1 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx | FileCheck %s --check-prefix=ALL --check-prefix=AVX1
2 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=ALL --check-prefix=AVX2
3 ; RUN: llc < %s -mcpu=knl -mattr=+avx512vl | FileCheck %s --check-prefix=ALL --check-prefix=AVX512VL
5 target triple = "x86_64-unknown-unknown"
7 define <4 x double> @shuffle_v4f64_0000(<4 x double> %a, <4 x double> %b) {
8 ; AVX1-LABEL: shuffle_v4f64_0000:
10 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
11 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
14 ; AVX2-LABEL: shuffle_v4f64_0000:
16 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
19 ; AVX512VL-LABEL: shuffle_v4f64_0000:
21 ; AVX512VL-NEXT: vbroadcastsd %xmm0, %ymm0
23 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
24 ret <4 x double> %shuffle
27 define <4 x double> @shuffle_v4f64_0001(<4 x double> %a, <4 x double> %b) {
28 ; AVX1-LABEL: shuffle_v4f64_0001:
30 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]
31 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
34 ; AVX2-LABEL: shuffle_v4f64_0001:
36 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
39 ; AVX512VL-LABEL: shuffle_v4f64_0001:
41 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
43 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
44 ret <4 x double> %shuffle
47 define <4 x double> @shuffle_v4f64_0020(<4 x double> %a, <4 x double> %b) {
48 ; AVX1-LABEL: shuffle_v4f64_0020:
50 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
51 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
52 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
53 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
56 ; AVX2-LABEL: shuffle_v4f64_0020:
58 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,2,0]
61 ; AVX512VL-LABEL: shuffle_v4f64_0020:
63 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,2,0]
65 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 0>
66 ret <4 x double> %shuffle
69 define <4 x double> @shuffle_v4f64_0300(<4 x double> %a, <4 x double> %b) {
70 ; AVX1-LABEL: shuffle_v4f64_0300:
72 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
73 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,1,2,2]
74 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
77 ; AVX2-LABEL: shuffle_v4f64_0300:
79 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,3,0,0]
82 ; AVX512VL-LABEL: shuffle_v4f64_0300:
84 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,3,0,0]
86 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
87 ret <4 x double> %shuffle
90 define <4 x double> @shuffle_v4f64_1000(<4 x double> %a, <4 x double> %b) {
91 ; AVX1-LABEL: shuffle_v4f64_1000:
93 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
94 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
95 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
98 ; AVX2-LABEL: shuffle_v4f64_1000:
100 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[1,0,0,0]
103 ; AVX512VL-LABEL: shuffle_v4f64_1000:
105 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[1,0,0,0]
106 ; AVX512VL-NEXT: retq
107 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 0, i32 0>
108 ret <4 x double> %shuffle
111 define <4 x double> @shuffle_v4f64_2200(<4 x double> %a, <4 x double> %b) {
112 ; AVX1-LABEL: shuffle_v4f64_2200:
114 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
115 ; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
118 ; AVX2-LABEL: shuffle_v4f64_2200:
120 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]
123 ; AVX512VL-LABEL: shuffle_v4f64_2200:
125 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[2,2,0,0]
126 ; AVX512VL-NEXT: retq
127 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 2, i32 2, i32 0, i32 0>
128 ret <4 x double> %shuffle
131 define <4 x double> @shuffle_v4f64_3330(<4 x double> %a, <4 x double> %b) {
132 ; AVX1-LABEL: shuffle_v4f64_3330:
134 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
135 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1,2],ymm0[3]
136 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,2]
139 ; AVX2-LABEL: shuffle_v4f64_3330:
141 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,0]
144 ; AVX512VL-LABEL: shuffle_v4f64_3330:
146 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,0]
147 ; AVX512VL-NEXT: retq
148 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
149 ret <4 x double> %shuffle
152 define <4 x double> @shuffle_v4f64_3210(<4 x double> %a, <4 x double> %b) {
153 ; AVX1-LABEL: shuffle_v4f64_3210:
155 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
156 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
159 ; AVX2-LABEL: shuffle_v4f64_3210:
161 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,1,0]
164 ; AVX512VL-LABEL: shuffle_v4f64_3210:
166 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,1,0]
167 ; AVX512VL-NEXT: retq
168 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
169 ret <4 x double> %shuffle
172 define <4 x double> @shuffle_v4f64_0023(<4 x double> %a, <4 x double> %b) {
173 ; ALL-LABEL: shuffle_v4f64_0023:
175 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3]
178 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 3>
179 ret <4 x double> %shuffle
182 define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) {
183 ; ALL-LABEL: shuffle_v4f64_0022:
185 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
187 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
