summaryrefslogtreecommitdiff
path: root/gcc/config/i386/xmmintrin.h
blob: a223562490ec9788d3abf2261389794cca04369b (plain)
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
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
/* Copyright (C) 2002-2013 Free Software Foundation, Inc.

   This file is part of GCC.

   GCC is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 3, or (at your option)
   any later version.

   GCC 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 General Public License for more details.

   Under Section 7 of GPL version 3, you are granted additional
   permissions described in the GCC Runtime Library Exception, version
   3.1, as published by the Free Software Foundation.

   You should have received a copy of the GNU General Public License and
   a copy of the GCC Runtime Library Exception along with this program;
   see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
   <http://www.gnu.org/licenses/>.  */

/* Implemented from the specification included in the Intel C++ Compiler
   User Guide and Reference, version 9.0.  */

#ifndef _XMMINTRIN_H_INCLUDED
#define _XMMINTRIN_H_INCLUDED

#ifndef __SSE__
# error "SSE instruction set not enabled"
#else

/* We need type definitions from the MMX header file.  */
#include <mmintrin.h>

/* Get _mm_malloc () and _mm_free ().  */
#include <mm_malloc.h>

/* The Intel API is flexible enough that we must allow aliasing with other
   vector types, and their scalar components.  */
typedef float __m128 __attribute__ ((__vector_size__ (16), __may_alias__));

/* Internal data types for implementing the intrinsics.  */
typedef float __v4sf __attribute__ ((__vector_size__ (16)));

/* Create a selector for use with the SHUFPS instruction.  */
#define _MM_SHUFFLE(fp3,fp2,fp1,fp0) \
 (((fp3) << 6) | ((fp2) << 4) | ((fp1) << 2) | (fp0))

/* Constants for use with _mm_prefetch.  */
enum _mm_hint
{
  _MM_HINT_T0 = 3,
  _MM_HINT_T1 = 2,
  _MM_HINT_T2 = 1,
  _MM_HINT_NTA = 0
};

/* Bits in the MXCSR.  */
#define _MM_EXCEPT_MASK       0x003f
#define _MM_EXCEPT_INVALID    0x0001
#define _MM_EXCEPT_DENORM     0x0002
#define _MM_EXCEPT_DIV_ZERO   0x0004
#define _MM_EXCEPT_OVERFLOW   0x0008
#define _MM_EXCEPT_UNDERFLOW  0x0010
#define _MM_EXCEPT_INEXACT    0x0020

#define _MM_MASK_MASK         0x1f80
#define _MM_MASK_INVALID      0x0080
#define _MM_MASK_DENORM       0x0100
#define _MM_MASK_DIV_ZERO     0x0200
#define _MM_MASK_OVERFLOW     0x0400
#define _MM_MASK_UNDERFLOW    0x0800
#define _MM_MASK_INEXACT      0x1000

#define _MM_ROUND_MASK        0x6000
#define _MM_ROUND_NEAREST     0x0000
#define _MM_ROUND_DOWN        0x2000
#define _MM_ROUND_UP          0x4000
#define _MM_ROUND_TOWARD_ZERO 0x6000

#define _MM_FLUSH_ZERO_MASK   0x8000
#define _MM_FLUSH_ZERO_ON     0x8000
#define _MM_FLUSH_ZERO_OFF    0x0000

/* Create a vector of zeros.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_setzero_ps (void)
{
  return __extension__ (__m128){ 0.0f, 0.0f, 0.0f, 0.0f };
}

/* Perform the respective operation on the lower SPFP (single-precision
   floating-point) values of A and B; the upper three SPFP values are
   passed through from A.  */

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_add_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_addss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sub_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_subss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mul_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_mulss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_div_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_divss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sqrt_ss (__m128 __A)
{
  return (__m128) __builtin_ia32_sqrtss ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_rcp_ss (__m128 __A)
{
  return (__m128) __builtin_ia32_rcpss ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_rsqrt_ss (__m128 __A)
{
  return (__m128) __builtin_ia32_rsqrtss ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_min_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_minss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_max_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_maxss ((__v4sf)__A, (__v4sf)__B);
}

