-
Notifications
You must be signed in to change notification settings - Fork 96
/
Copy pathquickstart.html
1110 lines (968 loc) · 108 KB
/
quickstart.html
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
<!DOCTYPE html>
<html lang="en" >
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" /><meta name="generator" content="Docutils 0.18.1: http://docutils.sourceforge.net/" />
<title>Quick Start — matx documentation</title>
<script data-cfasync="false">
document.documentElement.dataset.mode = localStorage.getItem("mode") || "";
document.documentElement.dataset.theme = localStorage.getItem("theme") || "light";
</script>
<!-- Loaded before other Sphinx assets -->
<link href="_static/styles/theme.css?digest=e353d410970836974a52" rel="stylesheet" />
<link href="_static/styles/bootstrap.css?digest=e353d410970836974a52" rel="stylesheet" />
<link href="_static/styles/pydata-sphinx-theme.css?digest=e353d410970836974a52" rel="stylesheet" />
<link href="_static/vendor/fontawesome/6.1.2/css/all.min.css?digest=e353d410970836974a52" rel="stylesheet" />
<link rel="preload" as="font" type="font/woff2" crossorigin href="_static/vendor/fontawesome/6.1.2/webfonts/fa-solid-900.woff2" />
<link rel="preload" as="font" type="font/woff2" crossorigin href="_static/vendor/fontawesome/6.1.2/webfonts/fa-brands-400.woff2" />
<link rel="preload" as="font" type="font/woff2" crossorigin href="_static/vendor/fontawesome/6.1.2/webfonts/fa-regular-400.woff2" />
<link rel="stylesheet" type="text/css" href="_static/pygments.css" />
<link rel="stylesheet" href="_static/styles/sphinx-book-theme.css?digest=14f4ca6b54d191a8c7657f6c759bf11a5fb86285" type="text/css" />
<!-- Pre-loaded scripts that we'll load fully later -->
<link rel="preload" as="script" href="_static/scripts/bootstrap.js?digest=e353d410970836974a52" />
<link rel="preload" as="script" href="_static/scripts/pydata-sphinx-theme.js?digest=e353d410970836974a52" />
<script data-url_root="./" id="documentation_options" src="_static/documentation_options.js"></script>
<script src="_static/doctools.js"></script>
<script src="_static/sphinx_highlight.js"></script>
<script src="_static/jquery.js"></script>
<script src="_static/_sphinx_javascript_frameworks_compat.js"></script>
<script src="_static/scripts/sphinx-book-theme.js?digest=5a5c038af52cf7bc1a1ec88eea08e6366ee68824"></script>
<script>DOCUMENTATION_OPTIONS.pagename = 'quickstart';</script>
<link rel="index" title="Index" href="genindex.html" />
<link rel="search" title="Search" href="search.html" />
<link rel="next" title="Build Integration" href="build.html" />
<link rel="prev" title="Welcome To MatX’s Documentation!" href="index.html" />
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<meta name="docsearch:language" content="en"/>
</head>
<body data-bs-spy="scroll" data-bs-target=".bd-toc-nav" data-offset="180" data-bs-root-margin="0px 0px -60%" data-default-mode="">
<a class="skip-link" href="#main-content">Skip to main content</a>
<input type="checkbox"
class="sidebar-toggle"
name="__primary"
id="__primary"/>
<label class="overlay overlay-primary" for="__primary"></label>
<input type="checkbox"
class="sidebar-toggle"
name="__secondary"
id="__secondary"/>
<label class="overlay overlay-secondary" for="__secondary"></label>
<div class="search-button__wrapper">
<div class="search-button__overlay"></div>
<div class="search-button__search-container">
<form class="bd-search d-flex align-items-center"
action="search.html"
method="get">
<i class="fa-solid fa-magnifying-glass"></i>
<input type="search"
class="form-control"
name="q"
id="search-input"
placeholder="Search..."
aria-label="Search..."
autocomplete="off"
autocorrect="off"
autocapitalize="off"
spellcheck="false"/>
<span class="search-button__kbd-shortcut"><kbd class="kbd-shortcut__modifier">Ctrl</kbd>+<kbd>K</kbd></span>
</form></div>
</div>
<nav class="bd-header navbar navbar-expand-lg bd-navbar">
</nav>
<div class="bd-container">
<div class="bd-container__inner bd-page-width">
<div class="bd-sidebar-primary bd-sidebar">
<div class="sidebar-header-items sidebar-primary__section">
</div>
<div class="sidebar-primary-items__start sidebar-primary__section">
<div class="sidebar-primary-item">
<a class="navbar-brand logo" href="index.html">
<p class="title logo__title">matx documentation</p>
</a></div>
<div class="sidebar-primary-item"><nav class="bd-links" id="bd-docs-nav" aria-label="Main">
<div class="bd-toc-item navbar-nav active">
<p aria-level="2" class="caption" role="heading"><span class="caption-text">MatX Documentation</span></p>
<ul class="current nav bd-sidenav">
<li class="toctree-l1 current active"><a class="current reference internal" href="#">Quick Start</a></li>
<li class="toctree-l1"><a class="reference internal" href="build.html">Build Integration</a></li>
<li class="toctree-l1 has-children"><a class="reference internal" href="basics/index.html">Basics</a><input class="toctree-checkbox" id="toctree-checkbox-1" name="toctree-checkbox-1" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-1"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l2"><a class="reference internal" href="basics/broadcast.html">Broadcasting</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/concepts.html">Concepts</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/creation.html">Creating Tensors</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/datatype.html">Data Types</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/fusion.html">Operator Fusion</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/matlabpython.html">MatX For MATLAB/NumPy Users</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/nvtx.html">NVTX Profiling</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/print.html">Printing</a></li>
<li class="toctree-l2"><a class="reference internal" href="basics/tensor.html">Tensor Type</a></li>
</ul>
</li>
<li class="toctree-l1 has-children"><a class="reference internal" href="api/index.html">API Reference</a><input class="toctree-checkbox" id="toctree-checkbox-2" name="toctree-checkbox-2" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-2"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/creation/index.html">Creating Operators and Tensors</a><input class="toctree-checkbox" id="toctree-checkbox-3" name="toctree-checkbox-3" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-3"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/creation/operators/index.html">Creation Operators</a><input class="toctree-checkbox" id="toctree-checkbox-4" name="toctree-checkbox-4" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-4"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/alternate.html">alternate</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/diag.html">diag</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/eye.html">eye</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/linspace.html">linspace</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/logspace.html">logspace</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/meshgrid.html">meshgrid</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/ones.html">ones</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/range.html">range</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/creation/operators/zeros.html">zeros</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/creation/tensors/index.html">Tensor Creation</a><input class="toctree-checkbox" id="toctree-checkbox-5" name="toctree-checkbox-5" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-5"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/creation/tensors/make.html">Making Tensors</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/manipulation/index.html">Manipulation Functions</a><input class="toctree-checkbox" id="toctree-checkbox-6" name="toctree-checkbox-6" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-6"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/manipulation/basic/index.html">Basic Manipulations</a><input class="toctree-checkbox" id="toctree-checkbox-7" name="toctree-checkbox-7" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-7"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/basic/copy.html">copy</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/basic/if.html">IF</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/basic/ifelse.html">IFELSE</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/basic/slice.html">slice</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/manipulation/joinrepeat/index.html">Joining and Repeating</a><input class="toctree-checkbox" id="toctree-checkbox-8" name="toctree-checkbox-8" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-8"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/clone.html">clone</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/concat.html">concat</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/flatten.html">flatten</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/lcollapse.html">lcollapse</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/rcollapse.html">rcollapse</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/repmat.html">repmat</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/joinrepeat/stack.html">stack</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/manipulation/rearranging/index.html">Rearranging</a><input class="toctree-checkbox" id="toctree-checkbox-9" name="toctree-checkbox-9" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-9"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/fliplr.html">fliplr</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/flipud.html">flipud</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/overlap.html">overlap</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/permute.html">permute</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/reshape.html">reshape</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/reverse.html">reverse</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/shift.html">shift</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/transpose.html">transpose</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/rearranging/transpose_matrix.html">transpose_matrix</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/manipulation/selecting/index.html">Selecting</a><input class="toctree-checkbox" id="toctree-checkbox-10" name="toctree-checkbox-10" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-10"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/selecting/at.html">at</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/selecting/reduce.html">reduce</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/selecting/remap.html">remap</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/manipulation/selecting/select.html">select</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/io/index.html">Input/Output</a><input