) {
101 ; AVX1-LABEL: shuffle_v4f64_3330:
103 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
104 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[1,1,2,2]
105 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,3,2]
106 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2],ymm1[3]
109 ; AVX2-LABEL: shuffle_v4f64_3330:
111 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,0]
113 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
114 ret <4 x double> %shuffle
117 define <4 x double> @shuffle_v4f64_3210(<4 x double> %a, <4 x double> %b) {
118 ; AVX1-LABEL: shuffle_v4f64_3210:
120 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
121 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
124 ; AVX2-LABEL: shuffle_v4f64_3210:
126 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,1,0]
128 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
129 ret <4 x double> %shuffle
132 define <4 x double> @shuffle_v4f64_0023(<4 x double> %a, <4 x double> %b) {
133 ; ALL-LABEL: shuffle_v4f64_0023:
135 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3]
137 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 3>
138 ret <4 x double> %shuffle
141 define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) {
142 ; ALL-LABEL: shuffle_v4f64_0022:
144 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
146 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
147 ret <4 x double> %shuffle
150 define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) {
151 ; ALL-LABEL: shuffle_v4f64_1032:
153 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
155 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
156 ret <4 x double> %shuffle
159 define <4 x double> @shuffle_v4f64_1133(<4 x double> %a, <4 x double> %b) {
160 ; ALL-LABEL: shuffle_v4f64_1133:
162 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3]
164 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
165 ret <4 x double> %shuffle
168 define <4 x double> @shuffle_v4f64_1023(<4 x double> %a, <4 x double> %b) {
169 ; ALL-LABEL: shuffle_v4f64_1023:
171 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3]
173 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 3>
174 ret <4 x double> %shuffle
177 define <4 x double> @shuffle_v4f64_1022(<4 x double> %a, <4 x double> %b) {
178 ; ALL-LABEL: shuffle_v4f64_1022:
180 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2]
182 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 2>
183 ret <4 x double> %shuffle
186 define <4 x double> @shuffle_v4f64_0423(<4 x double> %a, <4 x double> %b) {
187 ; AVX1-LABEL: shuffle_v4f64_0423:
189 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
190 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
193 ; AVX2-LABEL: shuffle_v4f64_0423:
195 ; AVX2-NEXT: vbroadcastsd %xmm1, %ymm1
196 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
198 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 3>
199 ret <4 x double> %shuffle
202 define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) {
203 ; ALL-LABEL: shuffle_v4f64_0462:
205 ; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
206 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
207 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3]
209 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 6, i32 2>
210 ret <4 x double> %shuffle
213 define <4 x double> @shuffle_v4f64_0426(<4 x double> %a, <4 x double> %b) {
214 ; ALL-LABEL: shuffle_v4f64_0426:
216 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
218 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
219 ret <4 x double> %shuffle
222 define <4 x double> @shuffle_v4f64_1537(<4 x double> %a, <4 x double> %b) {
223 ; ALL-LABEL: shuffle_v4f64_1537:
225 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
227 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
228 ret <4 x double> %shuffle
231 define <4 x double> @shuffle_v4f64_4062(<4 x double> %a, <4 x double> %b) {
232 ; ALL-LABEL: shuffle_v4f64_4062:
234 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
236 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 0, i32 6, i32 2>
237 ret <4 x double> %shuffle
240 define <4 x double> @shuffle_v4f64_5173(<4 x double> %a, <4 x double> %b) {
241 ; ALL-LABEL: shuffle_v4f64_5173:
243 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3]
245 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 7, i32 3>
246 ret <4 x double> %shuffle
249 define <4 x double> @shuffle_v4f64_5163(<4 x double> %a, <4 x double> %b) {
250 ; ALL-LABEL: shuffle_v4f64_5163:
252 ; ALL-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3]
254 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 6, i32 3>
255 ret <4 x double> %shuffle
258 define <4 x double> @shuffle_v4f64_0527(<4 x double> %a, <4 x double> %b) {
259 ; ALL-LABEL: shuffle_v4f64_0527:
261 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
263 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
264 ret <4 x double> %shuffle
267 define <4 x double> @shuffle_v4f64_4163(<4 x double> %a, <4 x double> %b) {
268 ; ALL-LABEL: shuffle_v4f64_4163:
270 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2],ymm0[3]
272 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
273 ret <4 x double> %shuffle
276 define <4 x double> @shuffle_v4f64_0145(<4 x double> %a, <4 x double> %b) {
277 ; ALL-LABEL: shuffle_v4f64_0145:
279 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
281 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
282 ret <4 x double> %shuffle
285 define <4 x double> @shuffle_v4f64_4501(<4 x double> %a, <4 x double> %b) {
286 ; ALL-LABEL: shuffle_v4f64_4501:
288 ; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
290 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
291 ret <4 x double> %shuffle
294 define <4 x double> @shuffle_v4f64_0167(<4 x double> %a, <4 x double> %b) {
295 ; ALL-LABEL: shuffle_v4f64_0167:
297 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
299 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
