File: GenericOpcode.html

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-20
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 1,496,436 kB
  • sloc: cpp: 5,593,990; ansic: 986,873; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,547; xml: 953; cs: 573; fortran: 567
file content (1036 lines) | stat: -rw-r--r-- 63,237 bytes parent folder | download | duplicates (7)
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


<!DOCTYPE html>

<html>
  <head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>Generic Opcodes &#8212; LLVM 13 documentation</title>
    <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
    <link rel="stylesheet" href="../_static/llvm-theme.css" type="text/css" />
    <script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
    <script src="../_static/jquery.js"></script>
    <script src="../_static/underscore.js"></script>
    <script src="../_static/doctools.js"></script>
    <link rel="index" title="Index" href="../genindex.html" />
    <link rel="search" title="Search" href="../search.html" />
    <link rel="next" title="Core Pipeline" href="Pipeline.html" />
    <link rel="prev" title="Generic Machine IR" href="GMIR.html" />
<style type="text/css">
  table.right { float: right; margin-left: 20px; }
  table.right td { border: 1px solid #ccc; }
</style>

  </head><body>
<div class="logo">
  <a href="../index.html">
    <img src="../_static/logo.png"
         alt="LLVM Logo" width="250" height="88"/></a>
</div>

    <div class="related" role="navigation" aria-label="related navigation">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="../genindex.html" title="General Index"
             accesskey="I">index</a></li>
        <li class="right" >
          <a href="Pipeline.html" title="Core Pipeline"
             accesskey="N">next</a> |</li>
        <li class="right" >
          <a href="GMIR.html" title="Generic Machine IR"
             accesskey="P">previous</a> |</li>
  <li><a href="https://llvm.org/">LLVM Home</a>&nbsp;|&nbsp;</li>
  <li><a href="../index.html">Documentation</a>&raquo;</li>

          <li class="nav-item nav-item-1"><a href="../Reference.html" >Reference</a> &#187;</li>
          <li class="nav-item nav-item-2"><a href="index.html" accesskey="U">Global Instruction Selection</a> &#187;</li>
        <li class="nav-item nav-item-this"><a href="">Generic Opcodes</a></li> 
      </ul>
    </div>

      <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
        <div class="sphinxsidebarwrapper">

<h3>Documentation</h3>

<ul class="want-points">
    <li><a href="https://llvm.org/docs/GettingStartedTutorials.html">Getting Started/Tutorials</a></li>
    <li><a href="https://llvm.org/docs/UserGuides.html">User Guides</a></li>
    <li><a href="https://llvm.org/docs/Reference.html">Reference</a></li>
</ul>

<h3>Getting Involved</h3>

<ul class="want-points">
    <li><a href="https://llvm.org/docs/Contributing.html">Contributing to LLVM</a></li>
    <li><a href="https://llvm.org/docs/HowToSubmitABug.html">Submitting Bug Reports</a></li>
    <li><a href="https://llvm.org/docs/GettingInvolved.html#mailing-lists">Mailing Lists</a></li>
    <li><a href="https://llvm.org/docs/GettingInvolved.html#irc">IRC</a></li>
    <li><a href="https://llvm.org/docs/GettingInvolved.html#meetups-and-social-events">Meetups and Social Events</a></li>
</ul>

<h3>Additional Links</h3>

<ul class="want-points">
    <li><a href="https://llvm.org/docs/FAQ.html">FAQ</a></li>
    <li><a href="https://llvm.org/docs/Lexicon.html">Glossary</a></li>
    <li><a href="https://llvm.org/pubs">Publications</a></li>
    <li><a href="https://github.com/llvm/llvm-project//">Github Repository</a></li>
</ul>
  <div role="note" aria-label="source link">
    <h3>This Page</h3>
    <ul class="this-page-menu">
      <li><a href="../_sources/GlobalISel/GenericOpcode.rst.txt"
            rel="nofollow">Show Source</a></li>
    </ul>
   </div>
<div id="searchbox" style="display: none" role="search">
  <h3 id="searchlabel">Quick search</h3>
    <div class="searchformwrapper">
    <form class="search" action="../search.html" method="get">
      <input type="text" name="q" aria-labelledby="searchlabel" />
      <input type="submit" value="Go" />
    </form>
    </div>
</div>
<script>$('#searchbox').show(0);</script>
        </div>
      </div>

    <div class="document">
      <div class="documentwrapper">
        <div class="bodywrapper">
          <div class="body" role="main">
            