class="toctree-checkbox" id="toctree-checkbox-11" name="toctree-checkbox-11" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-11"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/io/read_csv.html">read_csv</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/io/read_mat.html">read_mat</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/io/write_csv.html">write_csv</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/io/write_mat.html">write_mat</a></li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/math/index.html">Mathematical Functions</a><input class="toctree-checkbox" id="toctree-checkbox-12" name="toctree-checkbox-12" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-12"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/arithmetic/index.html">Arithmetic Operations</a><input class="toctree-checkbox" id="toctree-checkbox-13" name="toctree-checkbox-13" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-13"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/add.html">Addition (+)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/div.html">Divide (/)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/fmod.html">fmod</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/mod.html">Modulo (%)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/mul.html">Multiply (*)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/neg.html">Negate (-)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/pow.html">pow</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/arithmetic/sub.html">Subtract (-)</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/complex/index.html">Complex Numbers</a><input class="toctree-checkbox" id="toctree-checkbox-14" name="toctree-checkbox-14" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-14"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/angle.html">angle</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/conj.html">conj</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/hermitiant.html">hermitianT</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/imag.html">imag</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/norm.html">norm</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/complex/real.html">real</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/explog/index.html">Exponents and Logarithms</a><input class="toctree-checkbox" id="toctree-checkbox-15" name="toctree-checkbox-15" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-15"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/explog/exp.html">exp</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/explog/expj.html">expj</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/explog/log.html">log</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/explog/log10.html">log10</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/explog/log2.html">log2</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/extrema/index.html">Extrema Finding</a><input class="toctree-checkbox" id="toctree-checkbox-16" name="toctree-checkbox-16" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-16"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/extrema/max.html">max</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/extrema/min.html">min</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/extrema/rmax.html">rmax</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/extrema/rmin.html">rmin</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/misc/index.html">Miscellaneous</a><input class="toctree-checkbox" id="toctree-checkbox-17" name="toctree-checkbox-17" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-17"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/misc/abs.html">abs</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/misc/softmax.html">softmax</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/misc/sqrt.html">sqrt</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/round/index.html">Rounding</a><input class="toctree-checkbox" id="toctree-checkbox-18" name="toctree-checkbox-18" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-18"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/round/ceil.html">ceil</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/round/floor.html">floor</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/round/round.html">round</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/sumprod/index.html">Sums, Products, and Differences</a><input class="toctree-checkbox" id="toctree-checkbox-19" name="toctree-checkbox-19" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-19"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/sumprod/cumsum.html">cumsum</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/sumprod/prod.html">prod</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/sumprod/sum.html">sum</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/math/trig/index.html">Trigonometric Functions</a><input class="toctree-checkbox" id="toctree-checkbox-20" name="toctree-checkbox-20" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-20"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/acos.html">acos</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/acosh.html">acosh</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/asin.html">asin</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/asinh.html">asinh</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/atan.html">atan</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/atanh.html">atanh</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/cart2sph.html">cart2sph</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/cos.html">cos</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/cosh.html">cosh</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/sin.html">sin</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/sinh.html">sinh</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/sph2cart.html">sph2cart</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/tan.html">tan</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/math/trig/tanh.html">tanh</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/linalg/index.html">Linear Algebra</a><input class="toctree-checkbox" id="toctree-checkbox-21" name="toctree-checkbox-21" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-21"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/linalg/decomp/index.html">Decompositions</a><input class="toctree-checkbox" id="toctree-checkbox-22" name="toctree-checkbox-22" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-22"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/chol.html">chol</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/inverse.html">inv</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/lu.html">lu</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/qr.html">qr</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/svd.html">svd</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/svdbpi.html">svdbpi</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/decomp/svdpi.html">svdpi</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/linalg/eigenvalues/index.html">Eigenvalues and Eigenvectors</a><input class="toctree-checkbox" id="toctree-checkbox-23" name="toctree-checkbox-23" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-23"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/eigenvalues/eig.html">eig</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/linalg/matvec/index.html">Matrix and Vector Products</a><input class="toctree-checkbox" id="toctree-checkbox-24" name="toctree-checkbox-24" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-24"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/matvec/einsum.html">einsum</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/matvec/kron.html">kron</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/matvec/matmul.html">matmul</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/matvec/matvec.html">matvec</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/linalg/other/index.html">Other</a><input class="toctree-checkbox" id="toctree-checkbox-25" name="toctree-checkbox-25" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-25"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/other/cgsolve.html">cgsolve</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/linalg/other/trace.html">trace</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/stats/index.html">Statistics</a><input class="toctree-checkbox" id="toctree-checkbox-26" name="toctree-checkbox-26" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-26"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/stats/avgvar/index.html">Averages and Variances</a><input class="toctree-checkbox" id="toctree-checkbox-27" name="toctree-checkbox-27" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-27"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/stats/avgvar/mean.html">mean</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/stats/avgvar/median.html">median</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/stats/avgvar/stdd.html">stdd</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/stats/avgvar/var.html">var</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/stats/corr/index.html">Correlating</a><input class="toctree-checkbox" id="toctree-checkbox-28" name="toctree-checkbox-28" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-28"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/stats/corr/corr.html">corr</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/stats/corr/cov.html">cov</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/stats/hist/index.html">Histograms</a><input class="toctree-checkbox" id="toctree-checkbox-29" name="toctree-checkbox-29" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-29"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/stats/hist/hist.html">hist</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/stats/misc/index.html">Misc</a><input class="toctree-checkbox" id="toctree-checkbox-30" name="toctree-checkbox-30" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-30"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/stats/misc/percentile.html">percentile</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/searchsort/index.html">Searching and Sorting</a><input class="toctree-checkbox" id="toctree-checkbox-31" name="toctree-checkbox-31" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-31"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/argmax.html">argmax</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/argmin.html">argmin</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/find.html">find</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/findidx.html">find_idx</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/sort.html">sort</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/searchsort/unique.html">unique</a></li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/logic/index.html">Logic