300 ret <4 x double> %shuffle
303 define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) {
304 ; ALL-LABEL: shuffle_v4f64_1054:
306 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
307 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
309 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
310 ret <4 x double> %shuffle
313 define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) {
314 ; ALL-LABEL: shuffle_v4f64_3254:
316 ; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
317 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
319 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
320 ret <4 x double> %shuffle
323 define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) {
324 ; ALL-LABEL: shuffle_v4f64_3276:
326 ; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
327 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
329 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
330 ret <4 x double> %shuffle
333 define <4 x double> @shuffle_v4f64_1076(<4 x double> %a, <4 x double> %b) {
334 ; ALL-LABEL: shuffle_v4f64_1076:
336 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
337 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
339 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
340 ret <4 x double> %shuffle
343 define <4 x double> @shuffle_v4f64_0415(<4 x double> %a, <4 x double> %b) {
344 ; AVX1-LABEL: shuffle_v4f64_0415:
346 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
347 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
348 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
351 ; AVX2-LABEL: shuffle_v4f64_0415:
353 ; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,0,2,1]
354 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
355 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
357 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
358 ret <4 x double> %shuffle
361 define <4 x i64> @shuffle_v4i64_0000(<4 x i64> %a, <4 x i64> %b) {
362 ; AVX1-LABEL: shuffle_v4i64_0000:
364 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
365 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
368 ; AVX2-LABEL: shuffle_v4i64_0000:
370 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
372 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
373 ret <4 x i64> %shuffle
376 define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {
377 ; AVX1-LABEL: shuffle_v4i64_0001:
379 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]
380 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
383 ; AVX2-LABEL: shuffle_v4i64_0001:
385 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
387 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
388 ret <4 x i64> %shuffle
391 define <4 x i64> @shuffle_v4i64_0020(<4 x i64> %a, <4 x i64> %b) {
392 ; AVX1-LABEL: shuffle_v4i64_0020:
394 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
395 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
396 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
397 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
400 ; AVX2-LABEL: shuffle_v4i64_0020:
402 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
404 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 0>
405 ret <4 x i64> %shuffle
408 define <4 x i64> @shuffle_v4i64_0112(<4 x i64> %a, <4 x i64> %b) {
409 ; AVX1-LABEL: shuffle_v4i64_0112:
411 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
412 ; AVX1-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1],xmm1[0]
413 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
416 ; AVX2-LABEL: shuffle_v4i64_0112:
418 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
420 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 1, i32 2>
421 ret <4 x i64> %shuffle
424 define <4 x i64> @shuffle_v4i64_0300(<4 x i64> %a, <4 x i64> %b) {
425 ; AVX1-LABEL: shuffle_v4i64_0300:
427 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
428 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,1,2,2]
429 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
432 ; AVX2-LABEL: shuffle_v4i64_0300:
434 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,3,0,0]
436 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
437 ret <4 x i64> %shuffle
440 define <4 x i64> @shuffle_v4i64_1000(<4 x i64> %a, <4 x i64> %b) {
441 ; AVX1-LABEL: shuffle_v4i64_1000:
443 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
444 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
445 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
448 ; AVX2-LABEL: shuffle_v4i64_1000:
450 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
452 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 0, i32 0>
453 ret <4 x i64> %shuffle
456 define <4 x i64> @shuffle_v4i64_2200(<4 x i64> %a, <4 x i64> %b) {
457 ; AVX1-LABEL: shuffle_v4i64_2200:
459 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
460 ; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
463 ; AVX2-LABEL: shuffle_v4i64_2200:
465 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,2,0,0]
467 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 2, i32 0, i32 0>
468 ret <4 x i64> %shuffle
471 define <4 x i64> @shuffle_v4i64_3330(<4 x i64> %a, <4 x i64> %b) {
472 ; AVX1-LABEL: shuffle_v4i64_3330:
474 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
475 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[1,1,2,2]
476 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,3,2]
477 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2],ymm1[3]
480 ; AVX2-LABEL: shuffle_v4i64_3330:
482 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,3,3,0]
484 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
485 ret <4 x i64> %shuffle
488 define <4 x i64> @shuffle_v4i64_3210(<4 x i64> %a, <4 x i64> %b) {
489 ; AVX1-LABEL: shuffle_v4i64_3210:
491 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
492 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
495 ; AVX2-LABEL: shuffle_v4i64_3210:
497 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,1,0]
499 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
