1 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx -x86-experimental-vector-shuffle-legality | FileCheck %s --check-prefix=ALL --check-prefix=AVX --check-prefix=AVX1
2 ; RUN: llc < %s -mcpu=x86-64 -mattr=+avx2 -x86-experimental-vector-shuffle-legality | 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: vblendpd {{.*#+}} ymm0 = ymm1[0,1,2],ymm0[3]
105 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,2]
108 ; AVX2-LABEL: shuffle_v4f64_3330:
110 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,3,3,0]
112 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
113 ret <4 x double> %shuffle
116 define <4 x double> @shuffle_v4f64_3210(<4 x double> %a, <4 x double> %b) {
117 ; AVX1-LABEL: shuffle_v4f64_3210:
119 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
120 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
123 ; AVX2-LABEL: shuffle_v4f64_3210:
125 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[3,2,1,0]
127 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
128 ret <4 x double> %shuffle
131 define <4 x double> @shuffle_v4f64_0023(<4 x double> %a, <4 x double> %b) {
132 ; ALL-LABEL: shuffle_v4f64_0023:
134 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,0,2,3]
136 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 3>
137 ret <4 x double> %shuffle
140 define <4 x double> @shuffle_v4f64_0022(<4 x double> %a, <4 x double> %b) {
141 ; ALL-LABEL: shuffle_v4f64_0022:
143 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
145 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
146 ret <4 x double> %shuffle
149 define <4 x double> @shuffle_v4f64_1032(<4 x double> %a, <4 x double> %b) {
150 ; ALL-LABEL: shuffle_v4f64_1032:
152 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
154 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
155 ret <4 x double> %shuffle
158 define <4 x double> @shuffle_v4f64_1133(<4 x double> %a, <4 x double> %b) {
159 ; ALL-LABEL: shuffle_v4f64_1133:
161 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,3]
163 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
164 ret <4 x double> %shuffle
167 define <4 x double> @shuffle_v4f64_1023(<4 x double> %a, <4 x double> %b) {
168 ; ALL-LABEL: shuffle_v4f64_1023:
170 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,3]
172 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 3>
173 ret <4 x double> %shuffle
176 define <4 x double> @shuffle_v4f64_1022(<4 x double> %a, <4 x double> %b) {
177 ; ALL-LABEL: shuffle_v4f64_1022:
179 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,2,2]
181 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 2, i32 2>
182 ret <4 x double> %shuffle
185 define <4 x double> @shuffle_v4f64_0423(<4 x double> %a, <4 x double> %b) {
186 ; AVX1-LABEL: shuffle_v4f64_0423:
188 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
189 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
192 ; AVX2-LABEL: shuffle_v4f64_0423:
194 ; AVX2-NEXT: vbroadcastsd %xmm1, %ymm1
195 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
197 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 3>
198 ret <4 x double> %shuffle
201 define <4 x double> @shuffle_v4f64_0462(<4 x double> %a, <4 x double> %b) {
202 ; ALL-LABEL: shuffle_v4f64_0462:
204 ; ALL-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
205 ; ALL-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
206 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3]
208 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 6, i32 2>
209 ret <4 x double> %shuffle
212 define <4 x double> @shuffle_v4f64_0426(<4 x double> %a, <4 x double> %b) {
213 ; ALL-LABEL: shuffle_v4f64_0426:
215 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
217 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
218 ret <4 x double> %shuffle
221 define <4 x double> @shuffle_v4f64_1537(<4 x double> %a, <4 x double> %b) {
222 ; ALL-LABEL: shuffle_v4f64_1537:
224 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
226 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
227 ret <4 x double> %shuffle
230 define <4 x double> @shuffle_v4f64_4062(<4 x double> %a, <4 x double> %b) {
231 ; ALL-LABEL: shuffle_v4f64_4062:
233 ; ALL-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
235 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 0, i32 6, i32 2>