188 ret <4 x double> %shuffle
191 define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) {
192 ; ALL-LABEL: shuffle_v4f64_1032:
194 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
196 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
197 ret <4 x double> %shuffle
200 define <4 x double> @shuffle_v4f64_1133(<4 x double> %a, <4 x double> %b) {
201 ; ALL-LABEL: shuffle_v4f64_1133:
203 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3]
205 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
206 ret <4 x double> %shuffle
209 define <4 x double> @shuffle_v4f64_1023(<4 x double> %a, <4 x double> %b) {
210 ; ALL-LABEL: shuffle_v4f64_1023:
212 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3]
214 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 3>
215 ret <4 x double> %shuffle
218 define <4 x double> @shuffle_v4f64_1022(<4 x double> %a, <4 x double> %b) {
219 ; ALL-LABEL: shuffle_v4f64_1022:
221 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2]
223 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 2>
224 ret <4 x double> %shuffle
227 define <4 x double> @shuffle_v4f64_0423(<4 x double> %a, <4 x double> %b) {
228 ; AVX1-LABEL: shuffle_v4f64_0423:
230 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
231 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
234 ; AVX2-LABEL: shuffle_v4f64_0423:
236 ; AVX2-NEXT: vbroadcastsd %xmm1, %ymm1
237 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
240 ; AVX512VL-LABEL: shuffle_v4f64_0423:
242 ; AVX512VL-NEXT: vbroadcastsd %xmm1, %ymm1
243 ; AVX512VL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
244 ; AVX512VL-NEXT: retq
245 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 3>
246 ret <4 x double> %shuffle
249 define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) {
250 ; ALL-LABEL: shuffle_v4f64_0462:
252 ; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
253 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
254 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3]
256 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 6, i32 2>
257 ret <4 x double> %shuffle
260 define <4 x double> @shuffle_v4f64_0426(<4 x double> %a, <4 x double> %b) {
261 ; ALL-LABEL: shuffle_v4f64_0426:
263 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
265 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
266 ret <4 x double> %shuffle
269 define <4 x double> @shuffle_v4f64_1537(<4 x double> %a, <4 x double> %b) {
270 ; ALL-LABEL: shuffle_v4f64_1537:
272 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
274 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
275 ret <4 x double> %shuffle
278 define <4 x double> @shuffle_v4f64_4062(<4 x double> %a, <4 x double> %b) {
279 ; ALL-LABEL: shuffle_v4f64_4062:
281 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
283 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 0, i32 6, i32 2>
284 ret <4 x double> %shuffle
287 define <4 x double> @shuffle_v4f64_5173(<4 x double> %a, <4 x double> %b) {
288 ; ALL-LABEL: shuffle_v4f64_5173:
290 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3]
292 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 7, i32 3>
293 ret <4 x double> %shuffle
296 define <4 x double> @shuffle_v4f64_5163(<4 x double> %a, <4 x double> %b) {
297 ; ALL-LABEL: shuffle_v4f64_5163:
299 ; ALL-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3]
301 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 6, i32 3>
302 ret <4 x double> %shuffle
305 define <4 x double> @shuffle_v4f64_0527(<4 x double> %a, <4 x double> %b) {
306 ; ALL-LABEL: shuffle_v4f64_0527:
308 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
310 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
311 ret <4 x double> %shuffle
314 define <4 x double> @shuffle_v4f64_4163(<4 x double> %a, <4 x double> %b) {
315 ; ALL-LABEL: shuffle_v4f64_4163:
317 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2],ymm0[3]
319 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
320 ret <4 x double> %shuffle
323 define <4 x double> @shuffle_v4f64_0145(<4 x double> %a, <4 x double> %b) {
324 ; AVX1-LABEL: shuffle_v4f64_0145:
326 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
329 ; AVX2-LABEL: shuffle_v4f64_0145:
331 ; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
334 ; AVX512VL-LABEL: shuffle_v4f64_0145:
336 ; AVX512VL-NEXT: vinsertf32x4 $1, %xmm1, %ymm0, %ymm0
337 ; AVX512VL-NEXT: retq
338 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
339 ret <4 x double> %shuffle
342 define <4 x double> @shuffle_v4f64_4501(<4 x double> %a, <4 x double> %b) {
343 ; AVX1-LABEL: shuffle_v4f64_4501:
345 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
348 ; AVX2-LABEL: shuffle_v4f64_4501:
350 ; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
353 ; AVX512VL-LABEL: shuffle_v4f64_4501:
355 ; AVX512VL-NEXT: vinsertf32x4 $1, %xmm0, %ymm1, %ymm0
356 ; AVX512VL-NEXT: retq
357 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
358 ret <4 x double> %shuffle
361 define <4 x double> @shuffle_v4f64_0167(<4 x double> %a, <4 x double> %b) {
362 ; ALL-LABEL: shuffle_v4f64_0167:
364 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
366 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
367 ret <4 x double> %shuffle
370 define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) {
371 ; AVX1-LABEL: shuffle_v4f64_1054:
373 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