500 ret <4 x i64> %shuffle
503 define <4 x i64> @shuffle_v4i64_0124(<4 x i64> %a, <4 x i64> %b) {
504 ; AVX1-LABEL: shuffle_v4i64_0124:
506 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm1[0,0]
507 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
508 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],ymm1[3]
511 ; AVX2-LABEL: shuffle_v4i64_0124:
513 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
514 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
516 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 4>
517 ret <4 x i64> %shuffle
520 define <4 x i64> @shuffle_v4i64_0142(<4 x i64> %a, <4 x i64> %b) {
521 ; AVX1-LABEL: shuffle_v4i64_0142:
523 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1
524 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,1,2,2]
525 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
528 ; AVX2-LABEL: shuffle_v4i64_0142:
530 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
531 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,2,2]
532 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
534 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 2>
535 ret <4 x i64> %shuffle
538 define <4 x i64> @shuffle_v4i64_0412(<4 x i64> %a, <4 x i64> %b) {
539 ; AVX1-LABEL: shuffle_v4i64_0412:
541 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
542 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
543 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
544 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
545 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
548 ; AVX2-LABEL: shuffle_v4i64_0412:
550 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
551 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
552 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
554 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 2>
555 ret <4 x i64> %shuffle
558 define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) {
559 ; AVX1-LABEL: shuffle_v4i64_4012:
561 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
562 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
563 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
564 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
565 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
568 ; AVX2-LABEL: shuffle_v4i64_4012:
570 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,2]
571 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5,6,7]
573 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 2>
574 ret <4 x i64> %shuffle
577 define <4 x i64> @shuffle_v4i64_0145(<4 x i64> %a, <4 x i64> %b) {
578 ; ALL-LABEL: shuffle_v4i64_0145:
580 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
582 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
583 ret <4 x i64> %shuffle
586 define <4 x i64> @shuffle_v4i64_0451(<4 x i64> %a, <4 x i64> %b) {
587 ; AVX1-LABEL: shuffle_v4i64_0451:
589 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm1[1],xmm0[1]
590 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
591 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
594 ; AVX2-LABEL: shuffle_v4i64_0451:
596 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
597 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
598 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
600 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 5, i32 1>
601 ret <4 x i64> %shuffle
604 define <4 x i64> @shuffle_v4i64_4501(<4 x i64> %a, <4 x i64> %b) {
605 ; ALL-LABEL: shuffle_v4i64_4501:
607 ; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
609 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
610 ret <4 x i64> %shuffle
613 define <4 x i64> @shuffle_v4i64_4015(<4 x i64> %a, <4 x i64> %b) {
614 ; AVX1-LABEL: shuffle_v4i64_4015:
616 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
617 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
618 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
621 ; AVX2-LABEL: shuffle_v4i64_4015:
623 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
624 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
625 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
627 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 5>
628 ret <4 x i64> %shuffle
631 define <4 x i64> @shuffle_v4i64_2u35(<4 x i64> %a, <4 x i64> %b) {
632 ; AVX1-LABEL: shuffle_v4i64_2u35:
634 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
635 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm1 = xmm0[1],xmm1[1]
636 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
639 ; AVX2-LABEL: shuffle_v4i64_2u35:
641 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
642 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,1,3,3]
643 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
645 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 undef, i32 3, i32 5>
646 ret <4 x i64> %shuffle
649 define <4 x i64> @shuffle_v4i64_1251(<4 x i64> %a, <4 x i64> %b) {
650 ; AVX1-LABEL: shuffle_v4i64_1251:
652 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm2 = ymm0[2,3,0,1]
653 ; AVX1-NEXT: vshufpd {{.*#+}} ymm0 = ymm0[1],ymm2[0],ymm0[2],ymm2[3]
654 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
655 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
656 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
659 ; AVX2-LABEL: shuffle_v4i64_1251:
661 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,3]
662 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,2,2,1]
663 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
665 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 2, i32 5, i32 1>
666 ret <4 x i64> %shuffle
669 define <4 x i64> @shuffle_v4i64_1054(<4 x i64> %a, <4 x i64> %b) {
670 ; AVX1-LABEL: shuffle_v4i64_1054:
672 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
673 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
676 ; AVX2-LABEL: shuffle_v4i64_1054:
678 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
679 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