  <div class="section" id="generic-opcodes">
<span id="gmir-opcodes"></span><h1>Generic Opcodes<a class="headerlink" href="#generic-opcodes" title="Permalink to this headline">¶</a></h1>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><p><a class="reference internal" href="#constants" id="id1">Constants</a></p>
<ul>
<li><p><a class="reference internal" href="#g-implicit-def" id="id2">G_IMPLICIT_DEF</a></p></li>
<li><p><a class="reference internal" href="#g-constant" id="id3">G_CONSTANT</a></p></li>
<li><p><a class="reference internal" href="#g-fconstant" id="id4">G_FCONSTANT</a></p></li>
<li><p><a class="reference internal" href="#g-frame-index" id="id5">G_FRAME_INDEX</a></p></li>
<li><p><a class="reference internal" href="#g-global-value" id="id6">G_GLOBAL_VALUE</a></p></li>
<li><p><a class="reference internal" href="#g-block-addr" id="id7">G_BLOCK_ADDR</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#integer-extension-and-truncation" id="id8">Integer Extension and Truncation</a></p>
<ul>
<li><p><a class="reference internal" href="#g-anyext" id="id9">G_ANYEXT</a></p></li>
<li><p><a class="reference internal" href="#g-sext" id="id10">G_SEXT</a></p></li>
<li><p><a class="reference internal" href="#g-sext-inreg" id="id11">G_SEXT_INREG</a></p></li>
<li><p><a class="reference internal" href="#g-zext" id="id12">G_ZEXT</a></p></li>
<li><p><a class="reference internal" href="#g-trunc" id="id13">G_TRUNC</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#type-conversions" id="id14">Type Conversions</a></p>
<ul>
<li><p><a class="reference internal" href="#g-inttoptr" id="id15">G_INTTOPTR</a></p></li>
<li><p><a class="reference internal" href="#g-ptrtoint" id="id16">G_PTRTOINT</a></p></li>
<li><p><a class="reference internal" href="#g-bitcast" id="id17">G_BITCAST</a></p></li>
<li><p><a class="reference internal" href="#g-addrspace-cast" id="id18">G_ADDRSPACE_CAST</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#scalar-operations" id="id19">Scalar Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-extract" id="id20">G_EXTRACT</a></p></li>
<li><p><a class="reference internal" href="#g-insert" id="id21">G_INSERT</a></p></li>
<li><p><a class="reference internal" href="#g-merge-values" id="id22">G_MERGE_VALUES</a></p></li>
<li><p><a class="reference internal" href="#g-unmerge-values" id="id23">G_UNMERGE_VALUES</a></p></li>
<li><p><a class="reference internal" href="#g-bswap" id="id24">G_BSWAP</a></p></li>
<li><p><a class="reference internal" href="#g-bitreverse" id="id25">G_BITREVERSE</a></p></li>
<li><p><a class="reference internal" href="#g-sbfx-g-ubfx" id="id26">G_SBFX, G_UBFX</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#integer-operations" id="id27">Integer Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-add-g-sub-g-mul-g-and-g-or-g-xor-g-sdiv-g-udiv-g-srem-g-urem" id="id28">G_ADD, G_SUB, G_MUL, G_AND, G_OR, G_XOR, G_SDIV, G_UDIV, G_SREM, G_UREM</a></p></li>
<li><p><a class="reference internal" href="#g-sdivrem-g-udivrem" id="id29">G_SDIVREM, G_UDIVREM</a></p></li>
<li><p><a class="reference internal" href="#g-saddsat-g-uaddsat-g-ssubsat-g-usubsat-g-sshlsat-g-ushlsat" id="id30">G_SADDSAT, G_UADDSAT, G_SSUBSAT, G_USUBSAT, G_SSHLSAT, G_USHLSAT</a></p></li>
<li><p><a class="reference internal" href="#g-shl-g-lshr-g-ashr" id="id31">G_SHL, G_LSHR, G_ASHR</a></p></li>
<li><p><a class="reference internal" href="#g-rotr-g-rotl" id="id32">G_ROTR, G_ROTL</a></p></li>
<li><p><a class="reference internal" href="#g-icmp" id="id33">G_ICMP</a></p></li>
<li><p><a class="reference internal" href="#g-select" id="id34">G_SELECT</a></p></li>
<li><p><a class="reference internal" href="#g-ptr-add" id="id35">G_PTR_ADD</a></p></li>
<li><p><a class="reference internal" href="#g-ptrmask" id="id36">G_PTRMASK</a></p></li>
<li><p><a class="reference internal" href="#g-smin-g-smax-g-umin-g-umax" id="id37">G_SMIN, G_SMAX, G_UMIN, G_UMAX</a></p></li>
<li><p><a class="reference internal" href="#g-abs" id="id38">G_ABS</a></p></li>
<li><p><a class="reference internal" href="#g-uaddo-g-saddo-g-usubo-g-ssubo-g-smulo-g-umulo" id="id39">G_UADDO, G_SADDO, G_USUBO, G_SSUBO, G_SMULO, G_UMULO</a></p></li>
<li><p><a class="reference internal" href="#g-uadde-g-sadde-g-usube-g-ssube" id="id40">G_UADDE, G_SADDE, G_USUBE, G_SSUBE</a></p></li>
<li><p><a class="reference internal" href="#g-umulh-g-smulh" id="id41">G_UMULH, G_SMULH</a></p></li>
<li><p><a class="reference internal" href="#g-ctlz-g-cttz-g-ctpop" id="id42">G_CTLZ, G_CTTZ, G_CTPOP</a></p></li>
<li><p><a class="reference internal" href="#g-ctlz-zero-undef-g-cttz-zero-undef" id="id43">G_CTLZ_ZERO_UNDEF, G_CTTZ_ZERO_UNDEF</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#floating-point-operations" id="id44">Floating Point Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-fcmp" id="id45">G_FCMP</a></p></li>
<li><p><a class="reference internal" href="#g-fneg" id="id46">G_FNEG</a></p></li>
<li><p><a class="reference internal" href="#g-fpext" id="id47">G_FPEXT</a></p></li>
<li><p><a class="reference internal" href="#g-fptrunc" id="id48">G_FPTRUNC</a></p></li>
<li><p><a class="reference internal" href="#g-fptosi-g-fptoui-g-sitofp-g-uitofp" id="id49">G_FPTOSI, G_FPTOUI, G_SITOFP, G_UITOFP</a></p></li>
<li><p><a class="reference internal" href="#g-fabs" id="id50">G_FABS</a></p></li>
<li><p><a class="reference internal" href="#g-fcopysign" id="id51">G_FCOPYSIGN</a></p></li>
<li><p><a class="reference internal" href="#g-fcanonicalize" id="id52">G_FCANONICALIZE</a></p></li>
<li><p><a class="reference internal" href="#g-fminnum" id="id53">G_FMINNUM</a></p></li>
<li><p><a class="reference internal" href="#g-fmaxnum" id="id54">G_FMAXNUM</a></p></li>
<li><p><a class="reference internal" href="#g-fminnum-ieee" id="id55">G_FMINNUM_IEEE</a></p></li>
<li><p><a class="reference internal" href="#g-fmaxnum-ieee" id="id56">G_FMAXNUM_IEEE</a></p></li>
<li><p><a class="reference internal" href="#g-fminimum" id="id57">G_FMINIMUM</a></p></li>
<li><p><a class="reference internal" href="#g-fmaximum" id="id58">G_FMAXIMUM</a></p></li>
<li><p><a class="reference internal" href="#g-fadd-g-fsub-g-fmul-g-fdiv-g-frem" id="id59">G_FADD, G_FSUB, G_FMUL, G_FDIV, G_FREM</a></p></li>
<li><p><a class="reference internal" href="#g-fma" id="id60">G_FMA</a></p></li>
<li><p><a class="reference internal" href="#g-fmad" id="id61">G_FMAD</a></p></li>
<li><p><a class="reference internal" href="#g-fpow" id="id62">G_FPOW</a></p></li>
<li><p><a class="reference internal" href="#g-fexp-g-fexp2" id="id63">G_FEXP, G_FEXP2</a></p></li>
<li><p><a class="reference internal" href="#g-flog-g-flog2-g-flog10" id="id64">G_FLOG, G_FLOG2, G_FLOG10</a></p></li>
<li><p><a class="reference internal" href="#g-fceil-g-fcos-g-fsin-g-fsqrt-g-ffloor-g-frint-g-fnearbyint" id="id65">G_FCEIL, G_FCOS, G_FSIN, G_FSQRT, G_FFLOOR, G_FRINT, G_FNEARBYINT</a></p></li>
<li><p><a class="reference internal" href="#g-intrinsic-trunc" id="id66">G_INTRINSIC_TRUNC</a></p></li>
<li><p><a class="reference internal" href="#g-intrinsic-round" id="id67">G_INTRINSIC_ROUND</a></p></li>
<li><p><a class="reference internal" href="#g-lround-g-llround" id="id68">G_LROUND, G_LLROUND</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#vector-specific-operations" id="id69">Vector Specific Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-concat-vectors" id="id70">G_CONCAT_VECTORS</a></p></li>
<li><p><a class="reference internal" href="#g-build-vector-g-build-vector-trunc" id="id71">G_BUILD_VECTOR, G_BUILD_VECTOR_TRUNC</a></p></li>
<li><p><a class="reference internal" href="#g-insert-vector-elt" id="id72">G_INSERT_VECTOR_ELT</a></p></li>
<li><p><a class="reference internal" href="#g-extract-vector-elt" id="id73">G_EXTRACT_VECTOR_ELT</a></p></li>
<li><p><a class="reference internal" href="#g-shuffle-vector" id="id74">G_SHUFFLE_VECTOR</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#vector-reduction-operations" id="id75">Vector Reduction Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-vecreduce-seq-fadd-g-vecreduce-seq-fmul" id="id76">G_VECREDUCE_SEQ_FADD, G_VECREDUCE_SEQ_FMUL</a></p></li>
<li><p><a class="reference internal" href="#g-vecreduce-fadd-g-vecreduce-fmul" id="id77">G_VECREDUCE_FADD, G_VECREDUCE_FMUL</a></p></li>
<li><p><a class="reference internal" href="#g-vecreduce-fmax-g-vecreduce-fmin" id="id78">G_VECREDUCE_FMAX, G_VECREDUCE_FMIN</a></p></li>
<li><p><a class="reference internal" href="#integer-bitwise-reductions" id="id79">Integer/bitwise reductions</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#memory-operations" id="id80">Memory Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-load-g-sextload-g-zextload" id="id81">G_LOAD, G_SEXTLOAD, G_ZEXTLOAD</a></p></li>
<li><p><a class="reference internal" href="#g-indexed-load" id="id82">G_INDEXED_LOAD</a></p></li>
<li><p><a class="reference internal" href="#g-indexed-sextload" id="id83">G_INDEXED_SEXTLOAD</a></p></li>
<li><p><a class="reference internal" href="#g-indexed-zextload" id="id84">G_INDEXED_ZEXTLOAD</a></p></li>
<li><p><a class="reference internal" href="#g-store" id="id85">G_STORE</a></p></li>
<li><p><a class="reference internal" href="#g-indexed-store" id="id86">G_INDEXED_STORE</a></p></li>
<li><p><a class="reference internal" href="#g-atomic-cmpxchg-with-success" id="id87">G_ATOMIC_CMPXCHG_WITH_SUCCESS</a></p></li>
<li><p><a class="reference internal" href="#g-atomic-cmpxchg" id="id88">G_ATOMIC_CMPXCHG</a></p></li>
<li><p><a class="reference internal" href="#g-atomicrmw-xchg-g-atomicrmw-add-g-atomicrmw-sub-g-atomicrmw-and-g-atomicrmw-nand-g-atomicrmw-or-g-atomicrmw-xor-g-atomicrmw-max-g-atomicrmw-min-g-atomicrmw-umax-g-atomicrmw-umin-g-atomicrmw-fadd-g-atomicrmw-fsub" id="id89">G_ATOMICRMW_XCHG, G_ATOMICRMW_ADD, G_ATOMICRMW_SUB, G_ATOMICRMW_AND, G_ATOMICRMW_NAND, G_ATOMICRMW_OR, G_ATOMICRMW_XOR, G_ATOMICRMW_MAX, G_ATOMICRMW_MIN, G_ATOMICRMW_UMAX, G_ATOMICRMW_UMIN, G_ATOMICRMW_FADD, G_ATOMICRMW_FSUB</a></p></li>
<li><p><a class="reference internal" href="#g-fence" id="id90">G_FENCE</a></p></li>
<li><p><a class="reference internal" href="#g-memcpy" id="id91">G_MEMCPY</a></p></li>
<li><p><a class="reference internal" href="#g-memcpy-inline" id="id92">G_MEMCPY_INLINE</a></p></li>
<li><p><a class="reference internal" href="#g-memmove" id="id93">G_MEMMOVE</a></p></li>
<li><p><a class="reference internal" href="#g-memset" id="id94">G_MEMSET</a></p></li>
<li><p><a class="reference internal" href="#g-bzero" id="id95">G_BZERO</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#control-flow" id="id96">Control Flow</a></p>
<ul>
<li><p><a class="reference internal" href="#g-phi" id="id97">G_PHI</a></p></li>
<li><p><a class="reference internal" href="#g-br" id="id98">G_BR</a></p></li>
<li><p><a class="reference internal" href="#g-brcond" id="id99">G_BRCOND</a></p></li>
<li><p><a class="reference internal" href="#g-brindirect" id="id100">G_BRINDIRECT</a></p></li>
<li><p><a class="reference internal" href="#g-brjt" id="id101">G_BRJT</a></p></li>
<li><p><a class="reference internal" href="#g-jump-table" id="id102">G_JUMP_TABLE</a></p></li>
<li><p><a class="reference internal" href="#g-intrinsic-g-intrinsic-w-side-effects" id="id103">G_INTRINSIC, G_INTRINSIC_W_SIDE_EFFECTS</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#variadic-arguments" id="id104">Variadic Arguments</a></p>
<ul>
<li><p><a class="reference internal" href="#g-vastart" id="id105">G_VASTART</a></p></li>
<li><p><a class="reference internal" href="#g-vaarg" id="id106">G_VAARG</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#other-operations" id="id107">Other Operations</a></p>
<ul>
<li><p><a class="reference internal" href="#g-dyn-stackalloc" id="id108">G_DYN_STACKALLOC</a></p></li>