236 ret <4 x double> %shuffle
239 define <4 x double> @shuffle_v4f64_5173(<4 x double> %a, <4 x double> %b) {
240 ; ALL-LABEL: shuffle_v4f64_5173:
242 ; ALL-NEXT: vunpckhpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[3],ymm0[3]
244 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 7, i32 3>
245 ret <4 x double> %shuffle
248 define <4 x double> @shuffle_v4f64_5163(<4 x double> %a, <4 x double> %b) {
249 ; ALL-LABEL: shuffle_v4f64_5163:
251 ; ALL-NEXT: vshufpd {{.*#+}} ymm0 = ymm1[1],ymm0[1],ymm1[2],ymm0[3]
253 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 5, i32 1, i32 6, i32 3>
254 ret <4 x double> %shuffle
257 define <4 x double> @shuffle_v4f64_0527(<4 x double> %a, <4 x double> %b) {
258 ; ALL-LABEL: shuffle_v4f64_0527:
260 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
262 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 5, i32 2, i32 7>
263 ret <4 x double> %shuffle
266 define <4 x double> @shuffle_v4f64_4163(<4 x double> %a, <4 x double> %b) {
267 ; ALL-LABEL: shuffle_v4f64_4163:
269 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1],ymm1[2],ymm0[3]
271 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 1, i32 6, i32 3>
272 ret <4 x double> %shuffle
275 define <4 x double> @shuffle_v4f64_0145(<4 x double> %a, <4 x double> %b) {
276 ; ALL-LABEL: shuffle_v4f64_0145:
278 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
280 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
281 ret <4 x double> %shuffle
284 define <4 x double> @shuffle_v4f64_4501(<4 x double> %a, <4 x double> %b) {
285 ; ALL-LABEL: shuffle_v4f64_4501:
287 ; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
289 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
290 ret <4 x double> %shuffle
293 define <4 x double> @shuffle_v4f64_0167(<4 x double> %a, <4 x double> %b) {
294 ; ALL-LABEL: shuffle_v4f64_0167:
296 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
298 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 1, i32 6, i32 7>
299 ret <4 x double> %shuffle
302 define <4 x double> @shuffle_v4f64_1054(<4 x double> %a, <4 x double> %b) {
303 ; ALL-LABEL: shuffle_v4f64_1054:
305 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
306 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
308 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
309 ret <4 x double> %shuffle
312 define <4 x double> @shuffle_v4f64_3254(<4 x double> %a, <4 x double> %b) {
313 ; ALL-LABEL: shuffle_v4f64_3254:
315 ; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
316 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
318 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
319 ret <4 x double> %shuffle
322 define <4 x double> @shuffle_v4f64_3276(<4 x double> %a, <4 x double> %b) {
323 ; ALL-LABEL: shuffle_v4f64_3276:
325 ; ALL-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
326 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
328 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
329 ret <4 x double> %shuffle
332 define <4 x double> @shuffle_v4f64_1076(<4 x double> %a, <4 x double> %b) {
333 ; ALL-LABEL: shuffle_v4f64_1076:
335 ; ALL-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
336 ; ALL-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
338 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
339 ret <4 x double> %shuffle
342 define <4 x double> @shuffle_v4f64_0415(<4 x double> %a, <4 x double> %b) {
343 ; AVX1-LABEL: shuffle_v4f64_0415:
345 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
346 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
347 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
350 ; AVX2-LABEL: shuffle_v4f64_0415:
352 ; AVX2-NEXT: vpermpd {{.*#+}} ymm1 = ymm1[0,0,2,1]
353 ; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
354 ; AVX2-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2],ymm1[3]
356 %shuffle = shufflevector <4 x double> %a, <4 x double> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
357 ret <4 x double> %shuffle
360 define <4 x i64> @shuffle_v4i64_0000(<4 x i64> %a, <4 x i64> %b) {
361 ; AVX1-LABEL: shuffle_v4i64_0000:
363 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
364 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
367 ; AVX2-LABEL: shuffle_v4i64_0000:
369 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
371 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
372 ret <4 x i64> %shuffle
375 define <4 x i64> @shuffle_v4i64_0001(<4 x i64> %a, <4 x i64> %b) {
376 ; AVX1-LABEL: shuffle_v4i64_0001:
378 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm0[0,0]
379 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
382 ; AVX2-LABEL: shuffle_v4i64_0001:
384 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
386 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 0, i32 1>
387 ret <4 x i64> %shuffle
390 define <4 x i64> @shuffle_v4i64_0020(<4 x i64> %a, <4 x i64> %b) {
391 ; AVX1-LABEL: shuffle_v4i64_0020:
393 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
394 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm1 = xmm1[0],xmm0[0]
395 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
396 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
399 ; AVX2-LABEL: shuffle_v4i64_0020:
401 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,2,0]
403 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 0, i32 2, i32 0>
404 ret <4 x i64> %shuffle
407 define <4 x i64> @shuffle_v4i64_0112(<4 x i64> %a, <4 x i64> %b) {
408 ; AVX1-LABEL: shuffle_v4i64_0112:
410 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
411 ; AVX1-NEXT: vshufpd {{.*#+}} xmm1 = xmm0[1],xmm1[0]
412 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
415 ; AVX2-LABEL: shuffle_v4i64_0112:
417 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
419 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 1, i32 2>
420 ret <4 x i64> %shuffle
423 define <4 x i64> @shuffle_v4i64_0300(<4 x i64> %a, <4 x i64> %b) {
424 ; AVX1-LABEL: shuffle_v4i64_0300:
426 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
427 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,1,2,2]
428 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
431 ; AVX2-LABEL: shuffle_v4i64_0300:
433 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,3,0,0]
435 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 3, i32 0, i32 0>
436 ret <4 x i64> %shuffle
439 define <4 x i64> @shuffle_v4i64_1000(<4 x i64> %a, <4 x i64> %b) {
440 ; AVX1-LABEL: shuffle_v4i64_1000:
442 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
443 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
444 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
447 ; AVX2-LABEL: shuffle_v4i64_1000:
449 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,0,0,0]
451 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 0, i32 0>
452 ret <4 x i64> %shuffle
455 define <4 x i64> @shuffle_v4i64_2200(<4 x i64> %a, <4 x i64> %b) {
456 ; AVX1-LABEL: shuffle_v4i64_2200:
458 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
459 ; AVX1-NEXT: vmovddup {{.*#+}} ymm0 = ymm0[0,0,2,2]
462 ; AVX2-LABEL: shuffle_v4i64_2200:
464 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,2,0,0]
466 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 2, i32 0, i32 0>
467 ret <4 x i64> %shuffle
470 define <4 x i64> @shuffle_v4i64_3330(<4 x i64> %a, <4 x i64> %b) {
471 ; AVX1-LABEL: shuffle_v4i64_3330:
473 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1]
474 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0,1,2],ymm0[3]
475 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,1,3,2]
478 ; AVX2-LABEL: shuffle_v4i64_3330:
480 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,3,3,0]
482 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 0>
483 ret <4 x i64> %shuffle
486 define <4 x i64> @shuffle_v4i64_3210(<4 x i64> %a, <4 x i64> %b) {
487 ; AVX1-LABEL: shuffle_v4i64_3210:
489 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3,0,1]
490 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
493 ; AVX2-LABEL: shuffle_v4i64_3210:
495 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[3,2,1,0]
497 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
498 ret <4 x i64> %shuffle
501 define <4 x i64> @shuffle_v4i64_0124(<4 x i64> %a, <4 x i64> %b) {
502 ; AVX1-LABEL: shuffle_v4i64_0124:
504 ; AVX1-NEXT: vmovddup {{.*#+}} xmm1 = xmm1[0,0]
505 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
506 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],ymm1[3]
509 ; AVX2-LABEL: shuffle_v4i64_0124:
511 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