374 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
377 ; AVX2-LABEL: shuffle_v4f64_1054:
379 ; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
380 ; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
383 ; AVX512VL-LABEL: shuffle_v4f64_1054:
385 ; AVX512VL-NEXT: vinsertf32x4 $1, %xmm1, %ymm0, %ymm0
386 ; AVX512VL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
387 ; AVX512VL-NEXT: retq
388 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
389 ret <4 x double> %shuffle
392 define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) {
393 ; AVX1-LABEL: shuffle_v4f64_3254:
395 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
396 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
399 ; AVX2-LABEL: shuffle_v4f64_3254:
401 ; AVX2-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
402 ; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
405 ; AVX512VL-LABEL: shuffle_v4f64_3254:
407 ; AVX512VL-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
408 ; AVX512VL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
409 ; AVX512VL-NEXT: retq
410 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
411 ret <4 x double> %shuffle
414 define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) {
415 ; AVX1-LABEL: shuffle_v4f64_3276:
417 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
418 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
421 ; AVX2-LABEL: shuffle_v4f64_3276:
423 ; AVX2-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
424 ; AVX2-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
427 ; AVX512VL-LABEL: shuffle_v4f64_3276:
429 ; AVX512VL-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
430 ; AVX512VL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
431 ; AVX512VL-NEXT: retq
432 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
433 ret <4 x double> %shuffle
436 define <4 x double> @shuffle_v4f64_1076(<4 x double> %a, <4 x double> %b) {
437 ; ALL-LABEL: shuffle_v4f64_1076:
439 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
440 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
442 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
443 ret <4 x double> %shuffle
446 define <4 x double> @shuffle_v4f64_0415(<4 x double> %a, <4 x double> %b) {
447 ; AVX1-LABEL: shuffle_v4f64_0415:
449 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
450 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
451 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
454 ; AVX2-LABEL: shuffle_v4f64_0415:
456 ; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,0,2,1]
457 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
458 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
461 ; AVX512VL-LABEL: shuffle_v4f64_0415:
463 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,0,2,1]
464 ; AVX512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
465 ; AVX512VL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
466 ; AVX512VL-NEXT: retq
467 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
468 ret <4 x double> %shuffle
471 define <4 x double> @shuffle_v4f64_u062(<4 x double> %a, <4 x double> %b) {
472 ; ALL-LABEL: shuffle_v4f64_u062:
474 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
476 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 undef, i32 0, i32 6, i32 2>
477 ret <4 x double> %shuffle
480 define <4 x i64> @shuffle_v4i64_0000(<4 x i64> %a, <4 x i64> %b) {
481 ; AVX1-LABEL: shuffle_v4i64_0000:
483 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
484 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
487 ; AVX2-LABEL: shuffle_v4i64_0000:
489 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
492 ; AVX512VL-LABEL: shuffle_v4i64_0000:
494 ; AVX512VL-NEXT: vpbroadcastq %xmm0, %ymm0
495 ; AVX512VL-NEXT: retq
496 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
497 ret <4 x i64> %shuffle
500 define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {
501 ; AVX1-LABEL: shuffle_v4i64_0001:
503 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]
504 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
507 ; AVX2-LABEL: shuffle_v4i64_0001:
509 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
512 ; AVX512VL-LABEL: shuffle_v4i64_0001:
514 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
515 ; AVX512VL-NEXT: retq
516 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
517 ret <4 x i64> %shuffle
520 define <4 x i64> @shuffle_v4i64_0020(<4 x i64> %a, <4 x i64> %b) {
521 ; AVX1-LABEL: shuffle_v4i64_0020:
523 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
524 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
525 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
526 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
529 ; AVX2-LABEL: shuffle_v4i64_0020:
531 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
534 ; AVX512VL-LABEL: shuffle_v4i64_0020:
536 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
537 ; AVX512VL-NEXT: retq
538 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 0>
539 ret <4 x i64> %shuffle
542 define <4 x i64> @shuffle_v4i64_0112(<4 x i64> %a, <4 x i64> %b) {
543 ; AVX1-LABEL: shuffle_v4i64_0112:
545 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
546 ; AVX1-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1],xmm1[0]
547 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
550 ; AVX2-LABEL: shuffle_v4i64_0112:
552 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
