1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -mattr=+sve -force-streaming-compatible-sve < %s | FileCheck %s
4 target triple = "aarch64-unknown-linux-gnu"
10 define <8 x i8> @concat_v8i8(<4 x i8> %op1, <4 x i8> %op2) {
11 ; CHECK-LABEL: concat_v8i8:
13 ; CHECK-NEXT: sub sp, sp, #16
14 ; CHECK-NEXT: .cfi_def_cfa_offset 16
15 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
16 ; CHECK-NEXT: fmov w8, s1
17 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
18 ; CHECK-NEXT: mov z2.h, z1.h[3]
19 ; CHECK-NEXT: mov z3.h, z1.h[2]
20 ; CHECK-NEXT: mov z1.h, z1.h[1]
21 ; CHECK-NEXT: mov z4.h, z0.h[3]
22 ; CHECK-NEXT: strb w8, [sp, #12]
23 ; CHECK-NEXT: fmov w8, s0
24 ; CHECK-NEXT: strb w8, [sp, #8]
25 ; CHECK-NEXT: fmov w8, s2
26 ; CHECK-NEXT: mov z2.h, z0.h[2]
27 ; CHECK-NEXT: mov z0.h, z0.h[1]
28 ; CHECK-NEXT: strb w8, [sp, #15]
29 ; CHECK-NEXT: fmov w8, s3
30 ; CHECK-NEXT: strb w8, [sp, #14]
31 ; CHECK-NEXT: fmov w8, s1
32 ; CHECK-NEXT: strb w8, [sp, #13]
33 ; CHECK-NEXT: fmov w8, s4
34 ; CHECK-NEXT: strb w8, [sp, #11]
35 ; CHECK-NEXT: fmov w8, s2
36 ; CHECK-NEXT: strb w8, [sp, #10]
37 ; CHECK-NEXT: fmov w8, s0
38 ; CHECK-NEXT: strb w8, [sp, #9]
39 ; CHECK-NEXT: ldr d0, [sp, #8]
40 ; CHECK-NEXT: add sp, sp, #16
42 %res = shufflevector <4 x i8> %op1, <4 x i8> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
46 define <16 x i8> @concat_v16i8(<8 x i8> %op1, <8 x i8> %op2) {
47 ; CHECK-LABEL: concat_v16i8:
49 ; CHECK-NEXT: ptrue p0.b, vl8
50 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
51 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
52 ; CHECK-NEXT: splice z0.b, p0, z0.b, z1.b
53 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
55 %res = shufflevector <8 x i8> %op1, <8 x i8> %op2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
56 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
60 define void @concat_v32i8(ptr %a, ptr %b, ptr %c) {
61 ; CHECK-LABEL: concat_v32i8:
63 ; CHECK-NEXT: ldr q0, [x1]
64 ; CHECK-NEXT: ldr q1, [x0]
65 ; CHECK-NEXT: stp q1, q0, [x2]
67 %op1 = load <16 x i8>, ptr %a
68 %op2 = load <16 x i8>, ptr %b
69 %res = shufflevector <16 x i8> %op1, <16 x i8> %op2, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
70 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
71 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
72 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
73 store <32 x i8> %res, ptr %c
77 define void @concat_v64i8(ptr %a, ptr %b, ptr %c) {
78 ; CHECK-LABEL: concat_v64i8:
80 ; CHECK-NEXT: ldp q0, q1, [x1]
81 ; CHECK-NEXT: ldp q3, q2, [x0]
82 ; CHECK-NEXT: stp q0, q1, [x2, #32]
83 ; CHECK-NEXT: stp q3, q2, [x2]
85 %op1 = load <32 x i8>, ptr %a
86 %op2 = load <32 x i8>, ptr %b
87 %res = shufflevector <32 x i8> %op1, <32 x i8> %op2, <64 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
88 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
89 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
90 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31,
91 i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39,
92 i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47,
93 i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55,
94 i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
95 store <64 x i8> %res, ptr %c
103 define <4 x i16> @concat_v4i16(<2 x i16> %op1, <2 x i16> %op2) {
104 ; CHECK-LABEL: concat_v4i16:
106 ; CHECK-NEXT: sub sp, sp, #16
107 ; CHECK-NEXT: .cfi_def_cfa_offset 16
108 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
109 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
110 ; CHECK-NEXT: mov z2.s, z1.s[1]
111 ; CHECK-NEXT: fmov w8, s1
112 ; CHECK-NEXT: fmov w9, s0
113 ; CHECK-NEXT: mov z1.s, z0.s[1]
114 ; CHECK-NEXT: strh w8, [sp, #12]
115 ; CHECK-NEXT: fmov w8, s2
116 ; CHECK-NEXT: strh w9, [sp, #8]
117 ; CHECK-NEXT: fmov w9, s1
118 ; CHECK-NEXT: strh w8, [sp, #14]
119 ; CHECK-NEXT: strh w9, [sp, #10]
120 ; CHECK-NEXT: ldr d0, [sp, #8]
121 ; CHECK-NEXT: add sp, sp, #16
123 %res = shufflevector <2 x i16> %op1, <2 x i16> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
127 ; Don't use SVE for 128-bit vectors.