512 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7]
514 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 2, i32 4>
515 ret <4 x i64> %shuffle
518 define <4 x i64> @shuffle_v4i64_0142(<4 x i64> %a, <4 x i64> %b) {
519 ; AVX1-LABEL: shuffle_v4i64_0142:
521 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm1
522 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[0,1,2,2]
523 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
526 ; AVX2-LABEL: shuffle_v4i64_0142:
528 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
529 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,2,2]
530 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
532 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 2>
533 ret <4 x i64> %shuffle
536 define <4 x i64> @shuffle_v4i64_0412(<4 x i64> %a, <4 x i64> %b) {
537 ; AVX1-LABEL: shuffle_v4i64_0412:
539 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
540 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
541 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
542 ; AVX1-NEXT: vmovddup {{.*#+}} ymm1 = ymm1[0,0,2,2]
543 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1],ymm0[2,3]
546 ; AVX2-LABEL: shuffle_v4i64_0412:
548 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
549 ; AVX2-NEXT: vpbroadcastq %xmm1, %ymm1
550 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
552 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 2>
553 ret <4 x i64> %shuffle
556 define <4 x i64> @shuffle_v4i64_4012(<4 x i64> %a, <4 x i64> %b) {
557 ; AVX1-LABEL: shuffle_v4i64_4012:
559 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
560 ; AVX1-NEXT: vshufpd {{.*#+}} xmm2 = xmm0[1],xmm2[0]
561 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
562 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
563 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm1[0],ymm0[1,2,3]
566 ; AVX2-LABEL: shuffle_v4i64_4012:
568 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,2]
569 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5,6,7]
571 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 2>
572 ret <4 x i64> %shuffle
575 define <4 x i64> @shuffle_v4i64_0145(<4 x i64> %a, <4 x i64> %b) {
576 ; ALL-LABEL: shuffle_v4i64_0145:
578 ; ALL-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
580 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
581 ret <4 x i64> %shuffle
584 define <4 x i64> @shuffle_v4i64_0451(<4 x i64> %a, <4 x i64> %b) {
585 ; AVX1-LABEL: shuffle_v4i64_0451:
587 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm1[1],xmm0[1]
588 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
589 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
592 ; AVX2-LABEL: shuffle_v4i64_0451:
594 ; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0
595 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,1,3]
596 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5],ymm0[6,7]
598 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 5, i32 1>
599 ret <4 x i64> %shuffle
602 define <4 x i64> @shuffle_v4i64_4501(<4 x i64> %a, <4 x i64> %b) {
603 ; ALL-LABEL: shuffle_v4i64_4501:
605 ; ALL-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
607 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 5, i32 0, i32 1>
608 ret <4 x i64> %shuffle
611 define <4 x i64> @shuffle_v4i64_4015(<4 x i64> %a, <4 x i64> %b) {
612 ; AVX1-LABEL: shuffle_v4i64_4015:
614 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
615 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm1[0],xmm0[0]
616 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
619 ; AVX2-LABEL: shuffle_v4i64_4015:
621 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm1, %ymm1
622 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,3]
623 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm1[0,1],ymm0[2,3,4,5],ymm1[6,7]
625 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 4, i32 0, i32 1, i32 5>
626 ret <4 x i64> %shuffle
629 define <4 x i64> @shuffle_v4i64_2u35(<4 x i64> %a, <4 x i64> %b) {
630 ; AVX1-LABEL: shuffle_v4i64_2u35:
632 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
633 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm1 = xmm0[1],xmm1[1]
634 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
637 ; AVX2-LABEL: shuffle_v4i64_2u35:
639 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5,6,7]
640 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[2,1,3,1]
642 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 2, i32 undef, i32 3, i32 5>
643 ret <4 x i64> %shuffle
646 define <4 x i64> @shuffle_v4i64_1251(<4 x i64> %a, <4 x i64> %b) {
647 ; AVX1-LABEL: shuffle_v4i64_1251:
649 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm2 = ymm0[2,3,0,1]
650 ; AVX1-NEXT: vshufpd {{.*#+}} ymm0 = ymm0[1],ymm2[0],ymm0[2],ymm2[3]
651 ; AVX1-NEXT: vpermilpd {{.*#+}} xmm1 = xmm1[1,0]
652 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm1
653 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2],ymm0[3]
656 ; AVX2-LABEL: shuffle_v4i64_1251:
658 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,1,1,3]
659 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[1,2,2,1]
660 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7]
662 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 2, i32 5, i32 1>
663 ret <4 x i64> %shuffle
666 define <4 x i64> @shuffle_v4i64_1054(<4 x i64> %a, <4 x i64> %b) {
667 ; AVX1-LABEL: shuffle_v4i64_1054:
669 ; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
670 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
673 ; AVX2-LABEL: shuffle_v4i64_1054:
675 ; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm0, %ymm0
676 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
678 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 5, i32 4>
679 ret <4 x i64> %shuffle
682 define <4 x i64> @shuffle_v4i64_3254(<4 x i64> %a, <4 x i64> %b) {
683 ; AVX1-LABEL: shuffle_v4i64_3254:
685 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
686 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
689 ; AVX2-LABEL: shuffle_v4i64_3254:
691 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[0,1]
692 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
694 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 5, i32 4>
695 ret <4 x i64> %shuffle
698 define <4 x i64> @shuffle_v4i64_3276(<4 x i64> %a, <4 x i64> %b) {
699 ; AVX1-LABEL: shuffle_v4i64_3276:
701 ; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
702 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
705 ; AVX2-LABEL: shuffle_v4i64_3276:
707 ; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm1[2,3]
708 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
710 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 3, i32 2, i32 7, i32 6>
711 ret <4 x i64> %shuffle
714 define <4 x i64> @shuffle_v4i64_1076(<4 x i64> %a, <4 x i64> %b) {
715 ; AVX1-LABEL: shuffle_v4i64_1076:
717 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3]
718 ; AVX1-NEXT: vpermilpd {{.*#+}} ymm0 = ymm0[1,0,3,2]
721 ; AVX2-LABEL: shuffle_v4i64_1076:
723 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
724 ; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5]
726 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 1, i32 0, i32 7, i32 6>
727 ret <4 x i64> %shuffle
730 define <4 x i64> @shuffle_v4i64_0415(<4 x i64> %a, <4 x i64> %b) {
731 ; AVX1-LABEL: shuffle_v4i64_0415:
733 ; AVX1-NEXT: vunpckhpd {{.*#+}} xmm2 = xmm0[1],xmm1[1]
734 ; AVX1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0]
735 ; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0
738 ; AVX2-LABEL: shuffle_v4i64_0415:
740 ; AVX2-NEXT: vpermq {{.*#+}} ymm1 = ymm1[0,0,2,1]
741 ; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,1,3]
742 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
744 %shuffle = shufflevector <4 x i64> %a, <4 x i64> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
745 ret <4 x i64> %shuffle
748 define <4 x i64> @shuffle_v4i64_z4z6(<4 x i64> %a) {
749 ; AVX1-LABEL: shuffle_v4i64_z4z6:
751 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
752 ; AVX1-NEXT: vunpcklpd {{.*#+}} ymm0 = ymm1[0],ymm0[0],ymm1[2],ymm0[2]
755 ; AVX2-LABEL: shuffle_v4i64_z4z6:
757 ; 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]
759 %shuffle = shufflevector <4 x i64> zeroinitializer, <4 x i64> %a, <4 x i32> <i32 0, i32 4, i32 0, i32 6>
760 ret <4 x i64> %shuffle
763 define <4 x i64> @shuffle_v4i64_5zuz(<4 x i64> %a) {
764 ; AVX1-LABEL: shuffle_v4i64_5zuz:
766 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
767 ; AVX1-NEXT: vshufpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[2],ymm1[3]
