about summary refs log tree commit diff
path: root/sysdeps/x86_64/multiarch/memcmp-sse4.S
blob: cd57c1e2c7361c4c41fe412d5033d2f6f29ad5a1 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
/* memcmp with SSE4.1, wmemcmp with SSE4.1
   Copyright (C) 2010-2022 Free Software Foundation, Inc.
   This file is part of the GNU C Library.

   The GNU C Library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
   License as published by the Free Software Foundation; either
   version 2.1 of the License, or (at your option) any later version.

   The GNU C Library is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with the GNU C Library; if not, see
   <https://www.gnu.org/licenses/>.  */

#if IS_IN (libc)

# include <sysdep.h>

# ifndef MEMCMP
#  define MEMCMP	__memcmp_sse4_1
# endif

#ifdef USE_AS_WMEMCMP
# define CMPEQ	pcmpeqd
# define CHAR_SIZE	4
#else
# define CMPEQ	pcmpeqb
# define CHAR_SIZE	1
#endif


/* Warning!
           wmemcmp has to use SIGNED comparison for elements.
           memcmp has to use UNSIGNED comparison for elemnts.
*/

	.section .text.sse4.1,"ax",@progbits
ENTRY (MEMCMP)
# ifdef USE_AS_WMEMCMP
	shl	$2, %RDX_LP
# elif defined __ILP32__
	/* Clear the upper 32 bits.  */
	mov	%edx, %edx
# endif
	cmp	$79, %RDX_LP
	ja	L(79bytesormore)

	cmp	$CHAR_SIZE, %RDX_LP
	jbe	L(firstbyte)

	/* N in (CHAR_SIZE, 79) bytes.  */
	cmpl	$32, %edx
	ja	L(more_32_bytes)

	cmpl	$16, %edx
	jae	L(16_to_32_bytes)

# ifndef USE_AS_WMEMCMP
	cmpl	$8, %edx
	jae	L(8_to_16_bytes)

	cmpl	$4, %edx
	jb	L(2_to_3_bytes)

	movl	(%rdi), %eax
	movl	(%rsi), %ecx

	bswap	%eax
	bswap	%ecx

	shlq	$32, %rax
	shlq	$32, %rcx

	movl	-4(%rdi, %rdx), %edi
	movl	-4(%rsi, %rdx), %esi

	bswap	%edi
	bswap	%esi

	orq	%rdi, %rax
	orq	%rsi, %rcx
	subq	%rcx, %rax
	cmovne	%edx, %eax
	sbbl	%ecx, %ecx
	orl	%ecx, %eax
	ret

	.p2align 4,, 8
L(2_to_3_bytes):
	movzwl	(%rdi), %eax
	movzwl	(%rsi), %ecx
	shll	$8, %eax
	shll	$8, %ecx
	bswap	%eax
	bswap	%ecx
	movzbl	-1(%rdi, %rdx), %edi
	movzbl	-1(%rsi, %rdx), %esi
	orl	%edi, %eax
	orl	%esi, %ecx
	subl	%ecx, %eax
	ret

	.p2align 4,, 8
L(8_to_16_bytes):
	movq	(%rdi), %rax
	movq	(%rsi), %rcx

	bswap	%rax
	bswap	%rcx

	subq	%rcx, %rax
	jne	L(8_to_16_bytes_done)

	movq	-8(%rdi, %rdx), %rax
	movq	-8(%rsi, %rdx), %rcx

	bswap	%rax
	bswap	%rcx

	subq	%rcx, %rax

L(8_to_16_bytes_done):
	cmovne	%edx, %eax
	sbbl	%ecx, %ecx
	orl	%ecx, %eax
	ret
# else
	xorl	%eax, %eax
	movl	(%rdi), %ecx
	cmpl	(%rsi), %ecx
	jne	L(8_to_16_bytes_done)
	movl	4(%rdi), %ecx
	cmpl	4(%rsi), %ecx
	jne	L(8_to_16_bytes_done)
	movl	-4(%rdi, %rdx), %ecx
	cmpl	-4(%rsi, %rdx), %ecx
	jne	L(8_to_16_bytes_done)
	ret
# endif