128 define <8 x i16> @concat_v8i16(<4 x i16> %op1, <4 x i16> %op2) {
129 ; CHECK-LABEL: concat_v8i16:
131 ; CHECK-NEXT: ptrue p0.h, vl4
132 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
133 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
134 ; CHECK-NEXT: splice z0.h, p0, z0.h, z1.h
135 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
137 %res = shufflevector <4 x i16> %op1, <4 x i16> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
141 define void @concat_v16i16(ptr %a, ptr %b, ptr %c) {
142 ; CHECK-LABEL: concat_v16i16:
144 ; CHECK-NEXT: ldr q0, [x1]
145 ; CHECK-NEXT: ldr q1, [x0]
146 ; CHECK-NEXT: stp q1, q0, [x2]
148 %op1 = load <8 x i16>, ptr %a
149 %op2 = load <8 x i16>, ptr %b
150 %res = shufflevector <8 x i16> %op1, <8 x i16> %op2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
151 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
152 store <16 x i16> %res, ptr %c
156 define void @concat_v32i16(ptr %a, ptr %b, ptr %c) {
157 ; CHECK-LABEL: concat_v32i16:
159 ; CHECK-NEXT: ldp q0, q1, [x1]
160 ; CHECK-NEXT: ldp q3, q2, [x0]
161 ; CHECK-NEXT: stp q0, q1, [x2, #32]
162 ; CHECK-NEXT: stp q3, q2, [x2]
164 %op1 = load <16 x i16>, ptr %a
165 %op2 = load <16 x i16>, ptr %b
166 %res = shufflevector <16 x i16> %op1, <16 x i16> %op2, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
167 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
168 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
169 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
170 store <32 x i16> %res, ptr %c
178 ; Don't use SVE for 64-bit vectors.
179 define <2 x i32> @concat_v2i32(<1 x i32> %op1, <1 x i32> %op2) {
180 ; CHECK-LABEL: concat_v2i32:
182 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
183 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
184 ; CHECK-NEXT: zip1 z0.s, z0.s, z1.s
185 ; CHECK-NEXT: // kill: def $d0 killed $d0 killed $z0
187 %res = shufflevector <1 x i32> %op1, <1 x i32> %op2, <2 x i32> <i32 0, i32 1>
191 ; Don't use SVE for 128-bit vectors.
192 define <4 x i32> @concat_v4i32(<2 x i32> %op1, <2 x i32> %op2) {
193 ; CHECK-LABEL: concat_v4i32:
195 ; CHECK-NEXT: ptrue p0.s, vl2
196 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
197 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
198 ; CHECK-NEXT: splice z0.s, p0, z0.s, z1.s
199 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
201 %res = shufflevector <2 x i32> %op1, <2 x i32> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
205 define void @concat_v8i32(ptr %a, ptr %b, ptr %c) {
206 ; CHECK-LABEL: concat_v8i32:
208 ; CHECK-NEXT: ldr q0, [x1]
209 ; CHECK-NEXT: ldr q1, [x0]
210 ; CHECK-NEXT: stp q1, q0, [x2]
212 %op1 = load <4 x i32>, ptr %a
213 %op2 = load <4 x i32>, ptr %b
214 %res = shufflevector <4 x i32> %op1, <4 x i32> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
215 store <8 x i32> %res, ptr %c
219 define void @concat_v16i32(ptr %a, ptr %b, ptr %c) {
220 ; CHECK-LABEL: concat_v16i32:
222 ; CHECK-NEXT: ldp q0, q1, [x1]
223 ; CHECK-NEXT: ldp q3, q2, [x0]
224 ; CHECK-NEXT: stp q0, q1, [x2, #32]
225 ; CHECK-NEXT: stp q3, q2, [x2]
227 %op1 = load <8 x i32>, ptr %a
228 %op2 = load <8 x i32>, ptr %b
229 %res = shufflevector <8 x i32> %op1, <8 x i32> %op2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
230 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
231 store <16 x i32> %res, ptr %c
239 ; Don't use SVE for 128-bit vectors.