770 ; AVX2-LABEL: shuffle_v4i64_5zuz:
772 ; 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
774 %shuffle = shufflevector <4 x i64> zeroinitializer, <4 x i64> %a, <4 x i32> <i32 5, i32 0, i32 undef, i32 0>
775 ret <4 x i64> %shuffle
778 define <4 x i64> @stress_test1(<4 x i64> %a, <4 x i64> %b) {
779 ; ALL-LABEL: stress_test1:
781 %c = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 1, i32 1, i32 0>
782 %d = shufflevector <4 x i64> %c, <4 x i64> undef, <4 x i32> <i32 3, i32 undef, i32 2, i32 undef>
783 %e = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> <i32 3, i32 3, i32 1, i32 undef>
784 %f = shufflevector <4 x i64> %d, <4 x i64> %e, <4 x i32> <i32 5, i32 1, i32 1, i32 0>
789 define <4 x i64> @insert_reg_and_zero_v4i64(i64 %a) {
790 ; AVX1-LABEL: insert_reg_and_zero_v4i64:
792 ; AVX1-NEXT: vmovq %rdi, %xmm0
793 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
794 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
797 ; AVX2-LABEL: insert_reg_and_zero_v4i64:
799 ; AVX2-NEXT: vmovq %rdi, %xmm0
800 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
801 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
803 %v = insertelement <4 x i64> undef, i64 %a, i64 0
804 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
805 ret <4 x i64> %shuffle
808 define <4 x i64> @insert_mem_and_zero_v4i64(i64* %ptr) {
809 ; AVX1-LABEL: insert_mem_and_zero_v4i64:
811 ; AVX1-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
812 ; AVX1-NEXT: vxorpd %ymm1, %ymm1, %ymm1
813 ; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3]
816 ; AVX2-LABEL: insert_mem_and_zero_v4i64:
818 ; AVX2-NEXT: vmovq {{.*#+}} xmm0 = mem[0],zero
819 ; AVX2-NEXT: vpxor %ymm1, %ymm1, %ymm1
820 ; AVX2-NEXT: vpblendd {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3,4,5,6,7]
823 %v = insertelement <4 x i64> undef, i64 %a, i64 0
824 %shuffle = shufflevector <4 x i64> %v, <4 x i64> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
825 ret <4 x i64> %shuffle
828 define <4 x double> @insert_reg_and_zero_v4f64(double %a) {
829 ; ALL-LABEL: insert_reg_and_zero_v4f64:
831 ; ALL-NEXT: vxorpd %xmm1, %xmm1, %xmm1
832 ; ALL-NEXT: vmovsd {{.*#+}} xmm0 = xmm0[0],xmm1[1]
834 %v = insertelement <4 x double> undef, double %a, i32 0
835 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
836 ret <4 x double> %shuffle
839 define <4 x double> @insert_mem_and_zero_v4f64(double* %ptr) {
840 ; ALL-LABEL: insert_mem_and_zero_v4f64:
842 ; ALL-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
844 %a = load double* %ptr
845 %v = insertelement <4 x double> undef, double %a, i32 0
846 %shuffle = shufflevector <4 x double> %v, <4 x double> zeroinitializer, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
847 ret <4 x double> %shuffle
850 define <4 x double> @splat_mem_v4f64(double* %ptr) {
851 ; ALL-LABEL: splat_mem_v4f64:
853 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
855 %a = load double* %ptr
856 %v = insertelement <4 x double> undef, double %a, i32 0
857 %shuffle = shufflevector <4 x double> %v, <4 x double> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
858 ret <4 x double> %shuffle
861 define <4 x i64> @splat_mem_v4i64(i64* %ptr) {
862 ; AVX1-LABEL: splat_mem_v4i64:
864 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = mem[0,0]
865 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
868 ; AVX2-LABEL: splat_mem_v4i64:
870 ; AVX2-NEXT: vbroadcastsd (%rdi), %ymm0
873 %v = insertelement <4 x i64> undef, i64 %a, i64 0
874 %shuffle = shufflevector <4 x i64> %v, <4 x i64> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
875 ret <4 x i64> %shuffle
878 define <4 x double> @splat_mem_v4f64_2(double* %p) {
879 ; ALL-LABEL: splat_mem_v4f64_2:
881 ; ALL-NEXT: vbroadcastsd (%rdi), %ymm0
884 %2 = insertelement <2 x double> undef, double %1, i32 0
885 %3 = shufflevector <2 x double> %2, <2 x double> undef, <4 x i32> zeroinitializer
889 define <4 x double> @splat_v4f64(<2 x double> %r) {
890 ; AVX1-LABEL: splat_v4f64:
892 ; AVX1-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0]
893 ; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
896 ; AVX2-LABEL: splat_v4f64:
898 ; AVX2-NEXT: vbroadcastsd %xmm0, %ymm0
900 %1 = shufflevector <2 x double> %r, <2 x double> undef, <4 x i32> zeroinitializer