Functions</a><input class="toctree-checkbox" id="toctree-checkbox-32" name="toctree-checkbox-32" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-32"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/logic/bitwise/index.html">Bitwise Operations</a><input class="toctree-checkbox" id="toctree-checkbox-33" name="toctree-checkbox-33" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-33"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/logic/bitwise/and.html">Bitwise AND (&)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/bitwise/or.html">Bitwise OR (|)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/bitwise/xor.html">Bitwise XOR (^)</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/logic/comparison/index.html">Comparison Operations</a><input class="toctree-checkbox" id="toctree-checkbox-34" name="toctree-checkbox-34" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-34"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/equal.html">Equal (==)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/gt.html">Greater than (>)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/gte.html">Greater than or equal (>=)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/isclose.html">isclose</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/lt.html">Less than (<)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/lte.html">Less than or equal (<=)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/comparison/neq.html">Not-equal (!=)</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/logic/logical/index.html">Logical Operations</a><input class="toctree-checkbox" id="toctree-checkbox-35" name="toctree-checkbox-35" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-35"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/logic/logical/and.html">&& (AND)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/logical/not.html">! (NOT)</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/logical/or.html">|| (OR)</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/logic/truth/index.html">Truth Value Testing</a><input class="toctree-checkbox" id="toctree-checkbox-36" name="toctree-checkbox-36" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-36"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/logic/truth/all.html">all</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/truth/allclose.html">allclose</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/logic/truth/any.html">any</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/casting/index.html">Casting Operators</a><input class="toctree-checkbox" id="toctree-checkbox-37" name="toctree-checkbox-37" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-37"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_double.html">as_double</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_float.html">as_float</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_int16.html">as_int16</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_int32.html">as_int32</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_int8.html">as_int8</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_type.html">as_type</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_uint16.html">as_uint16</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_uint32.html">as_uint32</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/casting/as_uint8.html">as_uint8</a></li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/window/index.html">Windowing</a><input class="toctree-checkbox" id="toctree-checkbox-38" name="toctree-checkbox-38" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-38"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/window/bartlett.html">bartlett</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/window/blackman.html">blackman</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/window/flattop.html">flattop</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/window/hamming.html">hamming</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/window/hanning.html">hanning</a></li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/signalimage/index.html">Signal and Image Processing</a><input class="toctree-checkbox" id="toctree-checkbox-39" name="toctree-checkbox-39" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-39"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/signalimage/convolution/index.html">Convolution</a><input class="toctree-checkbox" id="toctree-checkbox-40" name="toctree-checkbox-40" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-40"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/convolution/conv1d.html">conv1d</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/convolution/conv2d.html">conv2d</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/signalimage/filtering/index.html">Filtering</a><input class="toctree-checkbox" id="toctree-checkbox-41" name="toctree-checkbox-41" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-41"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/filtering/channelize_poly.html">channelize_poly</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/filtering/filter.html">filter</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/filtering/resample_poly.html">resample_poly</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/signalimage/general/index.html">General</a><input class="toctree-checkbox" id="toctree-checkbox-42" name="toctree-checkbox-42" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-42"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/general/chirp.html">chirp</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/general/downsample.html">downsample</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/general/pwelch.html">pwelch</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/general/upsample.html">upsample</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/signalimage/radar/index.html">Radar</a><input class="toctree-checkbox" id="toctree-checkbox-43" name="toctree-checkbox-43" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-43"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/radar/ambgfun.html">ambgfun</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/signalimage/radar/mvdr.html">MVDR</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/polynomials/index.html">Polynomials</a><input class="toctree-checkbox" id="toctree-checkbox-44" name="toctree-checkbox-44" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-44"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/polynomials/legendre.html">legendre</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="api/random/random.html">Random Number Generation</a></li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/dft/index.html">Discrete Fourier Transform</a><input class="toctree-checkbox" id="toctree-checkbox-45" name="toctree-checkbox-45" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-45"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/dft/dct/index.html">DCT Functions</a><input class="toctree-checkbox" id="toctree-checkbox-46" name="toctree-checkbox-46" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-46"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/dft/dct/dct.html">dct</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/dft/fft/index.html">FFT Functions</a><input class="toctree-checkbox" id="toctree-checkbox-47" name="toctree-checkbox-47" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-47"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/dft/fft/fft.html">fft</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/fft/fft2d.html">fft2</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/fft/ifft.html">ifft</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/fft/ifft2.html">ifft2</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/fft/r2c.html">r2c</a></li>
</ul>
</li>
<li class="toctree-l3 has-children"><a class="reference internal" href="api/dft/utility/index.html">Utility Functions</a><input class="toctree-checkbox" id="toctree-checkbox-48" name="toctree-checkbox-48" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-48"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l4"><a class="reference internal" href="api/dft/utility/fftfreq.html">fftfreq</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/utility/fftshift1d.html">fftshift1D</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/utility/fftshift2d.html">fftshift2D</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/utility/ifftshift1d.html">ifftshift1D</a></li>
<li class="toctree-l4"><a class="reference internal" href="api/dft/utility/ifftshift2d.html">ifftshift2D</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2 has-children"><a class="reference internal" href="api/viz/index.html">Visualization</a><input class="toctree-checkbox" id="toctree-checkbox-49" name="toctree-checkbox-49" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-49"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l3"><a class="reference internal" href="api/viz/bar.html">bar</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/viz/contour.html">contour</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/viz/line.html">line</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/viz/scatter.html">scatter</a></li>
<li class="toctree-l3"><a class="reference internal" href="api/viz/surf.html">surf</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="api/type_traits.html">Type Traits</a></li>
</ul>
</li>
<li class="toctree-l1 has-children"><a class="reference internal" href="examples/index.html">Example Applications</a><input class="toctree-checkbox" id="toctree-checkbox-50" name="toctree-checkbox-50" type="checkbox"/><label class="toctree-toggle" for="toctree-checkbox-50"><i class="fa-solid fa-chevron-down"></i></label><ul>
<li class="toctree-l2"><a class="reference internal" href="examples/fftconv.html">FFT Convolution</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="limitations.html">Limitations</a></li>
</ul>
</div>
</nav></div>
</div>
<div class="sidebar-primary-items__end sidebar-primary__section">
</div>
<div id="rtd-footer-container"></div>
</div>
<main id="main-content" class="bd-main">
<div class="sbt-scroll-pixel-helper"></div>
<div class="bd-content">
<div class="bd-article-container">
<div class="bd-header-article">
<div class="header-article-items header-article__inner">
<div class="header-article-items__start">
<div class="header-article-item"><label class="sidebar-toggle primary-toggle btn btn-sm" for="__primary" title="Toggle primary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