</ul>
</li>
<li><p><a class="reference internal" href="#optimization-hints" id="id109">Optimization Hints</a></p>
<ul>
<li><p><a class="reference internal" href="#g-assert-sext-g-assert-zext" id="id110">G_ASSERT_SEXT, G_ASSERT_ZEXT</a></p></li>
</ul>
</li>
</ul>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>This documentation does not yet fully account for vectors. Many of the
scalar/integer/floating-point operations can also take vectors.</p>
</div>
<div class="section" id="constants">
<h2><a class="toc-backref" href="#id1">Constants</a><a class="headerlink" href="#constants" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-implicit-def">
<h3><a class="toc-backref" href="#id2">G_IMPLICIT_DEF</a><a class="headerlink" href="#g-implicit-def" title="Permalink to this headline">¶</a></h3>
<p>An undefined value.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0:_(s32) = G_IMPLICIT_DEF
</pre></div>
</div>
</div>
<div class="section" id="g-constant">
<h3><a class="toc-backref" href="#id3">G_CONSTANT</a><a class="headerlink" href="#g-constant" title="Permalink to this headline">¶</a></h3>
<p>An integer constant.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0:_(s32) = G_CONSTANT i32 1
</pre></div>
</div>
</div>
<div class="section" id="g-fconstant">
<h3><a class="toc-backref" href="#id4">G_FCONSTANT</a><a class="headerlink" href="#g-fconstant" title="Permalink to this headline">¶</a></h3>
<p>A floating point constant.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0:_(s32) = G_FCONSTANT float 1.0
</pre></div>
</div>
</div>
<div class="section" id="g-frame-index">
<h3><a class="toc-backref" href="#id5">G_FRAME_INDEX</a><a class="headerlink" href="#g-frame-index" title="Permalink to this headline">¶</a></h3>
<p>The address of an object in the stack frame.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(p0) = G_FRAME_INDEX %stack.0.ptr0
</pre></div>
</div>
</div>
<div class="section" id="g-global-value">
<h3><a class="toc-backref" href="#id6">G_GLOBAL_VALUE</a><a class="headerlink" href="#g-global-value" title="Permalink to this headline">¶</a></h3>
<p>The address of a global value.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0(p0) = G_GLOBAL_VALUE @var_local
</pre></div>
</div>
</div>
<div class="section" id="g-block-addr">
<h3><a class="toc-backref" href="#id7">G_BLOCK_ADDR</a><a class="headerlink" href="#g-block-addr" title="Permalink to this headline">¶</a></h3>
<p>The address of a basic block.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0:_(p0) = G_BLOCK_ADDR blockaddress(@test_blockaddress, %ir-block.block)
</pre></div>
</div>
</div>
</div>
<div class="section" id="integer-extension-and-truncation">
<h2><a class="toc-backref" href="#id8">Integer Extension and Truncation</a><a class="headerlink" href="#integer-extension-and-truncation" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-anyext">
<h3><a class="toc-backref" href="#id9">G_ANYEXT</a><a class="headerlink" href="#g-anyext" title="Permalink to this headline">¶</a></h3>
<p>Extend the underlying scalar type of an operation, leaving the high bits
unspecified.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_ANYEXT %0:_(s16)
</pre></div>
</div>
</div>
<div class="section" id="g-sext">
<h3><a class="toc-backref" href="#id10">G_SEXT</a><a class="headerlink" href="#g-sext" title="Permalink to this headline">¶</a></h3>
<p>Sign extend the underlying scalar type of an operation, copying the sign bit
into the newly-created space.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_SEXT %0:_(s16)
</pre></div>
</div>
</div>
<div class="section" id="g-sext-inreg">
<h3><a class="toc-backref" href="#id11">G_SEXT_INREG</a><a class="headerlink" href="#g-sext-inreg" title="Permalink to this headline">¶</a></h3>
<p>Sign extend the value from an arbitrary bit position, copying the sign bit
into all bits above it. This is equivalent to a shl + ashr pair with an
appropriate shift amount. $sz is an immediate (MachineOperand::isImm()
returns true) to allow targets to have some bitwidths legal and others
lowered. This opcode is particularly useful if the target has sign-extension
instructions that are cheaper than the constituent shifts as the optimizer is
able to make decisions on whether it’s better to hang on to the G_SEXT_INREG
or to lower it and optimize the individual shifts.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_SEXT_INREG %0:_(s32), 16
</pre></div>
</div>
</div>
<div class="section" id="g-zext">
<h3><a class="toc-backref" href="#id12">G_ZEXT</a><a class="headerlink" href="#g-zext" title="Permalink to this headline">¶</a></h3>
<p>Zero extend the underlying scalar type of an operation, putting zero bits
into the newly-created space.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_ZEXT %0:_(s16)
</pre></div>
</div>
</div>
<div class="section" id="g-trunc">
<h3><a class="toc-backref" href="#id13">G_TRUNC</a><a class="headerlink" href="#g-trunc" title="Permalink to this headline">¶</a></h3>
<p>Truncate the underlying scalar type of an operation. This is equivalent to
G_EXTRACT for scalar types, but acts elementwise on vectors.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s16) = G_TRUNC %0:_(s32)
</pre></div>
</div>
</div>
</div>
<div class="section" id="type-conversions">
<h2><a class="toc-backref" href="#id14">Type Conversions</a><a class="headerlink" href="#type-conversions" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-inttoptr">
<h3><a class="toc-backref" href="#id15">G_INTTOPTR</a><a class="headerlink" href="#g-inttoptr" title="Permalink to this headline">¶</a></h3>
<p>Convert an integer to a pointer.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(p0) = G_INTTOPTR %0:_(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-ptrtoint">
<h3><a class="toc-backref" href="#id16">G_PTRTOINT</a><a class="headerlink" href="#g-ptrtoint" title="Permalink to this headline">¶</a></h3>
<p>Convert a pointer to an integer.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_PTRTOINT %0:_(p0)
</pre></div>
</div>
</div>
<div class="section" id="g-bitcast">
<h3><a class="toc-backref" href="#id17">G_BITCAST</a><a class="headerlink" href="#g-bitcast" title="Permalink to this headline">¶</a></h3>
<p>Reinterpret a value as a new type. This is usually done without
changing any bits but this is not always the case due a subtlety in the
definition of the <a class="reference internal" href="../LangRef.html#i-bitcast"><span class="std std-ref">LLVM-IR Bitcast Instruction</span></a>. It
is allowed to bitcast between pointers with the same size, but
different address spaces.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s64) = G_BITCAST %0:_(&lt;2 x s32&gt;)
</pre></div>
</div>
</div>
<div class="section" id="g-addrspace-cast">
<h3><a class="toc-backref" href="#id18">G_ADDRSPACE_CAST</a><a class="headerlink" href="#g-addrspace-cast" title="Permalink to this headline">¶</a></h3>
<p>Convert a pointer to an address space to a pointer to another address space.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(p1) = G_ADDRSPACE_CAST %0:_(p0)
</pre></div>
</div>
<div class="admonition caution">
<p class="admonition-title">Caution</p>
<p><a class="reference internal" href="../LangRef.html#i-addrspacecast"><span class="std std-ref">‘addrspacecast .. to’ Instruction</span></a> doesn’t mention what happens if the cast is simply
invalid (i.e. if the address spaces are disjoint).</p>
</div>
</div>
</div>
<div class="section" id="scalar-operations">
<h2><a class="toc-backref" href="#id19">Scalar Operations</a><a class="headerlink" href="#scalar-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-extract">
<h3><a class="toc-backref" href="#id20">G_EXTRACT</a><a class="headerlink" href="#g-extract" title="Permalink to this headline">¶</a></h3>
<p>Extract a register of the specified size, starting from the block given by
index. This will almost certainly be mapped to sub-register COPYs after
register banks have been selected.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%3:_(s32) = G_EXTRACT %2:_(s64), 32
</pre></div>
</div>
</div>
<div class="section" id="g-insert">
<h3><a class="toc-backref" href="#id21">G_INSERT</a><a class="headerlink" href="#g-insert" title="Permalink to this headline">¶</a></h3>
<p>Insert a smaller register into a larger one at the specified bit-index.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%2:_(s64) = G_INSERT %0:(_s64), %1:_(s32), 0
</pre></div>
</div>
</div>
<div class="section" id="g-merge-values">
<h3><a class="toc-backref" href="#id22">G_MERGE_VALUES</a><a class="headerlink" href="#g-merge-values" title="Permalink to this headline">¶</a></h3>
<p>Concatenate multiple registers of the same size into a wider register.
The input operands are always ordered from lowest bits to highest:</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%0:(s32) = G_MERGE_VALUES %bits_0_7:(s8), %bits_8_15:(s8),
                          %bits_16_23:(s8), %bits_24_31:(s8)
</pre></div>
</div>
</div>
<div class="section" id="g-unmerge-values">
<h3><a class="toc-backref" href="#id23">G_UNMERGE_VALUES</a><a class="headerlink" href="#g-unmerge-values" title="Permalink to this headline">¶</a></h3>
<p>Extract multiple registers of the specified size, starting from blocks given by
indexes. This will almost certainly be mapped to sub-register COPYs after
register banks have been selected.
The output operands are always ordered from lowest bits to highest:</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%bits_0_7:(s8), %bits_8_15:(s8),
    %bits_16_23:(s8), %bits_24_31:(s8) = G_UNMERGE_VALUES %0:(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-bswap">
<h3><a class="toc-backref" href="#id24">G_BSWAP</a><a class="headerlink" href="#g-bswap" title="Permalink to this headline">¶</a></h3>
<p>Reverse the order of the bytes in a scalar.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_BSWAP %0:_(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-bitreverse">
<h3><a class="toc-backref" href="#id25">G_BITREVERSE</a><a class="headerlink" href="#g-bitreverse" title="Permalink to this headline">¶</a></h3>
<p>Reverse the order of the bits in a scalar.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_BITREVERSE %0:_(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-sbfx-g-ubfx">
<h3><a class="toc-backref" href="#id26">G_SBFX, G_UBFX</a><a class="headerlink" href="#g-sbfx-g-ubfx" title="Permalink to this headline">¶</a></h3>
<p>Extract a range of bits from a register.</p>
<p>The source operands are registers as follows:</p>
<ul class="simple">
<li><p>Source</p></li>
<li><p>The least-significant bit for the extraction</p></li>
<li><p>The width of the extraction</p></li>
</ul>
<p>The least-significant bit (lsb) and width operands are in the range:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="mi">0</span> <span class="o">&lt;=</span> <span class="n">lsb</span> <span class="o">&lt;</span> <span class="n">lsb</span> <span class="o">+</span> <span class="n">width</span> <span class="o">&lt;=</span> <span class="n">source</span> <span class="n">bitwidth</span><span class="p">,</span> <span class="n">where</span> <span class="nb">all</span> <span class="n">values</span> <span class="n">are</span> <span class="n">unsigned</span>
</pre></div>
</div>
<p>G_SBFX sign-extends the result, while G_UBFX zero-extends the result.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>; Extract 5 bits starting at bit 1 from %x and store them in %a.
; Sign-extend the result.
;
; Example:
; %x = 0...0000[10110]1 ---&gt; %a = 1...111111[10110]
%lsb_one = G_CONSTANT i32 1
%width_five = G_CONSTANT i32 5
%a:_(s32) = G_SBFX %x, %lsb_one, %width_five