/* Perform the respective operation on the four SPFP values in A and B.  */

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_add_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_addps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sub_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_subps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mul_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_mulps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_div_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_divps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sqrt_ps (__m128 __A)
{
  return (__m128) __builtin_ia32_sqrtps ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_rcp_ps (__m128 __A)
{
  return (__m128) __builtin_ia32_rcpps ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_rsqrt_ps (__m128 __A)
{
  return (__m128) __builtin_ia32_rsqrtps ((__v4sf)__A);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_min_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_minps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_max_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_maxps ((__v4sf)__A, (__v4sf)__B);
}

/* Perform logical bit-wise operations on 128-bit values.  */

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_and_ps (__m128 __A, __m128 __B)
{
  return __builtin_ia32_andps (__A, __B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_andnot_ps (__m128 __A, __m128 __B)
{
  return __builtin_ia32_andnps (__A, __B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_or_ps (__m128 __A, __m128 __B)
{
  return __builtin_ia32_orps (__A, __B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_xor_ps (__m128 __A, __m128 __B)
{
  return __builtin_ia32_xorps (__A, __B);
}

/* Perform a comparison on the lower SPFP values of A and B.  If the
   comparison is true, place a mask of all ones in the result, otherwise a
   mask of zeros.  The upper three SPFP values are passed through from A.  */

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpeq_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpeqss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmplt_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpltss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmple_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpless ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpgt_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movss ((__v4sf) __A,
					(__v4sf)
					__builtin_ia32_cmpltss ((__v4sf) __B,
								(__v4sf)
								__A));
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpge_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movss ((__v4sf) __A,
					(__v4sf)
					__builtin_ia32_cmpless ((__v4sf) __B,
								(__v4sf)
								__A));
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpneq_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpneqss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnlt_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpnltss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnle_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpnless ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpngt_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movss ((__v4sf) __A,
					(__v4sf)
					__builtin_ia32_cmpnltss ((__v4sf) __B,
								 (__v4sf)
								 __A));
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnge_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movss ((__v4sf) __A,
					(__v4sf)
					__builtin_ia32_cmpnless ((__v4sf) __B,
								 (__v4sf)
								 __A));
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpord_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpordss ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpunord_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpunordss ((__v4sf)__A, (__v4sf)__B);
}

/* Perform a comparison on the four SPFP values of A and B.  For each
   element, if the comparison is true, place a mask of all ones in the
   result, otherwise a mask of zeros.  */

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpeq_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpeqps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmplt_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpltps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmple_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpleps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpgt_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpgtps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpge_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpgeps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpneq_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpneqps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnlt_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpnltps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnle_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpnleps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpngt_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpngtps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpnge_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpngeps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpord_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpordps ((__v4sf)__A, (__v4sf)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cmpunord_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_cmpunordps ((__v4sf)__A, (__v4sf)__B);
}

/* Compare the lower SPFP values of A and B and return 1 if true
   and 0 if false.  */

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comieq_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comieq ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comilt_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comilt ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comile_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comile ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comigt_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comigt ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comige_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comige ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_comineq_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_comineq ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomieq_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomieq ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomilt_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomilt ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomile_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomile ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomigt_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomigt ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomige_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomige ((__v4sf)__A, (__v4sf)__B);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_ucomineq_ss (__m128 __A, __m128 __B)
{
  return __builtin_ia32_ucomineq ((__v4sf)__A, (__v4sf)__B);
}

/* Convert the lower SPFP value to a 32-bit integer according to the current
   rounding mode.  */
extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtss_si32 (__m128 __A)
{
  return __builtin_ia32_cvtss2si ((__v4sf) __A);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvt_ss2si (__m128 __A)
{
  return _mm_cvtss_si32 (__A);
}

#ifdef __x86_64__
/* Convert the lower SPFP value to a 32-bit integer according to the
   current rounding mode.  */

/* Intel intrinsic.  */
extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtss_si64 (__m128 __A)
{
  return __builtin_ia32_cvtss2si64 ((__v4sf) __A);
}

/* Microsoft intrinsic.  */
extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtss_si64x (__m128 __A)
{
  return __builtin_ia32_cvtss2si64 ((__v4sf) __A);
}
#endif

/* Convert the two lower SPFP values to 32-bit integers according to the
   current rounding mode.  Return the integers in packed form.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtps_pi32 (__m128 __A)
{
  return (__m64) __builtin_ia32_cvtps2pi ((__v4sf) __A);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvt_ps2pi (__m128 __A)
{
  return _mm_cvtps_pi32 (__A);
}