<span class="fa-solid fa-bars"></span>
</label></div>
</div>
<div class="header-article-items__end">
<div class="header-article-item">
<div class="article-header-buttons">
<div class="dropdown dropdown-download-buttons">
<button class="btn dropdown-toggle" type="button" data-bs-toggle="dropdown" aria-expanded="false" aria-label="Download this page">
<i class="fas fa-download"></i>
</button>
<ul class="dropdown-menu">
<li><a href="_sources/quickstart.rst" target="_blank"
class="btn btn-sm btn-download-source-button dropdown-item"
title="Download source file"
data-bs-placement="left" data-bs-toggle="tooltip"
>
<span class="btn__icon-container">
<i class="fas fa-file"></i>
</span>
<span class="btn__text-container">.rst</span>
</a>
</li>
<li>
<button onclick="window.print()"
class="btn btn-sm btn-download-pdf-button dropdown-item"
title="Print to PDF"
data-bs-placement="left" data-bs-toggle="tooltip"
>
<span class="btn__icon-container">
<i class="fas fa-file-pdf"></i>
</span>
<span class="btn__text-container">.pdf</span>
</button>
</li>
</ul>
</div>
<button onclick="toggleFullScreen()"
class="btn btn-sm btn-fullscreen-button"
title="Fullscreen mode"
data-bs-placement="bottom" data-bs-toggle="tooltip"
>
<span class="btn__icon-container">
<i class="fas fa-expand"></i>
</span>
</button>
<script>
document.write(`
<button class="theme-switch-button btn btn-sm btn-outline-primary navbar-btn rounded-circle" title="light/dark" aria-label="light/dark" data-bs-placement="bottom" data-bs-toggle="tooltip">
<span class="theme-switch" data-mode="light"><i class="fa-solid fa-sun"></i></span>
<span class="theme-switch" data-mode="dark"><i class="fa-solid fa-moon"></i></span>
<span class="theme-switch" data-mode="auto"><i class="fa-solid fa-circle-half-stroke"></i></span>
</button>
`);
</script>
<script>
document.write(`
<button class="btn btn-sm navbar-btn search-button search-button__button" title="Search" aria-label="Search" data-bs-placement="bottom" data-bs-toggle="tooltip">
<i class="fa-solid fa-magnifying-glass"></i>
</button>
`);
</script>
<label class="sidebar-toggle secondary-toggle btn btn-sm" for="__secondary"title="Toggle secondary sidebar" data-bs-placement="bottom" data-bs-toggle="tooltip">
<span class="fa-solid fa-list"></span>
</label>
</div></div>
</div>
</div>
</div>
<div id="jb-print-docs-body" class="onlyprint">
<h1>Quick Start</h1>
<!-- Table of contents -->
<div id="print-main-content">
<div id="jb-print-toc">
<div>
<h2> Contents </h2>
</div>
<nav aria-label="Page">
<ul class="visible nav section-nav flex-column">
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#adding-matx-to-your-project">Adding MatX to your project</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#tensor-views">Tensor Views</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#getting-shapes-and-sizes">Getting shapes and sizes</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#slicing-and-dicing">Slicing and dicing</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#permuting">Permuting</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#reshaping">Reshaping</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#increasing-dimensionality">Increasing dimensionality</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#creating-a-view-from-an-existing-pointer">Creating a view from an existing pointer</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#operator-expressions">Operator expressions</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#a-quick-note-about-assignment">A quick note about assignment</a><ul class="nav section-nav flex-column">
<li class="toc-h3 nav-item toc-entry"><a class="reference internal nav-link" href="#initialization-of-operators-and-generators">Initialization of operators and generators</a></li>
</ul>
</li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#transforms">Transforms</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#random-numbers">Random numbers</a></li>
<li class="toc-h2 nav-item toc-entry"><a class="reference internal nav-link" href="#that-s-it">That’s it!</a></li>
</ul>
</nav>
</div>
</div>
</div>
<div id="searchbox"></div>
<article class="bd-article" role="main">
<section id="quick-start">
<span id="quickstart"></span><h1>Quick Start<a class="headerlink" href="#quick-start" title="Permalink to this heading">#</a></h1>
<p>This guide walks through a quick introduction to MatX to get familiar with the types and basic functionality. For more extensive documentation, please
look at the following sources:</p>
<ol class="arabic simple">
<li><p>Example and API documentation</p></li>
<li><p>Example code in the examples/ directory</p></li>
<li><p>Jupyter notebook tutorials in the docs_input/notebooks/ directory</p></li>
</ol>
<section id="adding-matx-to-your-project">
<h2>Adding MatX to your project<a class="headerlink" href="#adding-matx-to-your-project" title="Permalink to this heading">#</a></h2>
<p>MatX is a single header file to include in your project called <code class="docutils literal notranslate"><span class="pre">matx.h</span></code>. Simply add the matx/include directory to your compiler’s
include search path, and add <code class="docutils literal notranslate"><span class="pre">#include</span> <span class="pre">"matx.h"</span></code>. All core MatX functions are in a top-level <code class="docutils literal notranslate"><span class="pre">matx</span></code> namespace, while more specific functions have
a nested namespace. For example, the visualization pieces of MatX are under <code class="docutils literal notranslate"><span class="pre">matx::viz</span></code>.</p>
</section>
<section id="tensor-views">
<h2>Tensor Views<a class="headerlink" href="#tensor-views" title="Permalink to this heading">#</a></h2>
<p>The most common data type in MatX is the tensor (tensor_t). The tensor is used for both viewing and managing any
underlying GPU or host memory. While dynamically-typed languages like Python or MATLAB will implicitly allocate and manage data for the user,
MatX requires either a one-time explicit memory allocation, or various ways of providing your own buffer. This gives more control over the lifetime
of the data, and allows reusing memory regions for different operations.</p>
<p>A tensor is created using the following syntax:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">});</span>
</pre></div>
</div>
<p>In this example we request a floating point tensor with rank 2 (2D array). The constructor arguments specify the shape of the tensor (10x20),
or the size of each dimension. The number of elements in the list determines the rank of the tensor. MatX supports any arbitrary rank tensor, so the
dimensions can be as long as you wish.</p>
<p>If the shape of the tensor is known at compile time, a static tensor can be created for a performance improvement when accessing elements of the
tensor:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_static_tensor</span><span class="o"><</span><span class="kt">float</span><span class="p">,</span><span class="w"> </span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="o">></span><span class="p">();</span>
</pre></div>
</div>
<p>Note that for a static tensor the shape is moved to the template parameters instead of function arguments.</p>
<p>After calling the make function, MatX will allocate CUDA managed memory large enough to accommodate the specified tensor size. Users can also
pass their own pointers in a different for of the <code class="docutils literal notranslate"><span class="pre">make_</span></code> family of functions to allow for more control over buffer types and ownership
semantics.</p>
<p>With our view <code class="docutils literal notranslate"><span class="pre">t</span></code> created above, we now have managed memory allocated sufficiently large to hold our values, but at this point the data
in the tensor is undefined. To set individual values in a view, we can use <code class="docutils literal notranslate"><span class="pre">operator()</span></code>:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="n">t</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span><span class="mi">2</span><span class="p">)</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mf">5.5</span><span class="p">;</span>
<span class="n">t</span><span class="p">(</span><span class="mi">4</span><span class="p">,</span><span class="mi">6</span><span class="p">)</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mi">-10</span><span class="n">f</span><span class="p">;</span>
</pre></div>
</div>
<p>The same operator can be used to get values:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="kt">float</span><span class="w"> </span><span class="n">f</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t</span><span class="p">(</span><span class="mi">2</span><span class="p">,</span><span class="mi">2</span><span class="p">)</span><span class="w"> </span><span class="c1">// f is now 5.5</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">operator()</span></code> takes as many parameters as the rank of the tensor:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">f0</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t0</span><span class="p">();</span><span class="w"> </span><span class="c1">// t0 is a rank-0 tensor (scalar)</span>
<span class="k">auto</span><span class="w"> </span><span class="n">f1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t1</span><span class="p">(</span><span class="mi">5</span><span class="p">);</span><span class="w"> </span><span class="c1">// t1 is a rank-1 tensor (scalar)</span>
</pre></div>
</div>
<p>Tensors can also be initialized using initializer list syntax using the <code class="docutils literal notranslate"><span class="pre">SetVals</span></code> function:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">myT</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">3</span><span class="p">,</span><span class="w"> </span><span class="mi">3</span><span class="p">});</span>
<span class="n">myT</span><span class="p">.</span><span class="n">SetVals</span><span class="p">({{</span><span class="mi">1</span><span class="p">,</span><span class="mi">2</span><span class="p">,</span><span class="mi">3</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">4</span><span class="p">,</span><span class="mi">5</span><span class="p">,</span><span class="mi">6</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">7</span><span class="p">,</span><span class="mi">8</span><span class="p">,</span><span class="mi">9</span><span class="p">}});</span>
</pre></div>
</div>
<p>In other languages it’s very common to initialize a tensor with a set of values on creation (ones, zeros, ranges). This will be covered later