; Extract 3 bits starting at bit 2 from %x and store them in %b. Zero-extend
; the result.
;
; Example:
; %x = 1...11111[100]11 ---&gt; %b = 0...00000[100]
%lsb_two = G_CONSTANT i32 2
%width_three = G_CONSTANT i32 3
%b:_(s32) = G_UBFX %x, %lsb_two, %width_three
</pre></div>
</div>
</div>
</div>
<div class="section" id="integer-operations">
<h2><a class="toc-backref" href="#id27">Integer Operations</a><a class="headerlink" href="#integer-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-add-g-sub-g-mul-g-and-g-or-g-xor-g-sdiv-g-udiv-g-srem-g-urem">
<h3><a class="toc-backref" href="#id28">G_ADD, G_SUB, G_MUL, G_AND, G_OR, G_XOR, G_SDIV, G_UDIV, G_SREM, G_UREM</a><a class="headerlink" href="#g-add-g-sub-g-mul-g-and-g-or-g-xor-g-sdiv-g-udiv-g-srem-g-urem" title="Permalink to this headline">¶</a></h3>
<p>These each perform their respective integer arithmetic on a scalar.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%dst:_(s32) = G_ADD %src0:_(s32), %src1:_(s32)
</pre></div>
</div>
<p>The above example adds %src1 to %src0 and stores the result in %dst.</p>
</div>
<div class="section" id="g-sdivrem-g-udivrem">
<h3><a class="toc-backref" href="#id29">G_SDIVREM, G_UDIVREM</a><a class="headerlink" href="#g-sdivrem-g-udivrem" title="Permalink to this headline">¶</a></h3>
<p>Perform integer division and remainder thereby producing two results.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%div:_(s32), %rem:_(s32) = G_SDIVREM %0:_(s32), %1:_(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-saddsat-g-uaddsat-g-ssubsat-g-usubsat-g-sshlsat-g-ushlsat">
<h3><a class="toc-backref" href="#id30">G_SADDSAT, G_UADDSAT, G_SSUBSAT, G_USUBSAT, G_SSHLSAT, G_USHLSAT</a><a class="headerlink" href="#g-saddsat-g-uaddsat-g-ssubsat-g-usubsat-g-sshlsat-g-ushlsat" title="Permalink to this headline">¶</a></h3>
<p>Signed and unsigned addition, subtraction and left shift with saturation.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%2:_(s32) = G_SADDSAT %0:_(s32), %1:_(s32)
</pre></div>
</div>
</div>
<div class="section" id="g-shl-g-lshr-g-ashr">
<h3><a class="toc-backref" href="#id31">G_SHL, G_LSHR, G_ASHR</a><a class="headerlink" href="#g-shl-g-lshr-g-ashr" title="Permalink to this headline">¶</a></h3>
<p>Shift the bits of a scalar left or right inserting zeros (sign-bit for G_ASHR).</p>
</div>
<div class="section" id="g-rotr-g-rotl">
<h3><a class="toc-backref" href="#id32">G_ROTR, G_ROTL</a><a class="headerlink" href="#g-rotr-g-rotl" title="Permalink to this headline">¶</a></h3>
<p>Rotate the bits right (G_ROTR) or left (G_ROTL).</p>
</div>
<div class="section" id="g-icmp">
<h3><a class="toc-backref" href="#id33">G_ICMP</a><a class="headerlink" href="#g-icmp" title="Permalink to this headline">¶</a></h3>
<p>Perform integer comparison producing non-zero (true) or zero (false). It’s
target specific whether a true value is 1, ~0U, or some other non-zero value.</p>
</div>
<div class="section" id="g-select">
<h3><a class="toc-backref" href="#id34">G_SELECT</a><a class="headerlink" href="#g-select" title="Permalink to this headline">¶</a></h3>
<p>Select between two values depending on a zero/non-zero value.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%5:_(s32) = G_SELECT %4(s1), %6, %2
</pre></div>
</div>
</div>
<div class="section" id="g-ptr-add">
<h3><a class="toc-backref" href="#id35">G_PTR_ADD</a><a class="headerlink" href="#g-ptr-add" title="Permalink to this headline">¶</a></h3>
<p>Add a scalar offset in addressible units to a pointer. Addressible units are
typically bytes but this may vary between targets.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(p0) = G_PTR_ADD %0:_(p0), %1:_(s32)
</pre></div>
</div>
<div class="admonition caution">
<p class="admonition-title">Caution</p>
<p>There are currently no in-tree targets that use this with addressable units
not equal to 8 bit.</p>
</div>
</div>
<div class="section" id="g-ptrmask">
<h3><a class="toc-backref" href="#id36">G_PTRMASK</a><a class="headerlink" href="#g-ptrmask" title="Permalink to this headline">¶</a></h3>
<p>Zero out an arbitrary mask of bits of a pointer. The mask type must be
an integer, and the number of vector elements must match for all
operands. This corresponds to <cite>i_intr_llvm_ptrmask</cite>.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%2:_(p0) = G_PTRMASK %0, %1
</pre></div>
</div>
</div>
<div class="section" id="g-smin-g-smax-g-umin-g-umax">
<h3><a class="toc-backref" href="#id37">G_SMIN, G_SMAX, G_UMIN, G_UMAX</a><a class="headerlink" href="#g-smin-g-smax-g-umin-g-umax" title="Permalink to this headline">¶</a></h3>
<p>Take the minimum/maximum of two values.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%5:_(s32) = G_SMIN %6, %2
</pre></div>
</div>
</div>
<div class="section" id="g-abs">
<h3><a class="toc-backref" href="#id38">G_ABS</a><a class="headerlink" href="#g-abs" title="Permalink to this headline">¶</a></h3>
<p>Take the absolute value of a signed integer. The absolute value of the minimum
negative value (e.g. the 8-bit value <cite>0x80</cite>) is defined to be itself.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%1:_(s32) = G_ABS %0
</pre></div>
</div>
</div>
<div class="section" id="g-uaddo-g-saddo-g-usubo-g-ssubo-g-smulo-g-umulo">
<h3><a class="toc-backref" href="#id39">G_UADDO, G_SADDO, G_USUBO, G_SSUBO, G_SMULO, G_UMULO</a><a class="headerlink" href="#g-uaddo-g-saddo-g-usubo-g-ssubo-g-smulo-g-umulo" title="Permalink to this headline">¶</a></h3>
<p>Perform the requested arithmetic and produce a carry output in addition to the
normal result.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%3:_(s32), %4:_(s1) = G_UADDO %0, %1
</pre></div>
</div>
</div>
<div class="section" id="g-uadde-g-sadde-g-usube-g-ssube">
<h3><a class="toc-backref" href="#id40">G_UADDE, G_SADDE, G_USUBE, G_SSUBE</a><a class="headerlink" href="#g-uadde-g-sadde-g-usube-g-ssube" title="Permalink to this headline">¶</a></h3>
<p>Perform the requested arithmetic and consume a carry input in addition to the
normal input. Also produce a carry output in addition to the normal result.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%4:_(s32), %5:_(s1) = G_UADDE %0, %1, %3:_(s1)
</pre></div>
</div>
</div>
<div class="section" id="g-umulh-g-smulh">
<h3><a class="toc-backref" href="#id41">G_UMULH, G_SMULH</a><a class="headerlink" href="#g-umulh-g-smulh" title="Permalink to this headline">¶</a></h3>
<p>Multiply two numbers at twice the incoming bit width (signed) and return
the high half of the result.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%3:_(s32) = G_UMULH %0, %1
</pre></div>
</div>
</div>
<div class="section" id="g-ctlz-g-cttz-g-ctpop">
<h3><a class="toc-backref" href="#id42">G_CTLZ, G_CTTZ, G_CTPOP</a><a class="headerlink" href="#g-ctlz-g-cttz-g-ctpop" title="Permalink to this headline">¶</a></h3>
<p>Count leading zeros, trailing zeros, or number of set bits.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%2:_(s33) = G_CTLZ_ZERO_UNDEF %1
%2:_(s33) = G_CTTZ_ZERO_UNDEF %1
%2:_(s33) = G_CTPOP %1
</pre></div>
</div>
</div>
<div class="section" id="g-ctlz-zero-undef-g-cttz-zero-undef">
<h3><a class="toc-backref" href="#id43">G_CTLZ_ZERO_UNDEF, G_CTTZ_ZERO_UNDEF</a><a class="headerlink" href="#g-ctlz-zero-undef-g-cttz-zero-undef" title="Permalink to this headline">¶</a></h3>
<p>Count leading zeros or trailing zeros. If the value is zero then the result is
undefined.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%2:_(s33) = G_CTLZ_ZERO_UNDEF %1
%2:_(s33) = G_CTTZ_ZERO_UNDEF %1
</pre></div>
</div>
</div>
</div>
<div class="section" id="floating-point-operations">
<h2><a class="toc-backref" href="#id44">Floating Point Operations</a><a class="headerlink" href="#floating-point-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-fcmp">
<h3><a class="toc-backref" href="#id45">G_FCMP</a><a class="headerlink" href="#g-fcmp" title="Permalink to this headline">¶</a></h3>