/* Truncate the lower SPFP value to a 32-bit integer.  */
extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvttss_si32 (__m128 __A)
{
  return __builtin_ia32_cvttss2si ((__v4sf) __A);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtt_ss2si (__m128 __A)
{
  return _mm_cvttss_si32 (__A);
}

#ifdef __x86_64__
/* Truncate the lower SPFP value to a 32-bit integer.  */

/* Intel intrinsic.  */
extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvttss_si64 (__m128 __A)
{
  return __builtin_ia32_cvttss2si64 ((__v4sf) __A);
}

/* Microsoft intrinsic.  */
extern __inline long long __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvttss_si64x (__m128 __A)
{
  return __builtin_ia32_cvttss2si64 ((__v4sf) __A);
}
#endif

/* Truncate the two lower SPFP values to 32-bit integers.  Return the
   integers in packed form.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvttps_pi32 (__m128 __A)
{
  return (__m64) __builtin_ia32_cvttps2pi ((__v4sf) __A);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtt_ps2pi (__m128 __A)
{
  return _mm_cvttps_pi32 (__A);
}

/* Convert B to a SPFP value and insert it as element zero in A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtsi32_ss (__m128 __A, int __B)
{
  return (__m128) __builtin_ia32_cvtsi2ss ((__v4sf) __A, __B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvt_si2ss (__m128 __A, int __B)
{
  return _mm_cvtsi32_ss (__A, __B);
}

#ifdef __x86_64__
/* Convert B to a SPFP value and insert it as element zero in A.  */

/* Intel intrinsic.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtsi64_ss (__m128 __A, long long __B)
{
  return (__m128) __builtin_ia32_cvtsi642ss ((__v4sf) __A, __B);
}

/* Microsoft intrinsic.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtsi64x_ss (__m128 __A, long long __B)
{
  return (__m128) __builtin_ia32_cvtsi642ss ((__v4sf) __A, __B);
}
#endif

/* Convert the two 32-bit values in B to SPFP form and insert them
   as the two lower elements in A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpi32_ps (__m128 __A, __m64 __B)
{
  return (__m128) __builtin_ia32_cvtpi2ps ((__v4sf) __A, (__v2si)__B);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvt_pi2ps (__m128 __A, __m64 __B)
{
  return _mm_cvtpi32_ps (__A, __B);
}

/* Convert the four signed 16-bit values in A to SPFP form.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpi16_ps (__m64 __A)
{
  __v4hi __sign;
  __v2si __hisi, __losi;
  __v4sf __zero, __ra, __rb;

  /* This comparison against zero gives us a mask that can be used to
     fill in the missing sign bits in the unpack operations below, so
     that we get signed values after unpacking.  */
  __sign = __builtin_ia32_pcmpgtw ((__v4hi)0LL, (__v4hi)__A);

  /* Convert the four words to doublewords.  */
  __losi = (__v2si) __builtin_ia32_punpcklwd ((__v4hi)__A, __sign);
  __hisi = (__v2si) __builtin_ia32_punpckhwd ((__v4hi)__A, __sign);

  /* Convert the doublewords to floating point two at a time.  */
  __zero = (__v4sf) _mm_setzero_ps ();
  __ra = __builtin_ia32_cvtpi2ps (__zero, __losi);
  __rb = __builtin_ia32_cvtpi2ps (__ra, __hisi);

  return (__m128) __builtin_ia32_movlhps (__ra, __rb);
}

/* Convert the four unsigned 16-bit values in A to SPFP form.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpu16_ps (__m64 __A)
{
  __v2si __hisi, __losi;
  __v4sf __zero, __ra, __rb;

  /* Convert the four words to doublewords.  */
  __losi = (__v2si) __builtin_ia32_punpcklwd ((__v4hi)__A, (__v4hi)0LL);
  __hisi = (__v2si) __builtin_ia32_punpckhwd ((__v4hi)__A, (__v4hi)0LL);

  /* Convert the doublewords to floating point two at a time.  */
  __zero = (__v4sf) _mm_setzero_ps ();
  __ra = __builtin_ia32_cvtpi2ps (__zero, __losi);
  __rb = __builtin_ia32_cvtpi2ps (__ra, __hisi);

  return (__m128) __builtin_ia32_movlhps (__ra, __rb);
}

/* Convert the low four signed 8-bit values in A to SPFP form.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpi8_ps (__m64 __A)
{
  __v8qi __sign;