in the tutorial when we discuss operators, and it should become clear why we initialize this way.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>For more information about creating tensors, including advanced usage, see the <a class="reference internal" href="basics/creation.html#creating"><span class="std std-ref">Creating Tensors</span></a> documentation</p>
</div>
</section>
<section id="getting-shapes-and-sizes">
<h2>Getting shapes and sizes<a class="headerlink" href="#getting-shapes-and-sizes" title="Permalink to this heading">#</a></h2>
<p>The dimensions of the tensor are stored internally in a type named tensorShape_t. This tensor shape contains the rank and dimensions of the
tensor view, but does not contain any information about type or storage. The shape can be retrieved using the <code class="docutils literal notranslate"><span class="pre">Shape</span></code> call:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">shape</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t</span><span class="p">.</span><span class="n">Shape</span><span class="p">();</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">Shape()</span></code> is similar to NumPy’s <code class="docutils literal notranslate"><span class="pre">shape</span></code> attribute.</p>
<p>The number of dimensions in a tensor can be retrieved using the <code class="docutils literal notranslate"><span class="pre">Rank()</span></code> member. Since the rank is known at compile time, this function
uses the <code class="docutils literal notranslate"><span class="pre">constexpr</span></code> modifier:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">r</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t</span><span class="p">.</span><span class="n">Rank</span><span class="p">();</span>
</pre></div>
</div>
<p>The size of each individual dimension can be fetched using <code class="docutils literal notranslate"><span class="pre">Size()</span></code>:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t1size</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t1</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">0</span><span class="p">);</span><span class="w"> </span><span class="c1">// Size of vector t1</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t2rows</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t2</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">0</span><span class="p">);</span><span class="w"> </span><span class="c1">// Rows in t2</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t2cols</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t2</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">1</span><span class="p">);</span><span class="w"> </span><span class="c1">// Cols in t2</span>
</pre></div>
</div>
</section>
<section id="slicing-and-dicing">
<h2>Slicing and dicing<a class="headerlink" href="#slicing-and-dicing" title="Permalink to this heading">#</a></h2>
<p>As the name implies, <code class="docutils literal notranslate"><span class="pre">t</span></code> is a view into a region of memory. When the initial view is created and memory is allocated, the tensor view is
of the entire 10x20 contiguous block of memory. Often we don’t want to see the entire block of memory, but only want to view a subset of the
underlying data. To do this, we use the <code class="docutils literal notranslate"><span class="pre">slice</span></code> operator:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">tCube</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">slice</span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">3</span><span class="p">,</span><span class="w"> </span><span class="mi">5</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">6</span><span class="p">,</span><span class="w"> </span><span class="mi">8</span><span class="p">});</span><span class="w"> </span><span class="c1">// Cube of t using rows 3-5 and cols 5-7</span>
<span class="k">auto</span><span class="w"> </span><span class="n">tRectS</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">slice</span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">0</span><span class="p">,</span><span class="w"> </span><span class="mi">0</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="n">matxEnd</span><span class="p">,</span><span class="w"> </span><span class="n">matxEnd</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">2</span><span class="p">,</span><span class="w"> </span><span class="mi">2</span><span class="p">});</span><span class="w"> </span><span class="c1">// Rectangle with stride of 2 in both dimensions</span>
<span class="k">auto</span><span class="w"> </span><span class="n">tCol</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">slice</span><span class="o"><</span><span class="mi">1</span><span class="o">></span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">0</span><span class="p">,</span><span class="w"> </span><span class="mi">4</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="n">matxEnd</span><span class="p">,</span><span class="w"> </span><span class="n">matxDropDim</span><span class="p">});</span><span class="w"> </span><span class="c1">// Create a 1D tensor with only column 5</span>
<span class="k">auto</span><span class="w"> </span><span class="n">tRow</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">slice</span><span class="o"><</span><span class="mi">1</span><span class="o">></span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">4</span><span class="p">,</span><span class="w"> </span><span class="mi">0</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="n">matxDropDim</span><span class="p">,</span><span class="w"> </span><span class="n">matxEnd</span><span class="p">});</span><span class="w"> </span><span class="c1">// Create a 1D tensor with only row 5</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">slice</span></code> returns a new view of the tensor using start, stop, and optional stride parameters. Since views are simply
light-weight views into memory, none of these variants modify the data; they return an object with new parameters describing
how the data is viewed. The resulting variables can be used exactly as the original view above:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">cubeRows</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">tCube</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">0</span><span class="p">);</span><span class="w"> </span><span class="c1">// 3</span>
<span class="k">auto</span><span class="w"> </span><span class="n">cubeCols</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">tCube</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">1</span><span class="p">);</span><span class="w"> </span><span class="c1">// 3</span>
<span class="k">auto</span><span class="w"> </span><span class="n">colSize</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">tCol</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">0</span><span class="p">);</span><span class="w"> </span><span class="c1">// 10 since the original tensor had 10 rows</span>
<span class="k">auto</span><span class="w"> </span><span class="n">rowSize</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">tRow</span><span class="p">.</span><span class="n">Size</span><span class="p">(</span><span class="mi">0</span><span class="p">);</span><span class="w"> </span><span class="c1">// 20 since the original tensor had 20 columns</span>
</pre></div>
</div>
<p>All view functions can be used on any type of existing view:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">tCubeP</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">permute</span><span class="p">(</span><span class="n">slice</span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">3</span><span class="p">,</span><span class="w"> </span><span class="mi">5</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">6</span><span class="p">,</span><span class="w"> </span><span class="mi">8</span><span class="p">}),</span><span class="w"> </span><span class="p">{</span><span class="mi">1</span><span class="p">,</span><span class="w"> </span><span class="mi">0</span><span class="p">});</span>
</pre></div>
</div>
<p>The above code takes the same cube as before, but permutes the cube view by swapping the two dimensions.</p>
<p><code class="docutils literal notranslate"><span class="pre">slice</span></code> is not limited to only tensors; it can be used on any operator as input:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="n">slice</span><span class="p">(</span><span class="n">eye</span><span class="p">(</span><span class="n">t</span><span class="p">.</span><span class="n">Shape</span><span class="p">()),</span><span class="w"> </span><span class="p">{</span><span class="mi">3</span><span class="p">,</span><span class="w"> </span><span class="mi">5</span><span class="p">},</span><span class="w"> </span><span class="p">{</span><span class="mi">6</span><span class="p">,</span><span class="w"> </span><span class="mi">8</span><span class="p">});</span>
</pre></div>
</div>
</section>
<section id="permuting">
<h2>Permuting<a class="headerlink" href="#permuting" title="Permalink to this heading">#</a></h2>
<p>Permuting a tensor is done using the <code class="docutils literal notranslate"><span class="pre">permute</span></code> function:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">tp</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">permute</span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">1</span><span class="p">,</span><span class="mi">0</span><span class="p">});</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">tp</span></code> is now a view into <code class="docutils literal notranslate"><span class="pre">t</span></code> where the rows and columns are swapped (transpose). <code class="docutils literal notranslate"><span class="pre">permute</span></code> is not limited to matrices, though:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t4</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">,</span><span class="w"> </span><span class="mi">5</span><span class="p">,</span><span class="w"> </span><span class="mi">2</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">tp4</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">permute</span><span class="p">(</span><span class="n">t</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">1</span><span class="p">,</span><span class="mi">3</span><span class="p">,</span><span class="mi">2</span><span class="p">,</span><span class="mi">0</span><span class="p">});</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">t4p</span></code> is now a permuted view of the original 4D tensor, but with the dimensions swapped as ordered in the initializer list. Just like