555 ; AVX512VL-LABEL: shuffle_v4i64_0112:
557 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
558 ; AVX512VL-NEXT: retq
559 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 1, i32 2>
560 ret <4 x i64> %shuffle
563 define <4 x i64> @shuffle_v4i64_0300(<4 x i64> %a, <4 x i64> %b) {
564 ; AVX1-LABEL: shuffle_v4i64_0300:
566 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
567 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,1,2,2]
568 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
571 ; AVX2-LABEL: shuffle_v4i64_0300:
573 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,3,0,0]
576 ; AVX512VL-LABEL: shuffle_v4i64_0300:
578 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,3,0,0]
579 ; AVX512VL-NEXT: retq
580 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
581 ret <4 x i64> %shuffle
584 define <4 x i64> @shuffle_v4i64_1000(<4 x i64> %a, <4 x i64> %b) {
585 ; AVX1-LABEL: shuffle_v4i64_1000:
587 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
588 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
589 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
592 ; AVX2-LABEL: shuffle_v4i64_1000:
594 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
597 ; AVX512VL-LABEL: shuffle_v4i64_1000:
599 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
600 ; AVX512VL-NEXT: retq
601 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 0, i32 0>
602 ret <4 x i64> %shuffle
605 define <4 x i64> @shuffle_v4i64_2200(<4 x i64> %a, <4 x i64> %b) {
606 ; AVX1-LABEL: shuffle_v4i64_2200:
608 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
609 ; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
612 ; AVX2-LABEL: shuffle_v4i64_2200:
614 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,2,0,0]
617 ; AVX512VL-LABEL: shuffle_v4i64_2200:
619 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,2,0,0]
620 ; AVX512VL-NEXT: retq
621 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 2, i32 0, i32 0>
622 ret <4 x i64> %shuffle
625 define <4 x i64> @shuffle_v4i64_3330(<4 x i64> %a, <4 x i64> %b) {
626 ; AVX1-LABEL: shuffle_v4i64_3330:
628 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
629 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1,2],ymm0[3]
630 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,2]
633 ; AVX2-LABEL: shuffle_v4i64_3330:
635 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,3,3,0]
638 ; AVX512VL-LABEL: shuffle_v4i64_3330:
640 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,3,3,0]
641 ; AVX512VL-NEXT: retq
642 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
643 ret <4 x i64> %shuffle
646 define <4 x i64> @shuffle_v4i64_3210(<4 x i64> %a, <4 x i64> %b) {
647 ; AVX1-LABEL: shuffle_v4i64_3210:
649 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
650 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
653 ; AVX2-LABEL: shuffle_v4i64_3210:
655 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,1,0]
658 ; AVX512VL-LABEL: shuffle_v4i64_3210:
660 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,1,0]
661 ; AVX512VL-NEXT: retq
662 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
663 ret <4 x i64> %shuffle
666 define <4 x i64> @shuffle_v4i64_0124(<4 x i64> %a, <4 x i64> %b) {
667 ; AVX1-LABEL: shuffle_v4i64_0124:
669 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm1[0,0]
670 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
671 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],ymm1[3]
674 ; AVX2-LABEL: shuffle_v4i64_0124:
676 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
677 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
680 ; AVX512VL-LABEL: shuffle_v4i64_0124:
682 ; AVX512VL-NEXT: vpbroadcastq %xmm1, %ymm1
683 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
684 ; AVX512VL-NEXT: retq
685 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 4>
686 ret <4 x i64> %shuffle
689 define <4 x i64> @shuffle_v4i64_0142(<4 x i64> %a, <4 x i64> %b) {
690 ; AVX1-LABEL: shuffle_v4i64_0142:
692 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1
693 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,1,2,2]
694 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
697 ; AVX2-LABEL: shuffle_v4i64_0142:
699 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
700 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,2,2]
701 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
704 ; AVX512VL-LABEL: shuffle_v4i64_0142:
706 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm1, %ymm1, %ymm1
707 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,2,2]
708 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
709 ; AVX512VL-NEXT: retq
710 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 2>
711 ret <4 x i64> %shuffle
714 define <4 x i64> @shuffle_v4i64_0412(<4 x i64> %a, <4 x i64> %b) {
715 ; AVX1-LABEL: shuffle_v4i64_0412:
717 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
718 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
719 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
720 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
721 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
724 ; AVX2-LABEL: shuffle_v4i64_0412:
726 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
727 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
728 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
731 ; AVX512VL-LABEL: shuffle_v4i64_0412:
733 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