<p>Perform floating point comparison producing non-zero (true) or zero
(false). It’s target specific whether a true value is 1, ~0U, or some other
non-zero value.</p>
</div>
<div class="section" id="g-fneg">
<h3><a class="toc-backref" href="#id46">G_FNEG</a><a class="headerlink" href="#g-fneg" title="Permalink to this headline">¶</a></h3>
<p>Floating point negation.</p>
</div>
<div class="section" id="g-fpext">
<h3><a class="toc-backref" href="#id47">G_FPEXT</a><a class="headerlink" href="#g-fpext" title="Permalink to this headline">¶</a></h3>
<p>Convert a floating point value to a larger type.</p>
</div>
<div class="section" id="g-fptrunc">
<h3><a class="toc-backref" href="#id48">G_FPTRUNC</a><a class="headerlink" href="#g-fptrunc" title="Permalink to this headline">¶</a></h3>
<p>Convert a floating point value to a narrower type.</p>
</div>
<div class="section" id="g-fptosi-g-fptoui-g-sitofp-g-uitofp">
<h3><a class="toc-backref" href="#id49">G_FPTOSI, G_FPTOUI, G_SITOFP, G_UITOFP</a><a class="headerlink" href="#g-fptosi-g-fptoui-g-sitofp-g-uitofp" title="Permalink to this headline">¶</a></h3>
<p>Convert between integer and floating point.</p>
</div>
<div class="section" id="g-fabs">
<h3><a class="toc-backref" href="#id50">G_FABS</a><a class="headerlink" href="#g-fabs" title="Permalink to this headline">¶</a></h3>
<p>Take the absolute value of a floating point value.</p>
</div>
<div class="section" id="g-fcopysign">
<h3><a class="toc-backref" href="#id51">G_FCOPYSIGN</a><a class="headerlink" href="#g-fcopysign" title="Permalink to this headline">¶</a></h3>
<p>Copy the value of the first operand, replacing the sign bit with that of the
second operand.</p>
</div>
<div class="section" id="g-fcanonicalize">
<h3><a class="toc-backref" href="#id52">G_FCANONICALIZE</a><a class="headerlink" href="#g-fcanonicalize" title="Permalink to this headline">¶</a></h3>
<p>See <a class="reference internal" href="../LangRef.html#i-intr-llvm-canonicalize"><span class="std std-ref">‘llvm.canonicalize.*’ Intrinsic</span></a>.</p>
</div>
<div class="section" id="g-fminnum">
<h3><a class="toc-backref" href="#id53">G_FMINNUM</a><a class="headerlink" href="#g-fminnum" title="Permalink to this headline">¶</a></h3>
<p>Perform floating-point minimum on two values.</p>
<p>In the case where a single input is a NaN (either signaling or quiet),
the non-NaN input is returned.</p>
<p>The return value of (FMINNUM 0.0, -0.0) could be either 0.0 or -0.0.</p>
</div>
<div class="section" id="g-fmaxnum">
<h3><a class="toc-backref" href="#id54">G_FMAXNUM</a><a class="headerlink" href="#g-fmaxnum" title="Permalink to this headline">¶</a></h3>
<p>Perform floating-point maximum on two values.</p>
<p>In the case where a single input is a NaN (either signaling or quiet),
the non-NaN input is returned.</p>
<p>The return value of (FMAXNUM 0.0, -0.0) could be either 0.0 or -0.0.</p>
</div>
<div class="section" id="g-fminnum-ieee">
<h3><a class="toc-backref" href="#id55">G_FMINNUM_IEEE</a><a class="headerlink" href="#g-fminnum-ieee" title="Permalink to this headline">¶</a></h3>
<p>Perform floating-point minimum on two values, following the IEEE-754 2008
definition. This differs from FMINNUM in the handling of signaling NaNs. If one
input is a signaling NaN, returns a quiet NaN.</p>
</div>
<div class="section" id="g-fmaxnum-ieee">
<h3><a class="toc-backref" href="#id56">G_FMAXNUM_IEEE</a><a class="headerlink" href="#g-fmaxnum-ieee" title="Permalink to this headline">¶</a></h3>
<p>Perform floating-point maximum on two values, following the IEEE-754 2008
definition. This differs from FMAXNUM in the handling of signaling NaNs. If one
input is a signaling NaN, returns a quiet NaN.</p>
</div>
<div class="section" id="g-fminimum">
<h3><a class="toc-backref" href="#id57">G_FMINIMUM</a><a class="headerlink" href="#g-fminimum" title="Permalink to this headline">¶</a></h3>
<p>NaN-propagating minimum that also treat -0.0 as less than 0.0. While
FMINNUM_IEEE follow IEEE 754-2008 semantics, FMINIMUM follows IEEE 754-2018
draft semantics.</p>
</div>
<div class="section" id="g-fmaximum">
<h3><a class="toc-backref" href="#id58">G_FMAXIMUM</a><a class="headerlink" href="#g-fmaximum" title="Permalink to this headline">¶</a></h3>
<p>NaN-propagating maximum that also treat -0.0 as less than 0.0. While
FMAXNUM_IEEE follow IEEE 754-2008 semantics, FMAXIMUM follows IEEE 754-2018
draft semantics.</p>
</div>
<div class="section" id="g-fadd-g-fsub-g-fmul-g-fdiv-g-frem">
<h3><a class="toc-backref" href="#id59">G_FADD, G_FSUB, G_FMUL, G_FDIV, G_FREM</a><a class="headerlink" href="#g-fadd-g-fsub-g-fmul-g-fdiv-g-frem" title="Permalink to this headline">¶</a></h3>
<p>Perform the specified floating point arithmetic.</p>
</div>
<div class="section" id="g-fma">
<h3><a class="toc-backref" href="#id60">G_FMA</a><a class="headerlink" href="#g-fma" title="Permalink to this headline">¶</a></h3>
<p>Perform a fused multiply add (i.e. without the intermediate rounding step).</p>
</div>
<div class="section" id="g-fmad">
<h3><a class="toc-backref" href="#id61">G_FMAD</a><a class="headerlink" href="#g-fmad" title="Permalink to this headline">¶</a></h3>
<p>Perform a non-fused multiply add (i.e. with the intermediate rounding step).</p>
</div>
<div class="section" id="g-fpow">
<h3><a class="toc-backref" href="#id62">G_FPOW</a><a class="headerlink" href="#g-fpow" title="Permalink to this headline">¶</a></h3>
<p>Raise the first operand to the power of the second.</p>
</div>
<div class="section" id="g-fexp-g-fexp2">
<h3><a class="toc-backref" href="#id63">G_FEXP, G_FEXP2</a><a class="headerlink" href="#g-fexp-g-fexp2" title="Permalink to this headline">¶</a></h3>
<p>Calculate the base-e or base-2 exponential of a value</p>
</div>
<div class="section" id="g-flog-g-flog2-g-flog10">
<h3><a class="toc-backref" href="#id64">G_FLOG, G_FLOG2, G_FLOG10</a><a class="headerlink" href="#g-flog-g-flog2-g-flog10" title="Permalink to this headline">¶</a></h3>
<p>Calculate the base-e, base-2, or base-10 respectively.</p>
</div>
<div class="section" id="g-fceil-g-fcos-g-fsin-g-fsqrt-g-ffloor-g-frint-g-fnearbyint">
<h3><a class="toc-backref" href="#id65">G_FCEIL, G_FCOS, G_FSIN, G_FSQRT, G_FFLOOR, G_FRINT, G_FNEARBYINT</a><a class="headerlink" href="#g-fceil-g-fcos-g-fsin-g-fsqrt-g-ffloor-g-frint-g-fnearbyint" title="Permalink to this headline">¶</a></h3>
<p>These correspond to the standard C functions of the same name.</p>
</div>
<div class="section" id="g-intrinsic-trunc">
<h3><a class="toc-backref" href="#id66">G_INTRINSIC_TRUNC</a><a class="headerlink" href="#g-intrinsic-trunc" title="Permalink to this headline">¶</a></h3>
<p>Returns the operand rounded to the nearest integer not larger in magnitude than the operand.</p>
</div>
<div class="section" id="g-intrinsic-round">
<h3><a class="toc-backref" href="#id67">G_INTRINSIC_ROUND</a><a class="headerlink" href="#g-intrinsic-round" title="Permalink to this headline">¶</a></h3>
<p>Returns the operand rounded to the nearest integer.</p>
</div>
<div class="section" id="g-lround-g-llround">
<h3><a class="toc-backref" href="#id68">G_LROUND, G_LLROUND</a><a class="headerlink" href="#g-lround-g-llround" title="Permalink to this headline">¶</a></h3>
<p>Returns the source operand rounded to the nearest integer with ties away from
zero.</p>
<p>See the LLVM LangRef entry on ‘<code class="docutils literal notranslate"><span class="pre">llvm.lround.*'</span></code> for details on behaviour.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%rounded_32:_(s32) = G_LROUND %round_me:_(s64)
%rounded_64:_(s64) = G_LLROUND %round_me:_(s64)
</pre></div>
</div>
</div>
</div>
<div class="section" id="vector-specific-operations">
<h2><a class="toc-backref" href="#id69">Vector Specific Operations</a><a class="headerlink" href="#vector-specific-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-concat-vectors">