with <code class="docutils literal notranslate"><span class="pre">slice</span></code>, <code class="docutils literal notranslate"><span class="pre">permute</span></code> works on operators as well.</p>
<p>Note that since no data is moved, permuting a tensor can be detrimental to performance, depending on the context. Permuting usually
changes the strides of dimensions such that the memory access patterns are no longer optimal, and accessing the permuted view
continuously can be very slow. If a permuted view will be accessed repeatedly, it’s recommended to copy the permuted view into
a new tensor so that the new layout is contiguous. Using the variables from above:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t4pc</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">(</span><span class="n">tp4</span><span class="p">.</span><span class="n">Shape</span><span class="p">());</span>
<span class="n">copy</span><span class="p">(</span><span class="n">t4pc</span><span class="p">,</span><span class="w"> </span><span class="n">t4p</span><span class="p">);</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">t4pc</span></code> will now contain the permuted data, but in contiguous memory. Copying a tensor (or operator) can also be done by the assignment
operator:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">t4pc</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">t4p</span><span class="p">).</span><span class="n">run</span><span class="p">();</span>
</pre></div>
</div>
</section>
<section id="reshaping">
<h2>Reshaping<a class="headerlink" href="#reshaping" title="Permalink to this heading">#</a></h2>
<p>Ultimately memory is always laid out linearly regardless of how we choose to view it. We can take advantage of this property by allowing
a reshaped view of an existing view. This is commonly done when we want to take a tensor of one rank and view the data as if it were
a tensor of a different rank. The product of dimensions in one rank must equal the product of dimensions in the other rank. For example,
to take a 1D tensor of size 16 and reshape into a 2D tensor of shape 4x4:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">auto</span> <span class="n">t1</span> <span class="o">=</span> <span class="n">make_tensor</span><span class="o"><</span><span class="nb">float</span><span class="o">></span><span class="p">({</span><span class="mi">16</span><span class="p">});</span>
<span class="n">auto</span> <span class="n">t2</span> <span class="o">=</span> <span class="n">t1</span><span class="o">.</span><span class="n">View</span><span class="p">({</span><span class="mi">4</span><span class="p">,</span><span class="mi">4</span><span class="p">});</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">t2</span></code> is now a view into the same memory as <code class="docutils literal notranslate"><span class="pre">t1</span></code>, but viewed as a different rank. Any modifications to one tensor will be seen in the
other since no data was copied.</p>
</section>
<section id="increasing-dimensionality">
<h2>Increasing dimensionality<a class="headerlink" href="#increasing-dimensionality" title="Permalink to this heading">#</a></h2>
<p>Sometimes it’s useful to increase the rank of an existing view to match the dimensions of another tensor. For example, to add a vector onto
all rows in a matrix, you can clone the tensor to a higher rank to match the other tensor:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">int</span><span class="o">></span><span class="p">({</span><span class="mi">16</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">16</span><span class="p">,</span><span class="w"> </span><span class="mi">16</span><span class="p">});</span>
<span class="c1">// ... Initialize tensors</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t1c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">clone</span><span class="o"><</span><span class="mi">2</span><span class="o">></span><span class="p">(</span><span class="n">t1</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">16</span><span class="p">,</span><span class="w"> </span><span class="n">matxKeepDim</span><span class="p">});</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">t1c</span></code> is now a new tensor view where each row is a replica of the tensor <code class="docutils literal notranslate"><span class="pre">t1</span></code>. Again, this is just a view and no data was modified or
allocated, so modifying a row/column in either of these tensors will affect the other.</p>
<p>The keyword <code class="docutils literal notranslate"><span class="pre">matxKeepDim</span></code> tells MatX which dimensions should be kept from the original tensor and where it should be in the new tensor.
In this example we used it in the columns place of the shape, but we also could have used <code class="docutils literal notranslate"><span class="pre">{matxKeepDim,</span> <span class="pre">16}</span></code> and we would have a 2D
view where all columns of <code class="docutils literal notranslate"><span class="pre">t1c</span></code> matches <code class="docutils literal notranslate"><span class="pre">t1</span></code>.</p>
<p>Note in some cases MatX’s <em>broadcasting</em> feature can be used instead of <code class="docutils literal notranslate"><span class="pre">clone</span></code>. This allows an implicit expansion of ranks during an
element-wise operation. For example, adding a 4D tensor to a 1D tensor will work as long as the outer dimension of the 4D tensor matches
that of the 1D tensor. Broadcasting is covered in the documentation. <code class="docutils literal notranslate"><span class="pre">clone</span></code> is much more powerful since it gives more control over which
dimensions are cloned instead of assuming the outer dimensions.</p>
</section>
<section id="creating-a-view-from-an-existing-pointer">
<h2>Creating a view from an existing pointer<a class="headerlink" href="#creating-a-view-from-an-existing-pointer" title="Permalink to this heading">#</a></h2>
<p>While using tensor views with CUDA managed memory is very convenient, there are situations where managed memory is not ideal. Integrating
MatX into an existing codebase, or wanting more control over the memory copies are both times when using standard CUDA memory allocations
is a better option. All constructors in the tensor_t class also allow a manually-allocated pointer to be passed in. MatX will not
attempt to allocate or free any memory when this constructor is used, and it is up to the caller to manage the memory lifecycle:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="kt">float</span><span class="w"> </span><span class="o">*</span><span class="n">my_device_ptr</span><span class="p">;</span><span class="w"> </span><span class="c1">// Assume my_device_ptr is allocated somewhere</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">(</span><span class="n">my_device_ptr</span><span class="p">,</span><span class="w"> </span><span class="p">{</span><span class="mi">20</span><span class="p">,</span><span class="mi">100</span><span class="p">});</span>
<span class="n">t2</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span><span class="mi">1</span><span class="p">)</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="mi">5</span><span class="p">;</span><span class="w"> </span><span class="c1">// Error! Don't do this!</span>
</pre></div>
</div>
<p>In the example above, <code class="docutils literal notranslate"><span class="pre">t2</span></code> is a new view pointing to the existing device-allocated memory. Unlike with managed memory, <code class="docutils literal notranslate"><span class="pre">operator()</span></code>
cannot be used on <code class="docutils literal notranslate"><span class="pre">t2</span></code> from the host side or the code may crash.</p>
</section>
<section id="operator-expressions">
<h2>Operator expressions<a class="headerlink" href="#operator-expressions" title="Permalink to this heading">#</a></h2>
<p>Tensors aren’t much use by themselves if all we can do is view them in various ways. MatX provides two main ways to perform computations on
tensor views: <em>operator expressions</em> and <em>executors</em>.</p>
<p>Operator expressions provide a way to use algebraic expressions using tensor views and operators to generate an element-wise GPU kernel at compile-time.
For example:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">a</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">b</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">});</span>
<span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">a</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">b</span><span class="p">).</span><span class="n">run</span><span class="p">();</span>
</pre></div>
</div>
<p>Ignoring that the data is unitialized, the first three lines simply create three 2D tensors with the same dimensions, while the last line runs an
operator for the equation c = a + b. In MatX terminology, an operator is a type that creates a CUDA kernel at compile-time to perform the
element-wise operation c = a + b. The = operator is used as a deferred assignment operator expressions to avoid ambiguity with the regular assignment
operator <code class="docutils literal notranslate"><span class="pre">=</span></code>. The <code class="docutils literal notranslate"><span class="pre">run</span></code> method takes an optional stream parameter, and executes the operation in the CUDA stream specified. Operators can use
expressions of any length, and normal precedence rules apply.</p>
<p>Tensor views can be mixed with scalars and operator functions:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">op</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">a</span><span class="o">*</span><span class="n">a</span><span class="p">)</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">b</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mf">2.0</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">abs</span><span class="p">(</span><span class="n">a</span><span class="p">));</span>
</pre></div>
</div>
<p>This expression squares each element in <code class="docutils literal notranslate"><span class="pre">a</span></code>, divides each element in <code class="docutils literal notranslate"><span class="pre">b</span></code> by 2, adds the result to <code class="docutils literal notranslate"><span class="pre">a</span></code>, and finally adds the resulting
tensor to the absolute value of every element in <code class="docutils literal notranslate"><span class="pre">a</span></code>. The result of the computation will be stored in the tensor view <code class="docutils literal notranslate"><span class="pre">c</span></code>.