  /* This comparison against zero gives us a mask that can be used to
     fill in the missing sign bits in the unpack operations below, so
     that we get signed values after unpacking.  */
  __sign = __builtin_ia32_pcmpgtb ((__v8qi)0LL, (__v8qi)__A);

  /* Convert the four low bytes to words.  */
  __A = (__m64) __builtin_ia32_punpcklbw ((__v8qi)__A, __sign);

  return _mm_cvtpi16_ps(__A);
}

/* Convert the low four unsigned 8-bit values in A to SPFP form.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpu8_ps(__m64 __A)
{
  __A = (__m64) __builtin_ia32_punpcklbw ((__v8qi)__A, (__v8qi)0LL);
  return _mm_cvtpu16_ps(__A);
}

/* Convert the four signed 32-bit values in A and B to SPFP form.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtpi32x2_ps(__m64 __A, __m64 __B)
{
  __v4sf __zero = (__v4sf) _mm_setzero_ps ();
  __v4sf __sfa = __builtin_ia32_cvtpi2ps (__zero, (__v2si)__A);
  __v4sf __sfb = __builtin_ia32_cvtpi2ps (__sfa, (__v2si)__B);
  return (__m128) __builtin_ia32_movlhps (__sfa, __sfb);
}

/* Convert the four SPFP values in A to four signed 16-bit integers.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtps_pi16(__m128 __A)
{
  __v4sf __hisf = (__v4sf)__A;
  __v4sf __losf = __builtin_ia32_movhlps (__hisf, __hisf);
  __v2si __hisi = __builtin_ia32_cvtps2pi (__hisf);
  __v2si __losi = __builtin_ia32_cvtps2pi (__losf);
  return (__m64) __builtin_ia32_packssdw (__hisi, __losi);
}

/* Convert the four SPFP values in A to four signed 8-bit integers.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtps_pi8(__m128 __A)
{
  __v4hi __tmp = (__v4hi) _mm_cvtps_pi16 (__A);
  return (__m64) __builtin_ia32_packsswb (__tmp, (__v4hi)0LL);
}

/* Selects four specific SPFP values from A and B based on MASK.  */
#ifdef __OPTIMIZE__
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_shuffle_ps (__m128 __A, __m128 __B, int const __mask)
{
  return (__m128) __builtin_ia32_shufps ((__v4sf)__A, (__v4sf)__B, __mask);
}
#else
#define _mm_shuffle_ps(A, B, MASK)					\
  ((__m128) __builtin_ia32_shufps ((__v4sf)(__m128)(A),			\
				   (__v4sf)(__m128)(B), (int)(MASK)))
#endif

/* Selects and interleaves the upper two SPFP values from A and B.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_unpackhi_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_unpckhps ((__v4sf)__A, (__v4sf)__B);
}

/* Selects and interleaves the lower two SPFP values from A and B.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_unpacklo_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_unpcklps ((__v4sf)__A, (__v4sf)__B);
}

/* Sets the upper two SPFP values with 64-bits of data loaded from P;
   the lower two values are passed through from A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_loadh_pi (__m128 __A, __m64 const *__P)
{
  return (__m128) __builtin_ia32_loadhps ((__v4sf)__A, (const __v2sf *)__P);
}

/* Stores the upper two SPFP values of A into P.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_storeh_pi (__m64 *__P, __m128 __A)
{
  __builtin_ia32_storehps ((__v2sf *)__P, (__v4sf)__A);
}

/* Moves the upper two values of B into the lower two values of A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_movehl_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movhlps ((__v4sf)__A, (__v4sf)__B);
}

/* Moves the lower two values of B into the upper two values of A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_movelh_ps (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movlhps ((__v4sf)__A, (__v4sf)__B);
}

/* Sets the lower two SPFP values with 64-bits of data loaded from P;
   the upper two values are passed through from A.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_loadl_pi (__m128 __A, __m64 const *__P)
{
  return (__m128) __builtin_ia32_loadlps ((__v4sf)__A, (const __v2sf *)__P);
}

/* Stores the lower two SPFP values of A into P.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_storel_pi (__m64 *__P, __m128 __A)
{
  __builtin_ia32_storelps ((__v2sf *)__P, (__v4sf)__A);
}

/* Creates a 4-bit mask from the most significant bits of the SPFP values.  */
extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_movemask_ps (__m128 __A)
{
  return __builtin_ia32_movmskps ((__v4sf)__A);
}