734 ; AVX512VL-NEXT: vpbroadcastq %xmm1, %ymm1
735 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
736 ; AVX512VL-NEXT: retq
737 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 2>
738 ret <4 x i64> %shuffle
741 define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) {
742 ; AVX1-LABEL: shuffle_v4i64_4012:
744 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
745 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
746 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
747 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
748 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
751 ; AVX2-LABEL: shuffle_v4i64_4012:
753 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,2]
754 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5,6,7]
757 ; AVX512VL-LABEL: shuffle_v4i64_4012:
759 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,2]
760 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5,6,7]
761 ; AVX512VL-NEXT: retq
762 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 2>
763 ret <4 x i64> %shuffle
766 define <4 x i64> @shuffle_v4i64_0145(<4 x i64> %a, <4 x i64> %b) {
767 ; AVX1-LABEL: shuffle_v4i64_0145:
769 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
772 ; AVX2-LABEL: shuffle_v4i64_0145:
774 ; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
777 ; AVX512VL-LABEL: shuffle_v4i64_0145:
779 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm1, %ymm0, %ymm0
780 ; AVX512VL-NEXT: retq
781 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
782 ret <4 x i64> %shuffle
785 define <4 x i64> @shuffle_v4i64_0451(<4 x i64> %a, <4 x i64> %b) {
786 ; AVX1-LABEL: shuffle_v4i64_0451:
788 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm1[1],xmm0[1]
789 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
790 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
793 ; AVX2-LABEL: shuffle_v4i64_0451:
795 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
796 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
797 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
800 ; AVX512VL-LABEL: shuffle_v4i64_0451:
802 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm0, %ymm0, %ymm0
803 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
804 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
805 ; AVX512VL-NEXT: retq
806 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 5, i32 1>
807 ret <4 x i64> %shuffle
810 define <4 x i64> @shuffle_v4i64_4501(<4 x i64> %a, <4 x i64> %b) {
811 ; AVX1-LABEL: shuffle_v4i64_4501:
813 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
816 ; AVX2-LABEL: shuffle_v4i64_4501:
818 ; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
821 ; AVX512VL-LABEL: shuffle_v4i64_4501:
823 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm0, %ymm1, %ymm0
824 ; AVX512VL-NEXT: retq
825 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
826 ret <4 x i64> %shuffle
829 define <4 x i64> @shuffle_v4i64_4015(<4 x i64> %a, <4 x i64> %b) {
830 ; AVX1-LABEL: shuffle_v4i64_4015:
832 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
833 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
834 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
837 ; AVX2-LABEL: shuffle_v4i64_4015:
839 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
840 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
841 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
844 ; AVX512VL-LABEL: shuffle_v4i64_4015:
846 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm1, %ymm1, %ymm1
847 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
848 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
849 ; AVX512VL-NEXT: retq
850 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 5>
851 ret <4 x i64> %shuffle
854 define <4 x i64> @shuffle_v4i64_2u35(<4 x i64> %a, <4 x i64> %b) {
855 ; AVX1-LABEL: shuffle_v4i64_2u35:
857 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
858 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm1 = xmm0[1],xmm1[1]
859 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
862 ; AVX2-LABEL: shuffle_v4i64_2u35:
864 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
865 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,1,3,1]
868 ; AVX512VL-LABEL: shuffle_v4i64_2u35:
870 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
871 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,1,3,1]
872 ; AVX512VL-NEXT: retq
873 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 undef, i32 3, i32 5>
874 ret <4 x i64> %shuffle
877 define <4 x i64> @shuffle_v4i64_1251(<4 x i64> %a, <4 x i64> %b) {
878 ; AVX1-LABEL: shuffle_v4i64_1251:
880 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm2 = ymm0[2,3,0,1]
881 ; AVX1-NEXT: vshufpd {{.*#+}} ymm0 = ymm0[1],ymm2[0],ymm0[2],ymm2[3]
882 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
883 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
884 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
887 ; AVX2-LABEL: shuffle_v4i64_1251:
889 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,3]
890 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,2,2,1]
891 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
894 ; AVX512VL-LABEL: shuffle_v4i64_1251:
896 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,3]
897 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,2,2,1]
898 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
899 ; AVX512VL-NEXT: retq