Again, the entire expression is generated at compile time and a kernel is stored in the variable <code class="docutils literal notranslate"><span class="pre">op</span></code>, but the kernel is not launched on the device.
To launch the operator in a CUDA stream, we use the <code class="docutils literal notranslate"><span class="pre">run</span></code> function:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="n">op</span><span class="p">.</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p><code class="docutils literal notranslate"><span class="pre">run</span></code> can be thought of as a way to launch the operator/kernel into a CUDA stream, similar to the traditional triple angle bracket notation (<<<>>>).
In MatX terminology, this is called an executor since it causes work to be executed on the device. It’s often not necessary to store the operator at
all if the execution is immediate, the two lines above can be combined:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">a</span><span class="o">*</span><span class="n">a</span><span class="p">)</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">b</span><span class="w"> </span><span class="o">/</span><span class="w"> </span><span class="mf">2.0</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">abs</span><span class="p">(</span><span class="n">a</span><span class="p">)).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>Sometimes the data we are using in an expression can be generated on-the-fly rather than coming from memory. Window functions, diagonal matrices, and
the identity matrix are all examples of this. MatX provides “generators” that can be used inside of expressions to generate data:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">a</span><span class="o">*</span><span class="n">a</span><span class="p">)</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">ones</span><span class="p">(</span><span class="n">a</span><span class="p">.</span><span class="n">Shape</span><span class="p">())).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>The example above uses the <code class="docutils literal notranslate"><span class="pre">ones</span></code> generator to create a tensor with only the value <code class="docutils literal notranslate"><span class="pre">1</span></code> matching the shape of a (10x20). <code class="docutils literal notranslate"><span class="pre">ones</span></code> simply returns the
value <code class="docutils literal notranslate"><span class="pre">1</span></code> any time an element of it is requested, and no data is ever loaded from memory.</p>
<p>Implicit in the <code class="docutils literal notranslate"><span class="pre">run</span></code> call above is a CUDA executor type. As a beta feature, MatX also supports executing code on the host using a different executor.
To run the same code on the host, a <code class="docutils literal notranslate"><span class="pre">SingleThreadHostExecutor</span></code> can be passed into <code class="docutils literal notranslate"><span class="pre">run</span></code>:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">a</span><span class="o">*</span><span class="n">a</span><span class="p">)</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">ones</span><span class="p">(</span><span class="n">a</span><span class="p">.</span><span class="n">Shape</span><span class="p">())).</span><span class="n">run</span><span class="p">(</span><span class="n">SingleThreadHostExecutor</span><span class="p">{});</span>
</pre></div>
</div>
<p>Instead of a CUDA stream, we pass an executor to <code class="docutils literal notranslate"><span class="pre">run</span></code> that instructs MatX to execute the code on the host instead of the device using a single CPU thread.
Unlike CUDA calls, host executors are synchronous, and the line above will block until finished executing.</p>
</section>
<section id="a-quick-note-about-assignment">
<h2>A quick note about assignment<a class="headerlink" href="#a-quick-note-about-assignment" title="Permalink to this heading">#</a></h2>
<p>MatX heavily relies on a deferred or lazy execution model where expressions are not executed at the time of assignment. This allows the library to
closely match the programming model of the GPU so that there are no surprises as to when code is executed. To facilitate the asynchronous model,
MatX overloads the assignment operator (=) to indicate a deferred execution. The deferred assignment can be executed using the <code class="docutils literal notranslate"><span class="pre">run()</span></code> method on
the expression. A statement as simple as the following:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">A</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">B</span><span class="p">).</span><span class="n">run</span><span class="p">()</span>
</pre></div>
</div>
<p>should be viewed as a deferred assignment of tensor B into tensor A (deep copy) that executes on the device when <code class="docutils literal notranslate"><span class="pre">run()</span></code> happens. The result of the
lazy assignment expression can also be assigned into a temporary variable:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">op</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="p">(</span><span class="n">A</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">B</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">C</span><span class="p">);</span>
</pre></div>
</div>
<p>In the code above, the <code class="docutils literal notranslate"><span class="pre">=</span></code> on the right side indicates lazy assignment, while the <code class="docutils literal notranslate"><span class="pre">=</span></code> on the left side executes the copy constructor on the new
variable <code class="docutils literal notranslate"><span class="pre">op</span></code>. The pattern above is expected to be infrequently used since expressions are typically executed on the same line as the definition,
but sometimes it’s useful for debugging purposes to look at the type of the expression. More complex expressions follow the same rules:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="n">IFELSE</span><span class="p">(</span><span class="n">A</span><span class="w"> </span><span class="o">></span><span class="w"> </span><span class="mi">5</span><span class="p">,</span><span class="w"> </span><span class="n">B</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">A</span><span class="p">,</span><span class="w"> </span><span class="n">C</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">B</span><span class="p">).</span><span class="n">run</span><span class="p">()</span>
</pre></div>
</div>
<p>Remember that since the assignment operator is deferred in both cases above, none of these assignments will happen until <code class="docutils literal notranslate"><span class="pre">A</span> <span class="pre">></span> <span class="pre">5</span></code> is executed on the
device, at which point only <em>one</em> of these assignments will occur.</p>
<section id="initialization-of-operators-and-generators">
<h3>Initialization of operators and generators<a class="headerlink" href="#initialization-of-operators-and-generators" title="Permalink to this heading">#</a></h3>
<p>As mentioned above, it’s common in high-level languages to initialize a tensor/array with a known set of values. For example, generating a range of linearly-
spaced values, all ones, or a diagonal matrix. These are all operations that do not need to be generated and stored in memory before using since they are
all generated from a formula. MatX calls these types of operators a <em>generator</em>, indicating that they generate data without storage.</p>
<p>Similar to high-level languages, generators can store their values in existing tensors like so:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">100</span><span class="p">});</span>
<span class="p">(</span><span class="n">t1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">linspace_x</span><span class="p">(</span><span class="n">t1</span><span class="p">.</span><span class="n">Shape</span><span class="p">(),</span><span class="w"> </span><span class="mf">1.0f</span><span class="p">,</span><span class="w"> </span><span class="mf">100.0f</span><span class="p">)).</span><span class="n">run</span><span class="p">();</span>
</pre></div>
</div>
<p>Similar to the <code class="docutils literal notranslate"><span class="pre">set</span></code> calls above, instead of an algebraic equation we are storing the output of generator <code class="docutils literal notranslate"><span class="pre">linspace_x</span></code> into the tensor <code class="docutils literal notranslate"><span class="pre">t1</span></code>.