<h3><a class="toc-backref" href="#id70">G_CONCAT_VECTORS</a><a class="headerlink" href="#g-concat-vectors" title="Permalink to this headline">¶</a></h3>
<p>Concatenate two vectors to form a longer vector.</p>
</div>
<div class="section" id="g-build-vector-g-build-vector-trunc">
<h3><a class="toc-backref" href="#id71">G_BUILD_VECTOR, G_BUILD_VECTOR_TRUNC</a><a class="headerlink" href="#g-build-vector-g-build-vector-trunc" title="Permalink to this headline">¶</a></h3>
<p>Create a vector from multiple scalar registers. No implicit
conversion is performed (i.e. the result element type must be the
same as all source operands)</p>
<p>The _TRUNC version truncates the larger operand types to fit the
destination vector elt type.</p>
</div>
<div class="section" id="g-insert-vector-elt">
<h3><a class="toc-backref" href="#id72">G_INSERT_VECTOR_ELT</a><a class="headerlink" href="#g-insert-vector-elt" title="Permalink to this headline">¶</a></h3>
<p>Insert an element into a vector</p>
</div>
<div class="section" id="g-extract-vector-elt">
<h3><a class="toc-backref" href="#id73">G_EXTRACT_VECTOR_ELT</a><a class="headerlink" href="#g-extract-vector-elt" title="Permalink to this headline">¶</a></h3>
<p>Extract an element from a vector</p>
</div>
<div class="section" id="g-shuffle-vector">
<h3><a class="toc-backref" href="#id74">G_SHUFFLE_VECTOR</a><a class="headerlink" href="#g-shuffle-vector" title="Permalink to this headline">¶</a></h3>
<p>Concatenate two vectors and shuffle the elements according to the mask operand.
The mask operand should be an IR Constant which exactly matches the
corresponding mask for the IR shufflevector instruction.</p>
</div>
</div>
<div class="section" id="vector-reduction-operations">
<h2><a class="toc-backref" href="#id75">Vector Reduction Operations</a><a class="headerlink" href="#vector-reduction-operations" title="Permalink to this headline">¶</a></h2>
<p>These operations represent horizontal vector reduction, producing a scalar result.</p>
<div class="section" id="g-vecreduce-seq-fadd-g-vecreduce-seq-fmul">
<h3><a class="toc-backref" href="#id76">G_VECREDUCE_SEQ_FADD, G_VECREDUCE_SEQ_FMUL</a><a class="headerlink" href="#g-vecreduce-seq-fadd-g-vecreduce-seq-fmul" title="Permalink to this headline">¶</a></h3>
<p>The SEQ variants perform reductions in sequential order. The first operand is
an initial scalar accumulator value, and the second operand is the vector to reduce.</p>
</div>
<div class="section" id="g-vecreduce-fadd-g-vecreduce-fmul">
<h3><a class="toc-backref" href="#id77">G_VECREDUCE_FADD, G_VECREDUCE_FMUL</a><a class="headerlink" href="#g-vecreduce-fadd-g-vecreduce-fmul" title="Permalink to this headline">¶</a></h3>
<p>These reductions are relaxed variants which may reduce the elements in any order.</p>
</div>
<div class="section" id="g-vecreduce-fmax-g-vecreduce-fmin">
<h3><a class="toc-backref" href="#id78">G_VECREDUCE_FMAX, G_VECREDUCE_FMIN</a><a class="headerlink" href="#g-vecreduce-fmax-g-vecreduce-fmin" title="Permalink to this headline">¶</a></h3>
<p>FMIN/FMAX nodes can have flags, for NaN/NoNaN variants.</p>
</div>
<div class="section" id="integer-bitwise-reductions">
<h3><a class="toc-backref" href="#id79">Integer/bitwise reductions</a><a class="headerlink" href="#integer-bitwise-reductions" title="Permalink to this headline">¶</a></h3>
<ul class="simple">
<li><p>G_VECREDUCE_ADD</p></li>
<li><p>G_VECREDUCE_MUL</p></li>
<li><p>G_VECREDUCE_AND</p></li>
<li><p>G_VECREDUCE_OR</p></li>
<li><p>G_VECREDUCE_XOR</p></li>
<li><p>G_VECREDUCE_SMAX</p></li>
<li><p>G_VECREDUCE_SMIN</p></li>
<li><p>G_VECREDUCE_UMAX</p></li>
<li><p>G_VECREDUCE_UMIN</p></li>
</ul>
<p>Integer reductions may have a result type larger than the vector element type.
However, the reduction is performed using the vector element type and the value
in the top bits is unspecified.</p>
</div>
</div>
<div class="section" id="memory-operations">
<h2><a class="toc-backref" href="#id80">Memory Operations</a><a class="headerlink" href="#memory-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-load-g-sextload-g-zextload">
<h3><a class="toc-backref" href="#id81">G_LOAD, G_SEXTLOAD, G_ZEXTLOAD</a><a class="headerlink" href="#g-load-g-sextload-g-zextload" title="Permalink to this headline">¶</a></h3>
<p>Generic load. Expects a MachineMemOperand in addition to explicit
operands. If the result size is larger than the memory size, the
high bits are undefined, sign-extended, or zero-extended respectively.</p>
<p>Only G_LOAD is valid if the result is a vector type. If the result is larger
than the memory size, the high elements are undefined (i.e. this is not a
per-element, vector anyextload)</p>
</div>
<div class="section" id="g-indexed-load">
<h3><a class="toc-backref" href="#id82">G_INDEXED_LOAD</a><a class="headerlink" href="#g-indexed-load" title="Permalink to this headline">¶</a></h3>
<p>Generic indexed load. Combines a GEP with a load. $newaddr is set to $base + $offset.
If $am is 0 (post-indexed), then the value is loaded from $base; if $am is 1 (pre-indexed)
then the value is loaded from $newaddr.</p>
</div>
<div class="section" id="g-indexed-sextload">
<h3><a class="toc-backref" href="#id83">G_INDEXED_SEXTLOAD</a><a class="headerlink" href="#g-indexed-sextload" title="Permalink to this headline">¶</a></h3>
<p>Same as G_INDEXED_LOAD except that the load performed is sign-extending, as with G_SEXTLOAD.</p>
</div>
<div class="section" id="g-indexed-zextload">
<h3><a class="toc-backref" href="#id84">G_INDEXED_ZEXTLOAD</a><a class="headerlink" href="#g-indexed-zextload" title="Permalink to this headline">¶</a></h3>
<p>Same as G_INDEXED_LOAD except that the load performed is zero-extending, as with G_ZEXTLOAD.</p>
</div>
<div class="section" id="g-store">
<h3><a class="toc-backref" href="#id85">G_STORE</a><a class="headerlink" href="#g-store" title="Permalink to this headline">¶</a></h3>
<p>Generic store. Expects a MachineMemOperand in addition to explicit
operands. If the stored value size is greater than the memory size,
the high bits are implicitly truncated. If this is a vector store, the
high elements are discarded (i.e. this does not function as a per-lane
vector, truncating store)</p>
</div>
<div class="section" id="g-indexed-store">
<h3><a class="toc-backref" href="#id86">G_INDEXED_STORE</a><a class="headerlink" href="#g-indexed-store" title="Permalink to this headline">¶</a></h3>
<p>Combines a store with a GEP. See description of G_INDEXED_LOAD for indexing behaviour.</p>
</div>
<div class="section" id="g-atomic-cmpxchg-with-success">
<h3><a class="toc-backref" href="#id87">G_ATOMIC_CMPXCHG_WITH_SUCCESS</a><a class="headerlink" href="#g-atomic-cmpxchg-with-success" title="Permalink to this headline">¶</a></h3>
<p>Generic atomic cmpxchg with internal success check. Expects a
MachineMemOperand in addition to explicit operands.</p>
</div>
<div class="section" id="g-atomic-cmpxchg">
<h3><a class="toc-backref" href="#id88">G_ATOMIC_CMPXCHG</a><a class="headerlink" href="#g-atomic-cmpxchg" title="Permalink to this headline">¶</a></h3>
<p>Generic atomic cmpxchg. Expects a MachineMemOperand in addition to explicit
operands.</p>
</div>
<div class="section" id="g-atomicrmw-xchg-g-atomicrmw-add-g-atomicrmw-sub-g-atomicrmw-and-g-atomicrmw-nand-g-atomicrmw-or-g-atomicrmw-xor-g-atomicrmw-max-g-atomicrmw-min-g-atomicrmw-umax-g-atomicrmw-umin-g-atomicrmw-fadd-g-atomicrmw-fsub">
<h3><a class="toc-backref" href="#id89">G_ATOMICRMW_XCHG, G_ATOMICRMW_ADD, G_ATOMICRMW_SUB, G_ATOMICRMW_AND, G_ATOMICRMW_NAND, G_ATOMICRMW_OR, G_ATOMICRMW_XOR, G_ATOMICRMW_MAX, G_ATOMICRMW_MIN, G_ATOMICRMW_UMAX, G_ATOMICRMW_UMIN, G_ATOMICRMW_FADD, G_ATOMICRMW_FSUB</a><a class="headerlink" href="#g-atomicrmw-xchg-g-atomicrmw-add-g-atomicrmw-sub-g-atomicrmw-and-g-atomicrmw-nand-g-atomicrmw-or-g-atomicrmw-xor-g-atomicrmw-max-g-atomicrmw-min-g-atomicrmw-umax-g-atomicrmw-umin-g-atomicrmw-fadd-g-atomicrmw-fsub" title="Permalink to this headline">¶</a></h3>