900 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 2, i32 5, i32 1>
901 ret <4 x i64> %shuffle
904 define <4 x i64> @shuffle_v4i64_1054(<4 x i64> %a, <4 x i64> %b) {
905 ; AVX1-LABEL: shuffle_v4i64_1054:
907 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
908 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
911 ; AVX2-LABEL: shuffle_v4i64_1054:
913 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
914 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
917 ; AVX512VL-LABEL: shuffle_v4i64_1054:
919 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm1, %ymm0, %ymm0
920 ; AVX512VL-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
921 ; AVX512VL-NEXT: retq
922 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
923 ret <4 x i64> %shuffle
926 define <4 x i64> @shuffle_v4i64_3254(<4 x i64> %a, <4 x i64> %b) {
927 ; AVX1-LABEL: shuffle_v4i64_3254:
929 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
930 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
933 ; AVX2-LABEL: shuffle_v4i64_3254:
935 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
936 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
939 ; AVX512VL-LABEL: shuffle_v4i64_3254:
941 ; AVX512VL-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
942 ; AVX512VL-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
943 ; AVX512VL-NEXT: retq
944 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
945 ret <4 x i64> %shuffle
948 define <4 x i64> @shuffle_v4i64_3276(<4 x i64> %a, <4 x i64> %b) {
949 ; AVX1-LABEL: shuffle_v4i64_3276:
951 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
952 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
955 ; AVX2-LABEL: shuffle_v4i64_3276:
957 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
958 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
961 ; AVX512VL-LABEL: shuffle_v4i64_3276:
963 ; AVX512VL-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
964 ; AVX512VL-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
965 ; AVX512VL-NEXT: retq
966 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
967 ret <4 x i64> %shuffle
970 define <4 x i64> @shuffle_v4i64_1076(<4 x i64> %a, <4 x i64> %b) {
971 ; AVX1-LABEL: shuffle_v4i64_1076:
973 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
974 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
977 ; AVX2-LABEL: shuffle_v4i64_1076:
979 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
980 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
983 ; AVX512VL-LABEL: shuffle_v4i64_1076:
985 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
986 ; AVX512VL-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
987 ; AVX512VL-NEXT: retq
988 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
989 ret <4 x i64> %shuffle
992 define <4 x i64> @shuffle_v4i64_0415(<4 x i64> %a, <4 x i64> %b) {
993 ; AVX1-LABEL: shuffle_v4i64_0415:
995 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
996 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
997 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
1000 ; AVX2-LABEL: shuffle_v4i64_0415:
1002 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,2,1]
1003 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,3]
1004 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
1007 ; AVX512VL-LABEL: shuffle_v4i64_0415:
1009 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,2,1]
1010 ; AVX512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,3]
1011 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
1012 ; AVX512VL-NEXT: retq
1013 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
1014 ret <4 x i64> %shuffle
1017 define <4 x i64> @shuffle_v4i64_z4z6(<4 x i64> %a) {
1018 ; AVX1-LABEL: shuffle_v4i64_z4z6:
1020 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
1021 ; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
1024 ; AVX2-LABEL: shuffle_v4i64_z4z6:
1026 ; AVX2-NEXT: vpslldq {{.*#+}} ymm0 = zero,zero,zero,zero,zero,zero,zero,zero,ymm0[0,1,2,3,4,5,6,7],zero,zero,zero,zero,zero,zero,zero,zero,ymm0[16,17,18,19,20,21,22,23]
1029 ; AVX512VL-LABEL: shuffle_v4i64_z4z6:
1031 ; AVX512VL-NEXT: vpslldq {{.*#+}} ymm0 = zero,zero,zero,zero,zero,zero,zero,zero,ymm0[0,1,2,3,4,5,6,7],zero,zero,zero,zero,zero,zero,zero,zero,ymm0[16,17,18,19,20,21,22,23]
1032 ; AVX512VL-NEXT: retq
1033 %shuffle = shufflevector <4 x i64> zeroinitializer, <4 x i64> %a, <4 x i32> <i32 0, i32 4, i32 0, i32 6>
1034 ret <4 x i64> %shuffle
1037 define <4 x i64> @shuffle_v4i64_5zuz(<4 x i64> %a) {
1038 ; AVX1-LABEL: shuffle_v4i64_5zuz:
1040 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
1041 ; AVX1-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
1044 ; AVX2-LABEL: shuffle_v4i64_5zuz:
1046 ; AVX2-NEXT: vpsrldq {{.*#+}} ymm0 = ymm0[8,9,10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,ymm0[24,25,26,27,28,29,30,31],zero,zero,zero,zero,zero,zero,zero,zero
1049 ; AVX512VL-LABEL: shuffle_v4i64_5zuz:
1051 ; AVX512VL-NEXT: vpsrldq {{.*#+}} ymm0 = ymm0[8,9,10,11,12,13,14,15],zero,zero,zero,zero,zero,zero,zero,zero,ymm0[24,25,26,27,28,29,30,31],zero,zero,zero,zero,zero,zero,zero,zero
1052 ; AVX512VL-NEXT: retq
1053 %shuffle = shufflevector <4 x i64> zeroinitializer, <4 x i64> %a, <4 x i32> <i32 5, i32 0, i32 undef, i32 0>
1054 ret <4 x i64> %shuffle
1057 define <4 x i64> @shuffle_v4i64_40u2(<4 x i64> %a, <4 x i64> %b) {
1058 ; AVX1-LABEL: shuffle_v4i64_40u2:
1060 ; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