240 define <2 x i64> @concat_v2i64(<1 x i64> %op1, <1 x i64> %op2) {
241 ; CHECK-LABEL: concat_v2i64:
243 ; CHECK-NEXT: ptrue p0.d, vl1
244 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
245 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
246 ; CHECK-NEXT: splice z0.d, p0, z0.d, z1.d
247 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
249 %res = shufflevector <1 x i64> %op1, <1 x i64> %op2, <2 x i32> <i32 0, i32 1>
253 define void @concat_v4i64(ptr %a, ptr %b, ptr %c) {
254 ; CHECK-LABEL: concat_v4i64:
256 ; CHECK-NEXT: ldr q0, [x1]
257 ; CHECK-NEXT: ldr q1, [x0]
258 ; CHECK-NEXT: stp q1, q0, [x2]
260 %op1 = load <2 x i64>, ptr %a
261 %op2 = load <2 x i64>, ptr %b
262 %res = shufflevector <2 x i64> %op1, <2 x i64> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
263 store <4 x i64> %res, ptr %c
267 define void @concat_v8i64(ptr %a, ptr %b, ptr %c) {
268 ; CHECK-LABEL: concat_v8i64:
270 ; CHECK-NEXT: ldp q0, q1, [x1]
271 ; CHECK-NEXT: ldp q3, q2, [x0]
272 ; CHECK-NEXT: stp q0, q1, [x2, #32]
273 ; CHECK-NEXT: stp q3, q2, [x2]
275 %op1 = load <4 x i64>, ptr %a
276 %op2 = load <4 x i64>, ptr %b
277 %res = shufflevector <4 x i64> %op1, <4 x i64> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
278 store <8 x i64> %res, ptr %c
286 define <4 x half> @concat_v4f16(<2 x half> %op1, <2 x half> %op2) {
287 ; CHECK-LABEL: concat_v4f16:
289 ; CHECK-NEXT: sub sp, sp, #16
290 ; CHECK-NEXT: .cfi_def_cfa_offset 16
291 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
292 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
293 ; CHECK-NEXT: mov z2.h, z1.h[1]
294 ; CHECK-NEXT: str h1, [sp, #12]
295 ; CHECK-NEXT: mov z1.h, z0.h[1]
296 ; CHECK-NEXT: str h0, [sp, #8]
297 ; CHECK-NEXT: str h2, [sp, #14]
298 ; CHECK-NEXT: str h1, [sp, #10]
299 ; CHECK-NEXT: ldr d0, [sp, #8]
300 ; CHECK-NEXT: add sp, sp, #16
302 %res = shufflevector <2 x half> %op1, <2 x half> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
306 define <8 x half> @concat_v8f16(<4 x half> %op1, <4 x half> %op2) {
307 ; CHECK-LABEL: concat_v8f16:
309 ; CHECK-NEXT: ptrue p0.h, vl4
310 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
311 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
312 ; CHECK-NEXT: splice z0.h, p0, z0.h, z1.h
313 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
315 %res = shufflevector <4 x half> %op1, <4 x half> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
319 define void @concat_v16f16(ptr %a, ptr %b, ptr %c) {
320 ; CHECK-LABEL: concat_v16f16:
322 ; CHECK-NEXT: ldr q0, [x1]
323 ; CHECK-NEXT: ldr q1, [x0]
324 ; CHECK-NEXT: stp q1, q0, [x2]
326 %op1 = load <8 x half>, ptr %a
327 %op2 = load <8 x half>, ptr %b
328 %res = shufflevector <8 x half> %op1, <8 x half> %op2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
329 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
330 store <16 x half> %res, ptr %c
334 define void @concat_v32f16(ptr %a, ptr %b, ptr %c) {
335 ; CHECK-LABEL: concat_v32f16:
337 ; CHECK-NEXT: ldp q0, q1, [x1]
338 ; CHECK-NEXT: ldp q3, q2, [x0]
339 ; CHECK-NEXT: stp q0, q1, [x2, #32]
340 ; CHECK-NEXT: stp q3, q2, [x2]
342 %op1 = load <16 x half>, ptr %a
343 %op2 = load <16 x half>, ptr %b
344 %res = shufflevector <16 x half> %op1, <16 x half> %op2, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
345 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
346 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
347 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
348 store <32 x half> %res, ptr %c
356 ; Don't use SVE for 64-bit vectors.