<code class="docutils literal notranslate"><span class="pre">linspace_x</span></code> takes 3 parameters: the shape of the tensor (in this case we match t1), the start value, and the stop value. Since there are 100 elements
in our tensor, it will generate a sequence of 1.0, 2.0, 3.0, etc, and store it in <code class="docutils literal notranslate"><span class="pre">t1</span></code>.</p>
<p>Why not just make a shorthand version of <code class="docutils literal notranslate"><span class="pre">linspace_x</span></code> that stores directly in a tensor? The reason is that generators can be used as part of a larger
expression and are not limited to simply assigning to a tensor. Expanding on our last example:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">100</span><span class="p">});</span>
<span class="p">(</span><span class="n">t1</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">ones</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">(</span><span class="n">t1</span><span class="p">.</span><span class="n">Shape</span><span class="p">())</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">linspace_x</span><span class="p">(</span><span class="n">t1</span><span class="p">.</span><span class="n">Shape</span><span class="p">(),</span><span class="w"> </span><span class="mf">1.0f</span><span class="p">,</span><span class="w"> </span><span class="mf">100.0f</span><span class="p">)</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="mf">5.0</span><span class="p">).</span><span class="n">run</span><span class="p">();</span>
</pre></div>
</div>
<p>Instead of setting <code class="docutils literal notranslate"><span class="pre">t1</span></code> to a range, we multiply the range by 5.0, and add that range to a vector of ones using the <code class="docutils literal notranslate"><span class="pre">ones</span></code> generator. Without any
intermediate storage, we combined two generators, a multiply, and an add operator into a single kernel.</p>
</section>
</section>
<section id="transforms">
<h2>Transforms<a class="headerlink" href="#transforms" title="Permalink to this heading">#</a></h2>
<p>As mentioned above, the <code class="docutils literal notranslate"><span class="pre">run</span></code> function takes an executor describing where to launch the work. In the examples above <code class="docutils literal notranslate"><span class="pre">run</span></code> the operator
expressions created a single fused element-wise operation. Often the type of operation we are trying to do cannot be expressed as
an element-wise operator and therefor can’t be fused with other operations without synchronization. These classes of operators are called <em>transforms</em>.
Transforms can be used anywhere operators are used:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">B</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">fft</span><span class="p">(</span><span class="n">A</span><span class="p">)</span><span class="w"> </span><span class="o">*</span><span class="w"> </span><span class="n">C</span><span class="p">).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>The <code class="docutils literal notranslate"><span class="pre">fft</span></code> transform above performs a 1D FFT on the tensor <code class="docutils literal notranslate"><span class="pre">A</span></code>, multiplies the output by <code class="docutils literal notranslate"><span class="pre">C</span></code>, and stores it in <code class="docutils literal notranslate"><span class="pre">B</span></code>. Since the FFT
may require synchronizing before performing the multiply, MatX can internally create a temporary buffer for the FFT output and free it when
the expression goes out of scope.</p>
<p>Unless documented otherwise, transforms work on tensors of a specific size. Matrix multiplies require a 2D tensor (matrix), 1D FFTs require
a 1D tensor (vector), etc. If the dimension of the tensor is higher than the expected dimension, all higher dimensions will be batched. In the FFT
call above, if <code class="docutils literal notranslate"><span class="pre">A</span></code> and <code class="docutils literal notranslate"><span class="pre">B</span></code> are 4D tensors, the inner 3 dimensions will launch a batched 1D FFT with no change in syntax.</p>
<p>As mentioned above, the same tensor views can be used in operator expressions before or after transforms:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">a</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">b</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="mi">2</span><span class="p">).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
<span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">matmul</span><span class="p">(</span><span class="n">a</span><span class="p">,</span><span class="w"> </span><span class="n">d</span><span class="p">)).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>Or fused in a single line:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">c</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">matmul</span><span class="p">(</span><span class="n">b</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="mi">2</span><span class="p">,</span><span class="w"> </span><span class="n">d</span><span class="p">)).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>The code above executes a kernel to store the result of <code class="docutils literal notranslate"><span class="pre">b</span> <span class="pre">+</span> <span class="pre">2</span></code> into <code class="docutils literal notranslate"><span class="pre">a</span></code>, then subsequently performs the matrix multiply <code class="docutils literal notranslate"><span class="pre">C</span> <span class="pre">=</span> <span class="pre">A</span> <span class="pre">*</span> <span class="pre">B</span></code>. Since
the operator and matrix multiply are launched in the same CUDA stream, they will be executed serially.</p>
<p>Common reduction executors are also available, such as <code class="docutils literal notranslate"><span class="pre">sum()</span></code>, <code class="docutils literal notranslate"><span class="pre">mean()</span></code>, <code class="docutils literal notranslate"><span class="pre">max()</span></code>, etc:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t4</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">100</span><span class="p">,</span><span class="w"> </span><span class="mi">100</span><span class="p">,</span><span class="w"> </span><span class="mi">100</span><span class="p">,</span><span class="w"> </span><span class="mi">100</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">t0</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">();</span>
<span class="p">(</span><span class="n">t0</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">sum</span><span class="p">(</span><span class="n">t4</span><span class="p">)).</span><span class="n">run</span><span class="p">();</span>
</pre></div>
</div>
<p>The above code performs an optimized sum reduction of <code class="docutils literal notranslate"><span class="pre">t4</span></code> into <code class="docutils literal notranslate"><span class="pre">t0</span></code>. Currently reduction type exectors <em>can</em> take operators as an input. Please
see the documentation for a list of which ones are compatible.</p>
<p>For more information about operation fusion, see <a class="reference internal" href="basics/fusion.html#fusion"><span class="std std-ref">Operator Fusion</span></a>.</p>
</section>
<section id="random-numbers">
<h2>Random numbers<a class="headerlink" href="#random-numbers" title="Permalink to this heading">#</a></h2>
<p>MatX can generate random numbers using the cuRAND library as the backend. Random number generation consumes memory on the device, so the construction
is slightly different than other types above:</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="k">auto</span><span class="w"> </span><span class="n">t2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">make_tensor</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">({</span><span class="mi">100</span><span class="p">,</span><span class="w"> </span><span class="mi">50</span><span class="p">});</span>
<span class="k">auto</span><span class="w"> </span><span class="n">randOp</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">random</span><span class="o"><</span><span class="kt">float</span><span class="o">></span><span class="p">(</span><span class="n">t</span><span class="p">.</span><span class="n">Shape</span><span class="p">(),</span><span class="w"> </span><span class="n">NORMAL</span><span class="p">);</span>
</pre></div>
</div>
<p>The code above creates a 100x50 2D tensor, followed by a random operator that produces normally-distributed numbers with the same shape as <code class="docutils literal notranslate"><span class="pre">t2</span></code>.</p>
<p>Using the random operator above uses the same assignment as with any operator, and when the values are fetched on the device a new random number
will be generated for each element.</p>
<div class="highlight-cpp notranslate"><div class="highlight"><pre><span></span><span class="p">(</span><span class="n">t2</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">randOp</span><span class="o">*</span><span class="mi">5</span><span class="w"> </span><span class="o">+</span><span class="w"> </span><span class="n">randOp</span><span class="p">).</span><span class="n">run</span><span class="p">(</span><span class="n">stream</span><span class="p">);</span>
</pre></div>
</div>
<p>In the example above <code class="docutils literal notranslate"><span class="pre">randOp</span></code> is accessed twice. On each access a new random number is generated.</p>
</section>
<section id="that-s-it">
<h2>That’s it!<a class="headerlink" href="#that-s-it" title="Permalink to this heading">#</a></h2>
<p>This quick start guide was intended to give a very brief introduction to the concepts behind MatX, and how these concepts apply to the code. There’s a lot
more to explore in MatX and far more functions than could be listed here. For more examples we recommend browsing through the examples to see how to perform
real tasks using MatX, and the API guide to see an exhaustive list of functions and operators.</p>
</section>
</section>
</article>
<footer class="bd-footer-article">
<div class="footer-article-items footer-article__inner">
<div class="footer-article-item"><!-- Previous / next buttons -->
<div class="prev-next-area">
<a class="left-prev"
href="index.html"
title="previous page">
<i class="fa-solid fa-angle-left"></i>
<div class="prev-next-info">