1063 ; AVX2-LABEL: shuffle_v4i64_40u2:
1065 ; AVX2-NEXT: vpunpcklqdq {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
1068 ; AVX512VL-LABEL: shuffle_v4i64_40u2:
1070 ; AVX512VL-NEXT: vpunpcklqdq {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
1071 ; AVX512VL-NEXT: retq
1072 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 undef, i32 2>
1073 ret <4 x i64> %shuffle
1076 define <4 x i64> @stress_test1(<4 x i64> %a, <4 x i64> %b) {
1077 ; ALL-LABEL: stress_test1:
1079 %c = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 1, i32 1, i32 0>
1080 %d = shufflevector <4 x i64> %c, <4 x i64> undef, <4 x i32> <i32 3, i32 undef, i32 2, i32 undef>
1081 %e = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 undef>
1082 %f = shufflevector <4 x i64> %d, <4 x i64> %e, <4 x i32> <i32 5, i32 1, i32 1, i32 0>
1087 define <4 x i64> @insert_reg_and_zero_v4i64(i64 %a) {
1088 ; ALL-LABEL: insert_reg_and_zero_v4i64:
1090 ; ALL-NEXT: vmovq %rdi, %xmm0
1092 %v = insertelement <4 x i64> undef, i64 %a, i64 0
1093 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
1094 ret <4 x i64> %shuffle
1097 define <4 x i64> @insert_mem_and_zero_v4i64(i64* %ptr) {
1098 ; AVX1-LABEL: insert_mem_and_zero_v4i64:
1100 ; AVX1-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
1103 ; AVX2-LABEL: insert_mem_and_zero_v4i64:
1105 ; AVX2-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
1108 ; AVX512VL-LABEL: insert_mem_and_zero_v4i64:
1110 ; AVX512VL-NEXT: vmovq (%rdi), %xmm0
1111 ; AVX512VL-NEXT: retq
1112 %a = load i64, i64* %ptr
1113 %v = insertelement <4 x i64> undef, i64 %a, i64 0
1114 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
1115 ret <4 x i64> %shuffle
1118 define <4 x double> @insert_reg_and_zero_v4f64(double %a) {
1119 ; AVX1-LABEL: insert_reg_and_zero_v4f64:
1121 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
1122 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
1125 ; AVX2-LABEL: insert_reg_and_zero_v4f64:
1127 ; AVX2-NEXT: vxorpd %ymm1, %ymm1, %ymm1
1128 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
1131 ; AVX512VL-LABEL: insert_reg_and_zero_v4f64:
1133 ; AVX512VL-NEXT: vxorpd %xmm1, %xmm1, %xmm1
1134 ; AVX512VL-NEXT: vmovsd %xmm0, %xmm1, %xmm0
1135 ; AVX512VL-NEXT: retq
1136 %v = insertelement <4 x double> undef, double %a, i32 0
1137 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
1138 ret <4 x double> %shuffle
1141 define <4 x double> @insert_mem_and_zero_v4f64(double* %ptr) {
1142 ; AVX1-LABEL: insert_mem_and_zero_v4f64:
1144 ; AVX1-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
1147 ; AVX2-LABEL: insert_mem_and_zero_v4f64:
1149 ; AVX2-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
1152 ; AVX512VL-LABEL: insert_mem_and_zero_v4f64:
1154 ; AVX512VL-NEXT: vmovsd (%rdi), %xmm0
1155 ; AVX512VL-NEXT: retq
1156 %a = load double, double* %ptr
1157 %v = insertelement <4 x double> undef, double %a, i32 0
1158 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
1159 ret <4 x double> %shuffle
1162 define <4 x double> @splat_mem_v4f64(double* %ptr) {
1163 ; ALL-LABEL: splat_mem_v4f64:
1165 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
1167 %a = load double, double* %ptr
1168 %v = insertelement <4 x double> undef, double %a, i32 0
1169 %shuffle = shufflevector <4 x double> %v, <4 x double> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
1170 ret <4 x double> %shuffle
1173 define <4 x i64> @splat_mem_v4i64(i64* %ptr) {
1174 ; AVX1-LABEL: splat_mem_v4i64:
1176 ; AVX1-NEXT: vbroadcastsd (%rdi), %ymm0
1179 ; AVX2-LABEL: splat_mem_v4i64:
1181 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
1184 ; AVX512VL-LABEL: splat_mem_v4i64:
1186 ; AVX512VL-NEXT: vpbroadcastq (%rdi), %ymm0
1187 ; AVX512VL-NEXT: retq
1188 %a = load i64, i64* %ptr
1189 %v = insertelement <4 x i64> undef, i64 %a, i64 0
1190 %shuffle = shufflevector <4 x i64> %v, <4 x i64> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
1191 ret <4 x i64> %shuffle
1194 define <4 x double> @splat_mem_v4f64_2(double* %p) {
1195 ; ALL-LABEL: splat_mem_v4f64_2:
1197 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
1199 %1 = load double, double* %p
1200 %2 = insertelement <2 x double> undef, double %1, i32 0
1201 %3 = shufflevector <2 x double> %2, <2 x double> undef, <4 x i32> zeroinitializer
1205 define <4 x double> @splat_v4f64(<2 x double> %r) {
1206 ; AVX1-LABEL: splat_v4f64:
1208 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
1209 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1212 ; AVX2-LABEL: splat_v4f64:
1214 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
1217 ; AVX512VL-LABEL: splat_v4f64:
1219 ; AVX512VL-NEXT: vbroadcastsd %xmm0, %ymm0
1220 ; AVX512VL-NEXT: retq
1221 %1 = shufflevector <2 x double> %r, <2 x double> undef, <4 x i32> zeroinitializer
1225 define <4 x i64> @splat_mem_v4i64_from_v2i64(<2 x i64>* %ptr) {
1226 ; AVX1-LABEL: splat_mem_v4i64_from_v2i64:
1228 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
1229 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1232 ; AVX2-LABEL: splat_mem_v4i64_from_v2i64:
1234 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
1237 ; AVX512VL-LABEL: splat_mem_v4i64_from_v2i64:
1239 ; AVX512VL-NEXT: vmovdqa64 (%rdi), %xmm0
1240 ; AVX512VL-NEXT: vpbroadcastq %xmm0, %ymm0
1241 ; AVX512VL-NEXT: retq
1242 %v = load <2 x i64>, <2 x i64>* %ptr
1243 %shuffle = shufflevector <2 x i64> %v, <2 x i64> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