681 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
682 ret <4 x i64> %shuffle
685 define <4 x i64> @shuffle_v4i64_3254(<4 x i64> %a, <4 x i64> %b) {
686 ; AVX1-LABEL: shuffle_v4i64_3254:
688 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
689 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
692 ; AVX2-LABEL: shuffle_v4i64_3254:
694 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
695 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
697 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
698 ret <4 x i64> %shuffle
701 define <4 x i64> @shuffle_v4i64_3276(<4 x i64> %a, <4 x i64> %b) {
702 ; AVX1-LABEL: shuffle_v4i64_3276:
704 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
705 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
708 ; AVX2-LABEL: shuffle_v4i64_3276:
710 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
711 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
713 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
714 ret <4 x i64> %shuffle
717 define <4 x i64> @shuffle_v4i64_1076(<4 x i64> %a, <4 x i64> %b) {
718 ; AVX1-LABEL: shuffle_v4i64_1076:
720 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
721 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
724 ; AVX2-LABEL: shuffle_v4i64_1076:
726 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
727 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
729 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
730 ret <4 x i64> %shuffle
733 define <4 x i64> @shuffle_v4i64_0415(<4 x i64> %a, <4 x i64> %b) {
734 ; AVX1-LABEL: shuffle_v4i64_0415:
736 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
737 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
738 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
741 ; AVX2-LABEL: shuffle_v4i64_0415:
743 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,2,1]
744 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,3]
745 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
747 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
748 ret <4 x i64> %shuffle
751 define <4 x i64> @stress_test1(<4 x i64> %a, <4 x i64> %b) {
752 ; ALL-LABEL: stress_test1:
754 %c = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 1, i32 1, i32 0>
755 %d = shufflevector <4 x i64> %c, <4 x i64> undef, <4 x i32> <i32 3, i32 undef, i32 2, i32 undef>
756 %e = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 undef>
757 %f = shufflevector <4 x i64> %d, <4 x i64> %e, <4 x i32> <i32 5, i32 1, i32 1, i32 0>
762 define <4 x i64> @insert_reg_and_zero_v4i64(i64 %a) {
763 ; AVX1-LABEL: insert_reg_and_zero_v4i64:
765 ; AVX1-NEXT: vmovq %rdi, %xmm0
766 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
767 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
770 ; AVX2-LABEL: insert_reg_and_zero_v4i64:
772 ; AVX2-NEXT: vmovq %rdi, %xmm0
773 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
774 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
776 %v = insertelement <4 x i64> undef, i64 %a, i64 0
777 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
778 ret <4 x i64> %shuffle
781 define <4 x i64> @insert_mem_and_zero_v4i64(i64* %ptr) {
782 ; AVX1-LABEL: insert_mem_and_zero_v4i64:
784 ; AVX1-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
785 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
786 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
789 ; AVX2-LABEL: insert_mem_and_zero_v4i64:
791 ; AVX2-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
792 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
793 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
796 %v = insertelement <4 x i64> undef, i64 %a, i64 0
797 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
798 ret <4 x i64> %shuffle
801 define <4 x double> @insert_reg_and_zero_v4f64(double %a) {
802 ; ALL-LABEL: insert_reg_and_zero_v4f64:
804 ; ALL-NEXT: vxorpd %xmm1, %xmm1, %xmm1
805 ; ALL-NEXT: vmovsd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
807 %v = insertelement <4 x double> undef, double %a, i32 0
808 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
809 ret <4 x double> %shuffle
812 define <4 x double> @insert_mem_and_zero_v4f64(double* %ptr) {
813 ; ALL-LABEL: insert_mem_and_zero_v4f64:
815 ; ALL-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
817 %a = load double* %ptr
818 %v = insertelement <4 x double> undef, double %a, i32 0
819 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
820 ret <4 x double> %shuffle
823 define <4 x double> @splat_mem_v4f64(double* %ptr) {
824 ; ALL-LABEL: splat_mem_v4f64:
826 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
828 %a = load double* %ptr
829 %v = insertelement <4 x double> undef, double %a, i32 0
830 %shuffle = shufflevector <4 x double> %v, <4 x double> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
831 ret <4 x double> %shuffle
834 define <4 x i64> @splat_mem_v4i64(i64* %ptr) {
835 ; AVX1-LABEL: splat_mem_v4i64:
837 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
838 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
841 ; AVX2-LABEL: splat_mem_v4i64:
843 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
846 %v = insertelement <4 x i64> undef, i64 %a, i64 0
847 %shuffle = shufflevector <4 x i64> %v, <4 x i64> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
848 ret <4 x i64> %shuffle
851 define <4 x double> @splat_mem_v4f64_2(double* %p) {
852 ; ALL-LABEL: splat_mem_v4f64_2:
854 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
857 %2 = insertelement <2 x double> undef, double %1, i32 0
858 %3 = shufflevector <2 x double> %2, <2 x double> undef, <4 x i32> zeroinitializer
862 define <4 x double> @splat_v4f64(<2 x double> %r) {
863 ; AVX1-LABEL: splat_v4f64:
865 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
866 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
869 ; AVX2-LABEL: splat_v4f64:
871 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
873 %1 = shufflevector <2 x double> %r, <2 x double> undef, <4 x i32> zeroinitializer