	.p2align 4,, 3
L(ret_zero):
	xorl	%eax, %eax
L(zero):
	ret

	.p2align 4,, 8
L(firstbyte):
	jb	L(ret_zero)
# ifdef USE_AS_WMEMCMP
	xorl	%eax, %eax
	movl	(%rdi), %ecx
	cmpl	(%rsi), %ecx
	je	L(zero)
L(8_to_16_bytes_done):
	setg	%al
	leal	-1(%rax, %rax), %eax
# else
	movzbl	(%rdi), %eax
	movzbl	(%rsi), %ecx
	sub	%ecx, %eax
# endif
	ret

	.p2align 4
L(vec_return_begin_48):
	addq	$16, %rdi
	addq	$16, %rsi
L(vec_return_begin_32):
	bsfl	%eax, %eax
# ifdef USE_AS_WMEMCMP
	movl	32(%rdi, %rax), %ecx
	xorl	%edx, %edx
	cmpl	32(%rsi, %rax), %ecx
	setg	%dl
	leal	-1(%rdx, %rdx), %eax
# else
	movzbl	32(%rsi, %rax), %ecx
	movzbl	32(%rdi, %rax), %eax
	subl	%ecx, %eax
# endif
	ret

	.p2align 4
L(vec_return_begin_16):
	addq	$16, %rdi
	addq	$16, %rsi
L(vec_return_begin):
	bsfl	%eax, %eax
# ifdef USE_AS_WMEMCMP
	movl	(%rdi, %rax), %ecx
	xorl	%edx, %edx
	cmpl	(%rsi, %rax), %ecx
	setg	%dl
	leal	-1(%rdx, %rdx), %eax
# else
	movzbl	(%rsi, %rax), %ecx
	movzbl	(%rdi, %rax), %eax
	subl	%ecx, %eax
# endif
	ret

	.p2align 4
L(vec_return_end_16):
	subl	$16, %edx
L(vec_return_end):
	bsfl	%eax, %eax
	addl	%edx, %eax
# ifdef USE_AS_WMEMCMP
	movl	-16(%rdi, %rax), %ecx
	xorl	%edx, %edx
	cmpl	-16(%rsi, %rax), %ecx
	setg	%dl
	leal	-1(%rdx, %rdx), %eax
# else
	movzbl	-16(%rsi, %rax), %ecx
	movzbl	-16(%rdi, %rax), %eax
	subl	%ecx, %eax
# endif
	ret

	.p2align 4,, 8
L(more_32_bytes):
	movdqu	(%rdi), %xmm0
	movdqu	(%rsi), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	16(%rdi), %xmm0
	movdqu	16(%rsi), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	cmpl	$64, %edx
	jbe	L(32_to_64_bytes)
	movdqu	32(%rdi), %xmm0
	movdqu	32(%rsi), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	.p2align 4,, 6
L(32_to_64_bytes):
	movdqu	-32(%rdi, %rdx), %xmm0
	movdqu	-32(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end_16)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret

	.p2align 4
L(16_to_32_bytes):
	movdqu	(%rdi), %xmm0
	movdqu	(%rsi), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret


	.p2align 4
L(79bytesormore):
	movdqu	(%rdi), %xmm0
	movdqu	(%rsi), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)


	mov	%rsi, %rcx
	and	$-16, %rsi
	add	$16, %rsi
	sub	%rsi, %rcx

	sub	%rcx, %rdi
	add	%rcx, %rdx
	test	$0xf, %rdi
	jz	L(2aligned)

	cmp	$128, %rdx
	ja	L(128bytesormore)

	.p2align 4,, 6
L(less128bytes):
	movdqu	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqu	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqu	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	cmp	$96, %rdx
	jb	L(32_to_64_bytes)

	addq	$64, %rdi
	addq	$64, %rsi
	subq	$64, %rdx

	.p2align 4,, 6
L(last_64_bytes):
	movdqu	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqu	-32(%rdi, %rdx), %xmm0
	movdqu	-32(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end_16)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret

	.p2align 4
L(128bytesormore):
	cmp	$256, %rdx
	ja	L(unaligned_loop)
L(less256bytes):
	movdqu	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqu	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqu	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	addq	$64, %rdi
	addq	$64, %rsi

	movdqu	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqu	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqu	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqu	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	addq	$-128, %rdx
	subq	$-64, %rsi
	subq	$-64, %rdi

	cmp	$64, %rdx
	ja	L(less128bytes)

	cmp	$32, %rdx
	ja	L(last_64_bytes)

	movdqu	-32(%rdi, %rdx), %xmm0
	movdqu	-32(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end_16)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret

	.p2align 4
L(unaligned_loop):
# ifdef DATA_CACHE_SIZE_HALF
	mov	$DATA_CACHE_SIZE_HALF, %R8_LP
# else
	mov	__x86_data_cache_size_half(%rip), %R8_LP
# endif
	movq	%r8, %r9
	addq	%r8, %r8
	addq	%r9, %r8
	cmpq	%r8, %rdx
	ja	L(L2_L3_cache_unaligned)
	sub	$64, %rdx
	.p2align 4
L(64bytesormore_loop):
	movdqu	(%rdi), %xmm0
	movdqu	16(%rdi), %xmm1
	movdqu	32(%rdi), %xmm2
	movdqu	48(%rdi), %xmm3

	CMPEQ	(%rsi), %xmm0
	CMPEQ	16(%rsi), %xmm1
	CMPEQ	32(%rsi), %xmm2
	CMPEQ	48(%rsi), %xmm3

	pand	%xmm0, %xmm1
	pand	%xmm2, %xmm3
	pand	%xmm1, %xmm3

	pmovmskb %xmm3, %eax
	incw	%ax
	jnz	L(64bytesormore_loop_end)

	add	$64, %rsi
	add	$64, %rdi
	sub	$64, %rdx
	ja	L(64bytesormore_loop)

	.p2align 4,, 6
L(loop_tail):
	addq	%rdx, %rdi
	movdqu	(%rdi), %xmm0
	movdqu	16(%rdi), %xmm1
	movdqu	32(%rdi), %xmm2
	movdqu	48(%rdi), %xmm3

	addq	%rdx, %rsi
	movdqu	(%rsi), %xmm4
	movdqu	16(%rsi), %xmm5
	movdqu	32(%rsi), %xmm6
	movdqu	48(%rsi), %xmm7

	CMPEQ	%xmm4, %xmm0
	CMPEQ	%xmm5, %xmm1
	CMPEQ	%xmm6, %xmm2
	CMPEQ	%xmm7, %xmm3

	pand	%xmm0, %xmm1
	pand	%xmm2, %xmm3
	pand	%xmm1, %xmm3

	pmovmskb %xmm3, %eax
	incw	%ax
	jnz	L(64bytesormore_loop_end)
	ret

L(L2_L3_cache_unaligned):
	subq	$64, %rdx
	.p2align 4
L(L2_L3_unaligned_128bytes_loop):
	prefetchnta 0x1c0(%rdi)
	prefetchnta 0x1c0(%rsi)

	movdqu	(%rdi), %xmm0
	movdqu	16(%rdi), %xmm1
	movdqu	32(%rdi), %xmm2
	movdqu	48(%rdi), %xmm3

	CMPEQ	(%rsi), %xmm0
	CMPEQ	16(%rsi), %xmm1
	CMPEQ	32(%rsi), %xmm2
	CMPEQ	48(%rsi), %xmm3

	pand	%xmm0, %xmm1
	pand	%xmm2, %xmm3
	pand	%xmm1, %xmm3

	pmovmskb %xmm3, %eax
	incw	%ax
	jnz	L(64bytesormore_loop_end)

	add	$64, %rsi
	add	$64, %rdi
	sub	$64, %rdx
	ja	L(L2_L3_unaligned_128bytes_loop)
	jmp	L(loop_tail)


	/* This case is for machines which are sensitive for unaligned
	 * instructions.  */
	.p2align 4
L(2aligned):
	cmp	$128, %rdx
	ja	L(128bytesormorein2aligned)
L(less128bytesin2aligned):
	movdqa	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqa	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqa	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqa	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	cmp	$96, %rdx
	jb	L(32_to_64_bytes)

	addq	$64, %rdi
	addq	$64, %rsi
	subq	$64, %rdx

	.p2align 4,, 6
L(aligned_last_64_bytes):
	movdqa	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqa	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqu	-32(%rdi, %rdx), %xmm0
	movdqu	-32(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end_16)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret

	.p2align 4
L(128bytesormorein2aligned):
	cmp	$256, %rdx
	ja	L(aligned_loop)
L(less256bytesin2alinged):
	movdqa	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqa	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqa	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqa	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	addq	$64, %rdi
	addq	$64, %rsi

	movdqa	(%rdi), %xmm1
	CMPEQ	(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin)

	movdqa	16(%rdi), %xmm1
	CMPEQ	16(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_16)

	movdqa	32(%rdi), %xmm1
	CMPEQ	32(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_32)

	movdqa	48(%rdi), %xmm1
	CMPEQ	48(%rsi), %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_begin_48)

	addq	$-128, %rdx
	subq	$-64, %rsi
	subq	$-64, %rdi

	cmp	$64, %rdx
	ja	L(less128bytesin2aligned)

	cmp	$32, %rdx
	ja	L(aligned_last_64_bytes)

	movdqu	-32(%rdi, %rdx), %xmm0
	movdqu	-32(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end_16)

	movdqu	-16(%rdi, %rdx), %xmm0
	movdqu	-16(%rsi, %rdx), %xmm1
	CMPEQ	%xmm0, %xmm1
	pmovmskb %xmm1, %eax
	incw	%ax
	jnz	L(vec_return_end)
	ret

	.p2align 4
L(aligned_loop):
# ifdef DATA_CACHE_SIZE_HALF
	mov	$DATA_CACHE_SIZE_HALF, %R8_LP
# else
	mov	__x86_data_cache_size_half(%rip), %R8_LP
# endif
	movq	%r8, %r9
	addq	%r8, %r8
	addq	%r9, %r8
	cmpq	%r8, %rdx
	ja	L(L2_L3_cache_aligned)

	sub	$64, %rdx
	.p2align 4
L(64bytesormore_loopin2aligned):
	movdqa	(%rdi), %xmm0
	movdqa	16(%rdi), %xmm1
	movdqa	32(%rdi), %xmm2
	movdqa	48(%rdi), %xmm3

	CMPEQ	(%rsi), %xmm0
	CMPEQ	16(%rsi), %xmm1
	CMPEQ	32(%rsi), %xmm2
	CMPEQ	48(%rsi), %xmm3

	pand	%xmm0, %xmm1
	pand	%xmm2, %xmm3
	pand	%xmm1, %xmm3

	pmovmskb %xmm3, %eax
	incw	%ax
	jnz	L(64bytesormore_loop_end)
	add	$64, %rsi
	add	$64, %rdi
	sub	$64, %rdx
	ja	L(64bytesormore_loopin2aligned)
	jmp	L(loop_tail)

L(L2_L3_cache_aligned):
	subq	$64, %rdx
	.p2align 4
L(L2_L3_aligned_128bytes_loop):
	prefetchnta 0x1c0(%rdi)
	prefetchnta 0x1c0(%rsi)
	movdqa	(%rdi), %xmm0
	movdqa	16(%rdi), %xmm1
	movdqa	32(%rdi), %xmm2
	movdqa	48(%rdi), %xmm3

	CMPEQ	(%rsi), %xmm0
	CMPEQ	16(%rsi), %xmm1
	CMPEQ	32(%rsi), %xmm2
	CMPEQ	48(%rsi), %xmm3

	pand	%xmm0, %xmm1
	pand	%xmm2, %xmm3
	pand	%xmm1, %xmm3

	pmovmskb %xmm3, %eax
	incw	%ax
	jnz	L(64bytesormore_loop_end)

	addq	$64, %rsi
	addq	$64, %rdi
	subq	$64, %rdx
	ja	L(L2_L3_aligned_128bytes_loop)
	jmp	L(loop_tail)

	.p2align 4
L(64bytesormore_loop_end):
	pmovmskb %xmm0, %ecx
	incw	%cx
	jnz	L(loop_end_ret)

	pmovmskb %xmm1, %ecx
	notw	%cx
	sall	$16, %ecx
	jnz	L(loop_end_ret)

	pmovmskb %xmm2, %ecx
	notw	%cx
	shlq	$32, %rcx
	jnz	L(loop_end_ret)

	addq	$48, %rdi
	addq	$48, %rsi
	movq	%rax, %rcx

	.p2align 4,, 6
L(loop_end_ret):
	bsfq	%rcx, %rcx
# ifdef USE_AS_WMEMCMP
	movl	(%rdi, %rcx), %eax
	xorl	%edx, %edx
	cmpl	(%rsi, %rcx), %eax
	setg	%dl
	leal	-1(%rdx, %rdx), %eax
# else
	movzbl	(%rdi, %rcx), %eax
	movzbl	(%rsi, %rcx), %ecx
	subl	%ecx, %eax
# endif
	ret
END (MEMCMP)
#endif