<p>Generic atomicrmw. Expects a MachineMemOperand in addition to explicit
operands.</p>
</div>
<div class="section" id="g-fence">
<h3><a class="toc-backref" href="#id90">G_FENCE</a><a class="headerlink" href="#g-fence" title="Permalink to this headline">¶</a></h3>
<div class="admonition caution">
<p class="admonition-title">Caution</p>
<p>I couldn’t find any documentation on this at the time of writing.</p>
</div>
</div>
<div class="section" id="g-memcpy">
<h3><a class="toc-backref" href="#id91">G_MEMCPY</a><a class="headerlink" href="#g-memcpy" title="Permalink to this headline">¶</a></h3>
<p>Generic memcpy. Expects two MachineMemOperands covering the store and load
respectively, in addition to explicit operands.</p>
</div>
<div class="section" id="g-memcpy-inline">
<h3><a class="toc-backref" href="#id92">G_MEMCPY_INLINE</a><a class="headerlink" href="#g-memcpy-inline" title="Permalink to this headline">¶</a></h3>
<p>Generic inlined memcpy. Like G_MEMCPY, but it is guaranteed that this version
will not be lowered as a call to an external function. Currently the size
operand is required to evaluate as a constant (not an immediate), though that is
expected to change when llvm.memcpy.inline is taught to support dynamic sizes.</p>
</div>
<div class="section" id="g-memmove">
<h3><a class="toc-backref" href="#id93">G_MEMMOVE</a><a class="headerlink" href="#g-memmove" title="Permalink to this headline">¶</a></h3>
<p>Generic memmove. Similar to G_MEMCPY, but the source and destination memory
ranges are allowed to overlap.</p>
</div>
<div class="section" id="g-memset">
<h3><a class="toc-backref" href="#id94">G_MEMSET</a><a class="headerlink" href="#g-memset" title="Permalink to this headline">¶</a></h3>
<p>Generic memset. Expects a MachineMemOperand in addition to explicit operands.</p>
</div>
<div class="section" id="g-bzero">
<h3><a class="toc-backref" href="#id95">G_BZERO</a><a class="headerlink" href="#g-bzero" title="Permalink to this headline">¶</a></h3>
<p>Generic bzero. Expects a MachineMemOperand in addition to explicit operands.</p>
</div>
</div>
<div class="section" id="control-flow">
<h2><a class="toc-backref" href="#id96">Control Flow</a><a class="headerlink" href="#control-flow" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-phi">
<h3><a class="toc-backref" href="#id97">G_PHI</a><a class="headerlink" href="#g-phi" title="Permalink to this headline">¶</a></h3>
<p>Implement the φ node in the SSA graph representing the function.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%dst(s8) = G_PHI %src1(s8), %bb.&lt;id1&gt;, %src2(s8), %bb.&lt;id2&gt;
</pre></div>
</div>
</div>
<div class="section" id="g-br">
<h3><a class="toc-backref" href="#id98">G_BR</a><a class="headerlink" href="#g-br" title="Permalink to this headline">¶</a></h3>
<p>Unconditional branch</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>G_BR %bb.&lt;id&gt;
</pre></div>
</div>
</div>
<div class="section" id="g-brcond">
<h3><a class="toc-backref" href="#id99">G_BRCOND</a><a class="headerlink" href="#g-brcond" title="Permalink to this headline">¶</a></h3>
<p>Conditional branch</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>G_BRCOND %condition, %basicblock.&lt;id&gt;
</pre></div>
</div>
</div>
<div class="section" id="g-brindirect">
<h3><a class="toc-backref" href="#id100">G_BRINDIRECT</a><a class="headerlink" href="#g-brindirect" title="Permalink to this headline">¶</a></h3>
<p>Indirect branch</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>G_BRINDIRECT %src(p0)
</pre></div>
</div>
</div>
<div class="section" id="g-brjt">
<h3><a class="toc-backref" href="#id101">G_BRJT</a><a class="headerlink" href="#g-brjt" title="Permalink to this headline">¶</a></h3>
<p>Indirect branch to jump table entry</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>G_BRJT %ptr(p0), %jti, %idx(s64)
</pre></div>
</div>
</div>
<div class="section" id="g-jump-table">
<h3><a class="toc-backref" href="#id102">G_JUMP_TABLE</a><a class="headerlink" href="#g-jump-table" title="Permalink to this headline">¶</a></h3>
<p>Generates a pointer to the address of the jump table specified by the source
operand. The source operand is a jump table index.
G_JUMP_TABLE can be used in conjunction with G_BRJT to support jump table
codegen with GlobalISel.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%dst:_(p0) = G_JUMP_TABLE %jump-table.0
</pre></div>
</div>
<p>The above example generates a pointer to the source jump table index.</p>
</div>
<div class="section" id="g-intrinsic-g-intrinsic-w-side-effects">
<h3><a class="toc-backref" href="#id103">G_INTRINSIC, G_INTRINSIC_W_SIDE_EFFECTS</a><a class="headerlink" href="#g-intrinsic-g-intrinsic-w-side-effects" title="Permalink to this headline">¶</a></h3>
<p>Call an intrinsic</p>
<p>The _W_SIDE_EFFECTS version is considered to have unknown side-effects and
as such cannot be reordered across other side-effecting instructions.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Unlike SelectionDAG, there is no _VOID variant. Both of these are permitted
to have zero, one, or multiple results.</p>
</div>
</div>
</div>
<div class="section" id="variadic-arguments">
<h2><a class="toc-backref" href="#id104">Variadic Arguments</a><a class="headerlink" href="#variadic-arguments" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-vastart">
<h3><a class="toc-backref" href="#id105">G_VASTART</a><a class="headerlink" href="#g-vastart" title="Permalink to this headline">¶</a></h3>
<div class="admonition caution">
<p class="admonition-title">Caution</p>
<p>I found no documentation for this instruction at the time of writing.</p>
</div>
</div>
<div class="section" id="g-vaarg">
<h3><a class="toc-backref" href="#id106">G_VAARG</a><a class="headerlink" href="#g-vaarg" title="Permalink to this headline">¶</a></h3>
<div class="admonition caution">
<p class="admonition-title">Caution</p>
<p>I found no documentation for this instruction at the time of writing.</p>
</div>
</div>
</div>
<div class="section" id="other-operations">
<h2><a class="toc-backref" href="#id107">Other Operations</a><a class="headerlink" href="#other-operations" title="Permalink to this headline">¶</a></h2>
<div class="section" id="g-dyn-stackalloc">
<h3><a class="toc-backref" href="#id108">G_DYN_STACKALLOC</a><a class="headerlink" href="#g-dyn-stackalloc" title="Permalink to this headline">¶</a></h3>
<p>Dynamically realigns the stack pointer to the specified size and alignment.
An alignment value of <cite>0</cite> or <cite>1</cite> means no specific alignment.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%8:_(p0) = G_DYN_STACKALLOC %7(s64), 32
</pre></div>
</div>
</div>
</div>
<div class="section" id="optimization-hints">
<h2><a class="toc-backref" href="#id109">Optimization Hints</a><a class="headerlink" href="#optimization-hints" title="Permalink to this headline">¶</a></h2>
<p>These instructions do not correspond to any target instructions. They act as
hints for various combines.</p>
<div class="section" id="g-assert-sext-g-assert-zext">
<h3><a class="toc-backref" href="#id110">G_ASSERT_SEXT, G_ASSERT_ZEXT</a><a class="headerlink" href="#g-assert-sext-g-assert-zext" title="Permalink to this headline">¶</a></h3>
<p>This signifies that the contents of a register were previously extended from a
smaller type.</p>
<p>The smaller type is denoted using an immediate operand. For scalars, this is the
width of the entire smaller type. For vectors, this is the width of the smaller
element type.</p>
<div class="highlight-none notranslate"><div class="highlight"><pre><span></span>%x_was_zexted:_(s32) = G_ASSERT_ZEXT %x(s32), 16
%y_was_zexted:_(&lt;2 x s32&gt;) = G_ASSERT_ZEXT %y(&lt;2 x s32&gt;), 16