/* Return the contents of the control register.  */
extern __inline unsigned int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_getcsr (void)
{
  return __builtin_ia32_stmxcsr ();
}

/* Read exception bits from the control register.  */
extern __inline unsigned int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_GET_EXCEPTION_STATE (void)
{
  return _mm_getcsr() & _MM_EXCEPT_MASK;
}

extern __inline unsigned int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_GET_EXCEPTION_MASK (void)
{
  return _mm_getcsr() & _MM_MASK_MASK;
}

extern __inline unsigned int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_GET_ROUNDING_MODE (void)
{
  return _mm_getcsr() & _MM_ROUND_MASK;
}

extern __inline unsigned int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_GET_FLUSH_ZERO_MODE (void)
{
  return _mm_getcsr() & _MM_FLUSH_ZERO_MASK;
}

/* Set the control register to I.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_setcsr (unsigned int __I)
{
  __builtin_ia32_ldmxcsr (__I);
}

/* Set exception bits in the control register.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_SET_EXCEPTION_STATE(unsigned int __mask)
{
  _mm_setcsr((_mm_getcsr() & ~_MM_EXCEPT_MASK) | __mask);
}

extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_SET_EXCEPTION_MASK (unsigned int __mask)
{
  _mm_setcsr((_mm_getcsr() & ~_MM_MASK_MASK) | __mask);
}

extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_SET_ROUNDING_MODE (unsigned int __mode)
{
  _mm_setcsr((_mm_getcsr() & ~_MM_ROUND_MASK) | __mode);
}

extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_MM_SET_FLUSH_ZERO_MODE (unsigned int __mode)
{
  _mm_setcsr((_mm_getcsr() & ~_MM_FLUSH_ZERO_MASK) | __mode);
}

/* Create a vector with element 0 as F and the rest zero.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_set_ss (float __F)
{
  return __extension__ (__m128)(__v4sf){ __F, 0.0f, 0.0f, 0.0f };
}

/* Create a vector with all four elements equal to F.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_set1_ps (float __F)
{
  return __extension__ (__m128)(__v4sf){ __F, __F, __F, __F };
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_set_ps1 (float __F)
{
  return _mm_set1_ps (__F);
}

/* Create a vector with element 0 as *P and the rest zero.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_load_ss (float const *__P)
{
  return _mm_set_ss (*__P);
}

/* Create a vector with all four elements equal to *P.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_load1_ps (float const *__P)
{
  return _mm_set1_ps (*__P);
}

extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_load_ps1 (float const *__P)
{
  return _mm_load1_ps (__P);
}

/* Load four SPFP values from P.  The address must be 16-byte aligned.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_load_ps (float const *__P)
{
  return (__m128) *(__v4sf *)__P;
}

/* Load four SPFP values from P.  The address need not be 16-byte aligned.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_loadu_ps (float const *__P)
{
  return (__m128) __builtin_ia32_loadups (__P);
}

/* Load four SPFP values in reverse order.  The address must be aligned.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_loadr_ps (float const *__P)
{
  __v4sf __tmp = *(__v4sf *)__P;
  return (__m128) __builtin_ia32_shufps (__tmp, __tmp, _MM_SHUFFLE (0,1,2,3));
}

/* Create the vector [Z Y X W].  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_set_ps (const float __Z, const float __Y, const float __X, const float __W)
{
  return __extension__ (__m128)(__v4sf){ __W, __X, __Y, __Z };
}

/* Create the vector [W X Y Z].  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_setr_ps (float __Z, float __Y, float __X, float __W)
{
  return __extension__ (__m128)(__v4sf){ __Z, __Y, __X, __W };
}

/* Stores the lower SPFP value.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_store_ss (float *__P, __m128 __A)
{
  *__P = __builtin_ia32_vec_ext_v4sf ((__v4sf)__A, 0);
}

extern __inline float __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_cvtss_f32 (__m128 __A)
{
  return __builtin_ia32_vec_ext_v4sf ((__v4sf)__A, 0);
}

/* Store four SPFP values.  The address must be 16-byte aligned.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_store_ps (float *__P, __m128 __A)
{
  *(__v4sf *)__P = (__v4sf)__A;
}