357 define <2 x float> @concat_v2f32(<1 x float> %op1, <1 x float> %op2) {
358 ; CHECK-LABEL: concat_v2f32:
360 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
361 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
362 ; CHECK-NEXT: zip1 z0.s, z0.s, z1.s
363 ; CHECK-NEXT: // kill: def $d0 killed $d0 killed $z0
365 %res = shufflevector <1 x float> %op1, <1 x float> %op2, <2 x i32> <i32 0, i32 1>
369 ; Don't use SVE for 128-bit vectors.
370 define <4 x float> @concat_v4f32(<2 x float> %op1, <2 x float> %op2) {
371 ; CHECK-LABEL: concat_v4f32:
373 ; CHECK-NEXT: ptrue p0.s, vl2
374 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
375 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
376 ; CHECK-NEXT: splice z0.s, p0, z0.s, z1.s
377 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
379 %res = shufflevector <2 x float> %op1, <2 x float> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
383 define void @concat_v8f32(ptr %a, ptr %b, ptr %c) {
384 ; CHECK-LABEL: concat_v8f32:
386 ; CHECK-NEXT: ldr q0, [x1]
387 ; CHECK-NEXT: ldr q1, [x0]
388 ; CHECK-NEXT: stp q1, q0, [x2]
390 %op1 = load <4 x float>, ptr %a
391 %op2 = load <4 x float>, ptr %b
392 %res = shufflevector <4 x float> %op1, <4 x float> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
393 store <8 x float> %res, ptr %c
397 define void @concat_v16f32(ptr %a, ptr %b, ptr %c) {
398 ; CHECK-LABEL: concat_v16f32:
400 ; CHECK-NEXT: ldp q0, q1, [x1]
401 ; CHECK-NEXT: ldp q3, q2, [x0]
402 ; CHECK-NEXT: stp q0, q1, [x2, #32]
403 ; CHECK-NEXT: stp q3, q2, [x2]
405 %op1 = load <8 x float>, ptr %a
406 %op2 = load <8 x float>, ptr %b
407 %res = shufflevector <8 x float> %op1, <8 x float> %op2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
408 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
409 store <16 x float> %res, ptr %c
417 ; Don't use SVE for 128-bit vectors.
418 define <2 x double> @concat_v2f64(<1 x double> %op1, <1 x double> %op2) {
419 ; CHECK-LABEL: concat_v2f64:
421 ; CHECK-NEXT: ptrue p0.d, vl1
422 ; CHECK-NEXT: // kill: def $d0 killed $d0 def $z0
423 ; CHECK-NEXT: // kill: def $d1 killed $d1 def $z1
424 ; CHECK-NEXT: splice z0.d, p0, z0.d, z1.d
425 ; CHECK-NEXT: // kill: def $q0 killed $q0 killed $z0
427 %res = shufflevector <1 x double> %op1, <1 x double> %op2, <2 x i32> <i32 0, i32 1>
428 ret <2 x double> %res
431 define void @concat_v4f64(ptr %a, ptr %b, ptr %c) {
432 ; CHECK-LABEL: concat_v4f64:
434 ; CHECK-NEXT: ldr q0, [x1]
435 ; CHECK-NEXT: ldr q1, [x0]
436 ; CHECK-NEXT: stp q1, q0, [x2]
438 %op1 = load <2 x double>, ptr %a
439 %op2 = load <2 x double>, ptr %b
440 %res = shufflevector <2 x double> %op1, <2 x double> %op2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
441 store <4 x double> %res, ptr %c
445 define void @concat_v8f64(ptr %a, ptr %b, ptr %c) {
446 ; CHECK-LABEL: concat_v8f64:
448 ; CHECK-NEXT: ldp q0, q1, [x1]
449 ; CHECK-NEXT: ldp q3, q2, [x0]
450 ; CHECK-NEXT: stp q0, q1, [x2, #32]
451 ; CHECK-NEXT: stp q3, q2, [x2]
453 %op1 = load <4 x double>, ptr %a
454 %op2 = load <4 x double>, ptr %b
455 %res = shufflevector <4 x double> %op1, <4 x double> %op2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
456 store <8 x double> %res, ptr %c
464 define void @concat_v32i8_undef(ptr %a, ptr %b) {
465 ; CHECK-LABEL: concat_v32i8_undef:
467 ; CHECK-NEXT: ldr q0, [x0]