1244 ret <4 x i64> %shuffle
1247 define <4 x double> @splat_mem_v4f64_from_v2f64(<2 x double>* %ptr) {
1248 ; AVX1-LABEL: splat_mem_v4f64_from_v2f64:
1250 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
1251 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1254 ; AVX2-LABEL: splat_mem_v4f64_from_v2f64:
1256 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
1259 ; AVX512VL-LABEL: splat_mem_v4f64_from_v2f64:
1261 ; AVX512VL-NEXT: vbroadcastsd (%rdi), %ymm0
1262 ; AVX512VL-NEXT: retq
1263 %v = load <2 x double>, <2 x double>* %ptr
1264 %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
1265 ret <4 x double> %shuffle
1268 define <4 x i64> @splat128_mem_v4i64_from_v2i64(<2 x i64>* %ptr) {
1269 ; AVX1-LABEL: splat128_mem_v4i64_from_v2i64:
1271 ; AVX1-NEXT: vmovaps (%rdi), %xmm0
1272 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1275 ; AVX2-LABEL: splat128_mem_v4i64_from_v2i64:
1277 ; AVX2-NEXT: vmovaps (%rdi), %xmm0
1278 ; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1281 ; AVX512VL-LABEL: splat128_mem_v4i64_from_v2i64:
1283 ; AVX512VL-NEXT: vmovdqa64 (%rdi), %xmm0
1284 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm0, %ymm0, %ymm0
1285 ; AVX512VL-NEXT: retq
1286 %v = load <2 x i64>, <2 x i64>* %ptr
1287 %shuffle = shufflevector <2 x i64> %v, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
1288 ret <4 x i64> %shuffle
1291 define <4 x double> @splat128_mem_v4f64_from_v2f64(<2 x double>* %ptr) {
1292 ; AVX1-LABEL: splat128_mem_v4f64_from_v2f64:
1294 ; AVX1-NEXT: vmovaps (%rdi), %xmm0
1295 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1298 ; AVX2-LABEL: splat128_mem_v4f64_from_v2f64:
1300 ; AVX2-NEXT: vmovaps (%rdi), %xmm0
1301 ; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
1304 ; AVX512VL-LABEL: splat128_mem_v4f64_from_v2f64:
1306 ; AVX512VL-NEXT: vmovapd (%rdi), %xmm0
1307 ; AVX512VL-NEXT: vinsertf32x4 $1, %xmm0, %ymm0, %ymm0
1308 ; AVX512VL-NEXT: retq
1309 %v = load <2 x double>, <2 x double>* %ptr
1310 %shuffle = shufflevector <2 x double> %v, <2 x double> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
1311 ret <4 x double> %shuffle
1314 define <4 x double> @bitcast_v4f64_0426(<4 x double> %a, <4 x double> %b) {
1315 ; AVX1-LABEL: bitcast_v4f64_0426:
1317 ; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
1320 ; AVX2-LABEL: bitcast_v4f64_0426:
1322 ; AVX2-NEXT: vpunpcklqdq {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
1325 ; AVX512VL-LABEL: bitcast_v4f64_0426:
1327 ; AVX512VL-NEXT: vpunpcklqdq {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
1328 ; AVX512VL-NEXT: retq
1329 %shuffle64 = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 0, i32 6, i32 2>
1330 %bitcast32 = bitcast <4 x double> %shuffle64 to <8 x float>
1331 %shuffle32 = shufflevector <8 x float> %bitcast32, <8 x float> undef, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
1332 %bitcast16 = bitcast <8 x float> %shuffle32 to <16 x i16>
1333 %shuffle16 = shufflevector <16 x i16> %bitcast16, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 0, i32 1, i32 6, i32 7, i32 4, i32 5, i32 10, i32 11, i32 8, i32 9, i32 14, i32 15, i32 12, i32 13>
1334 %bitcast64 = bitcast <16 x i16> %shuffle16 to <4 x double>
1335 ret <4 x double> %bitcast64
1338 define <4 x i64> @concat_v4i64_0167(<4 x i64> %a0, <4 x i64> %a1) {
1339 ; AVX1-LABEL: concat_v4i64_0167:
1341 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
1344 ; AVX2-LABEL: concat_v4i64_0167:
1346 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
1349 ; AVX512VL-LABEL: concat_v4i64_0167:
1351 ; AVX512VL-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
1352 ; AVX512VL-NEXT: retq
1353 %a0lo = shufflevector <4 x i64> %a0, <4 x i64> %a1, <2 x i32> <i32 0, i32 1>
1354 %a1hi = shufflevector <4 x i64> %a0, <4 x i64> %a1, <2 x i32> <i32 6, i32 7>
1355 %shuffle64 = shufflevector <2 x i64> %a0lo, <2 x i64> %a1hi, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
1356 ret <4 x i64> %shuffle64
1359 define <4 x i64> @concat_v4i64_0145_bc(<4 x i64> %a0, <4 x i64> %a1) {
1360 ; AVX1-LABEL: concat_v4i64_0145_bc:
1362 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1365 ; AVX2-LABEL: concat_v4i64_0145_bc:
1367 ; AVX2-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
1370 ; AVX512VL-LABEL: concat_v4i64_0145_bc:
1372 ; AVX512VL-NEXT: vinserti32x4 $1, %xmm1, %ymm0, %ymm0
1373 ; AVX512VL-NEXT: retq
1374 %a0lo = shufflevector <4 x i64> %a0, <4 x i64> %a1, <2 x i32> <i32 0, i32 1>
1375 %a1lo = shufflevector <4 x i64> %a0, <4 x i64> %a1, <2 x i32> <i32 4, i32 5>
1376 %bc0lo = bitcast <2 x i64> %a0lo to <4 x i32>
1377 %bc1lo = bitcast <2 x i64> %a1lo to <4 x i32>
1378 %shuffle32 = shufflevector <4 x i32> %bc0lo, <4 x i32> %bc1lo, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
1379 %shuffle64 = bitcast <8 x i32> %shuffle32 to <4 x i64>
1380 ret <4 x i64> %shuffle64
1383 define <4 x i64> @insert_dup_mem_v4i64(i64* %ptr) {
1384 ; AVX1-LABEL: insert_dup_mem_v4i64:
1386 ; AVX1-NEXT: vbroadcastsd (%rdi), %ymm0
1389 ; AVX2-LABEL: insert_dup_mem_v4i64:
1391 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
1394 ; AVX512VL-LABEL: insert_dup_mem_v4i64:
1396 ; AVX512VL-NEXT: vpbroadcastq (%rdi), %ymm0
1397 ; AVX512VL-NEXT: retq
1398 %tmp = load i64, i64* %ptr, align 1
1399 %tmp1 = insertelement <2 x i64> undef, i64 %tmp, i32 0
1400 %tmp2 = shufflevector <2 x i64> %tmp1, <2 x i64> undef, <4 x i32> zeroinitializer