/* Store four SPFP values.  The address need not be 16-byte aligned.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_storeu_ps (float *__P, __m128 __A)
{
  __builtin_ia32_storeups (__P, (__v4sf)__A);
}

/* Store the lower SPFP value across four words.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_store1_ps (float *__P, __m128 __A)
{
  __v4sf __va = (__v4sf)__A;
  __v4sf __tmp = __builtin_ia32_shufps (__va, __va, _MM_SHUFFLE (0,0,0,0));
  _mm_storeu_ps (__P, __tmp);
}

extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_store_ps1 (float *__P, __m128 __A)
{
  _mm_store1_ps (__P, __A);
}

/* Store four SPFP values in reverse order.  The address must be aligned.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_storer_ps (float *__P, __m128 __A)
{
  __v4sf __va = (__v4sf)__A;
  __v4sf __tmp = __builtin_ia32_shufps (__va, __va, _MM_SHUFFLE (0,1,2,3));
  _mm_store_ps (__P, __tmp);
}

/* Sets the low SPFP value of A from the low value of B.  */
extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_move_ss (__m128 __A, __m128 __B)
{
  return (__m128) __builtin_ia32_movss ((__v4sf)__A, (__v4sf)__B);
}

/* Extracts one of the four words of A.  The selector N must be immediate.  */
#ifdef __OPTIMIZE__
extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_extract_pi16 (__m64 const __A, int const __N)
{
  return __builtin_ia32_vec_ext_v4hi ((__v4hi)__A, __N);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pextrw (__m64 const __A, int const __N)
{
  return _mm_extract_pi16 (__A, __N);
}
#else
#define _mm_extract_pi16(A, N)	\
  ((int) __builtin_ia32_vec_ext_v4hi ((__v4hi)(__m64)(A), (int)(N)))

#define _m_pextrw(A, N) _mm_extract_pi16(A, N)
#endif

/* Inserts word D into one of four words of A.  The selector N must be
   immediate.  */
#ifdef __OPTIMIZE__
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_insert_pi16 (__m64 const __A, int const __D, int const __N)
{
  return (__m64) __builtin_ia32_vec_set_v4hi ((__v4hi)__A, __D, __N);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pinsrw (__m64 const __A, int const __D, int const __N)
{
  return _mm_insert_pi16 (__A, __D, __N);
}
#else
#define _mm_insert_pi16(A, D, N)				\
  ((__m64) __builtin_ia32_vec_set_v4hi ((__v4hi)(__m64)(A),	\
					(int)(D), (int)(N)))

#define _m_pinsrw(A, D, N) _mm_insert_pi16(A, D, N)
#endif

/* Compute the element-wise maximum of signed 16-bit values.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_max_pi16 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pmaxsw ((__v4hi)__A, (__v4hi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pmaxsw (__m64 __A, __m64 __B)
{
  return _mm_max_pi16 (__A, __B);
}

/* Compute the element-wise maximum of unsigned 8-bit values.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_max_pu8 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pmaxub ((__v8qi)__A, (__v8qi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pmaxub (__m64 __A, __m64 __B)
{
  return _mm_max_pu8 (__A, __B);
}

/* Compute the element-wise minimum of signed 16-bit values.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_min_pi16 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pminsw ((__v4hi)__A, (__v4hi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pminsw (__m64 __A, __m64 __B)
{
  return _mm_min_pi16 (__A, __B);
}

/* Compute the element-wise minimum of unsigned 8-bit values.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_min_pu8 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pminub ((__v8qi)__A, (__v8qi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pminub (__m64 __A, __m64 __B)
{
  return _mm_min_pu8 (__A, __B);
}

/* Create an 8-bit mask of the signs of 8-bit values.  */
extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_movemask_pi8 (__m64 __A)
{
  return __builtin_ia32_pmovmskb ((__v8qi)__A);
}

extern __inline int __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pmovmskb (__m64 __A)
{
  return _mm_movemask_pi8 (__A);
}

/* Multiply four unsigned 16-bit values in A by four unsigned 16-bit values
   in B and produce the high 16 bits of the 32-bit results.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_mulhi_pu16 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pmulhuw ((__v4hi)__A, (__v4hi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pmulhuw (__m64 __A, __m64 __B)
{
  return _mm_mulhi_pu16 (__A, __B);
}