468 ; CHECK-NEXT: str q0, [x1]
470 %op1 = load <16 x i8>, ptr %a
471 %res = shufflevector <16 x i8> %op1, <16 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
472 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
473 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
474 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
475 store <32 x i8> %res, ptr %b
479 define void @concat_v16i16_undef(ptr %a, ptr %b) {
480 ; CHECK-LABEL: concat_v16i16_undef:
482 ; CHECK-NEXT: ldr q0, [x0]
483 ; CHECK-NEXT: str q0, [x1]
485 %op1 = load <8 x i16>, ptr %a
486 %res = shufflevector <8 x i16> %op1, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
487 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
488 store <16 x i16> %res, ptr %b
492 define void @concat_v8i32_undef(ptr %a, ptr %b) {
493 ; CHECK-LABEL: concat_v8i32_undef:
495 ; CHECK-NEXT: ldr q0, [x0]
496 ; CHECK-NEXT: str q0, [x1]
498 %op1 = load <4 x i32>, ptr %a
499 %res = shufflevector <4 x i32> %op1, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
500 store <8 x i32> %res, ptr %b
504 define void @concat_v4i64_undef(ptr %a, ptr %b) {
505 ; CHECK-LABEL: concat_v4i64_undef:
507 ; CHECK-NEXT: ldr q0, [x0]
508 ; CHECK-NEXT: str q0, [x1]
510 %op1 = load <2 x i64>, ptr %a
511 %res = shufflevector <2 x i64> %op1, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
512 store <4 x i64> %res, ptr %b
520 define void @concat_v32i8_4op(ptr %a, ptr %b) {
521 ; CHECK-LABEL: concat_v32i8_4op:
523 ; CHECK-NEXT: ldr d0, [x0]
524 ; CHECK-NEXT: str q0, [x1]
526 %op1 = load <8 x i8>, ptr %a
527 %shuffle = shufflevector <8 x i8> %op1, <8 x i8> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
528 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
529 %res = shufflevector <16 x i8> %shuffle, <16 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
530 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15,
531 i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23,
532 i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
533 store <32 x i8> %res, ptr %b
537 define void @concat_v16i16_4op(ptr %a, ptr %b) {
538 ; CHECK-LABEL: concat_v16i16_4op:
540 ; CHECK-NEXT: ldr d0, [x0]
541 ; CHECK-NEXT: str q0, [x1]
543 %op1 = load <4 x i16>, ptr %a
544 %shuffle = shufflevector <4 x i16> %op1, <4 x i16> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
545 %res = shufflevector <8 x i16> %shuffle, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7,
546 i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
547 store <16 x i16> %res, ptr %b
551 define void @concat_v8i32_4op(ptr %a, ptr %b) {
552 ; CHECK-LABEL: concat_v8i32_4op:
554 ; CHECK-NEXT: ldr d0, [x0]
555 ; CHECK-NEXT: str q0, [x1]
557 %op1 = load <2 x i32>, ptr %a
558 %shuffle = shufflevector <2 x i32> %op1, <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
559 %res = shufflevector <4 x i32> %shuffle, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
560 store <8 x i32> %res, ptr %b
564 define void @concat_v4i64_4op(ptr %a, ptr %b) {
565 ; CHECK-LABEL: concat_v4i64_4op:
567 ; CHECK-NEXT: ldr d0, [x0]
568 ; CHECK-NEXT: str q0, [x1]
570 %op1 = load <1 x i64>, ptr %a
571 %shuffle = shufflevector <1 x i64> %op1, <1 x i64> undef, <2 x i32> <i32 0, i32 1>
572 %res = shufflevector <2 x i64> %shuffle, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
573 store <4 x i64> %res, ptr %b