%z_was_sexted:_(s32) = G_ASSERT_SEXT %z(s32), 8
</pre></div>
</div>
<p>G_ASSERT_SEXT and G_ASSERT_ZEXT act like copies, albeit with some restrictions.</p>
<p>The source and destination registers must</p>
<ul class="simple">
<li><p>Be virtual</p></li>
<li><p>Belong to the same register class</p></li>
<li><p>Belong to the same register bank</p></li>
</ul>
<p>It should always be safe to</p>
<ul class="simple">
<li><p>Look through the source register</p></li>
<li><p>Replace the destination register with the source register</p></li>
</ul>
</div>
</div>
</div>


            <div class="clearer"></div>
          </div>
        </div>
      </div>
      <div class="clearer"></div>
    </div>
    <div class="related" role="navigation" aria-label="related navigation">
      <h3>Navigation</h3>
      <ul>
        <li class="right" style="margin-right: 10px">
          <a href="../genindex.html" title="General Index"
             >index</a></li>
        <li class="right" >
          <a href="Pipeline.html" title="Core Pipeline"
             >next</a> |</li>
        <li class="right" >
          <a href="GMIR.html" title="Generic Machine IR"
             >previous</a> |</li>
  <li><a href="https://llvm.org/">LLVM Home</a>&nbsp;|&nbsp;</li>
  <li><a href="../index.html">Documentation</a>&raquo;</li>

          <li class="nav-item nav-item-1"><a href="../Reference.html" >Reference</a> &#187;</li>
          <li class="nav-item nav-item-2"><a href="index.html" >Global Instruction Selection</a> &#187;</li>
        <li class="nav-item nav-item-this"><a href="">Generic Opcodes</a></li> 
      </ul>
    </div>
    <div class="footer" role="contentinfo">
        &#169; Copyright 2003-2021, LLVM Project.
      Last updated on 2021-09-18.
      Created using <a href="https://www.sphinx-doc.org/">Sphinx</a> 3.5.4.
    </div>
  </body>
</html>