/* Return a combination of the four 16-bit values in A.  The selector
   must be an immediate.  */
#ifdef __OPTIMIZE__
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_shuffle_pi16 (__m64 __A, int const __N)
{
  return (__m64) __builtin_ia32_pshufw ((__v4hi)__A, __N);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pshufw (__m64 __A, int const __N)
{
  return _mm_shuffle_pi16 (__A, __N);
}
#else
#define _mm_shuffle_pi16(A, N) \
  ((__m64) __builtin_ia32_pshufw ((__v4hi)(__m64)(A), (int)(N)))

#define _m_pshufw(A, N) _mm_shuffle_pi16 (A, N)
#endif

/* Conditionally store byte elements of A into P.  The high bit of each
   byte in the selector N determines whether the corresponding byte from
   A is stored.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_maskmove_si64 (__m64 __A, __m64 __N, char *__P)
{
  __builtin_ia32_maskmovq ((__v8qi)__A, (__v8qi)__N, __P);
}

extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_maskmovq (__m64 __A, __m64 __N, char *__P)
{
  _mm_maskmove_si64 (__A, __N, __P);
}

/* Compute the rounded averages of the unsigned 8-bit values in A and B.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_avg_pu8 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pavgb ((__v8qi)__A, (__v8qi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pavgb (__m64 __A, __m64 __B)
{
  return _mm_avg_pu8 (__A, __B);
}

/* Compute the rounded averages of the unsigned 16-bit values in A and B.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_avg_pu16 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_pavgw ((__v4hi)__A, (__v4hi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_pavgw (__m64 __A, __m64 __B)
{
  return _mm_avg_pu16 (__A, __B);
}

/* Compute the sum of the absolute differences of the unsigned 8-bit
   values in A and B.  Return the value in the lower 16-bit word; the
   upper words are cleared.  */
extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sad_pu8 (__m64 __A, __m64 __B)
{
  return (__m64) __builtin_ia32_psadbw ((__v8qi)__A, (__v8qi)__B);
}

extern __inline __m64 __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_m_psadbw (__m64 __A, __m64 __B)
{
  return _mm_sad_pu8 (__A, __B);
}

/* Loads one cache line from address P to a location "closer" to the
   processor.  The selector I specifies the type of prefetch operation.  */
#ifdef __OPTIMIZE__
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_prefetch (const void *__P, enum _mm_hint __I)
{
  __builtin_prefetch (__P, 0, __I);
}
#else
#define _mm_prefetch(P, I) \
  __builtin_prefetch ((P), 0, (I))
#endif

/* Stores the data in A to the address P without polluting the caches.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_stream_pi (__m64 *__P, __m64 __A)
{
  __builtin_ia32_movntq ((unsigned long long *)__P, (unsigned long long)__A);
}

/* Likewise.  The address must be 16-byte aligned.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_stream_ps (float *__P, __m128 __A)
{
  __builtin_ia32_movntps (__P, (__v4sf)__A);
}

/* Guarantees that every preceding store is globally visible before
   any subsequent store.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_sfence (void)
{
  __builtin_ia32_sfence ();
}

/* The execution of the next instruction is delayed by an implementation
   specific amount of time.  The instruction does not modify the
   architectural state.  */
extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__))
_mm_pause (void)
{
  __builtin_ia32_pause ();
}

/* Transpose the 4x4 matrix composed of row[0-3].  */
#define _MM_TRANSPOSE4_PS(row0, row1, row2, row3)			\
do {									\
  __v4sf __r0 = (row0), __r1 = (row1), __r2 = (row2), __r3 = (row3);	\
  __v4sf __t0 = __builtin_ia32_unpcklps (__r0, __r1);			\
  __v4sf __t1 = __builtin_ia32_unpcklps (__r2, __r3);			\
  __v4sf __t2 = __builtin_ia32_unpckhps (__r0, __r1);			\
  __v4sf __t3 = __builtin_ia32_unpckhps (__r2, __r3);			\
  (row0) = __builtin_ia32_movlhps (__t0, __t1);				\
  (row1) = __builtin_ia32_movhlps (__t1, __t0);				\
  (row2) = __builtin_ia32_movlhps (__t2, __t3);				\
  (row3) = __builtin_ia32_movhlps (__t3, __t2);				\
} while (0)

/* For backward source compatibility.  */
#ifdef __SSE2__
# include <emmintrin.h>
#endif

#endif /* __SSE__ */
#endif /* _XMMINTRIN_H_INCLUDED */