This file is indexed.

/usr/share/doc/python-pyopencl-doc/html/misc.html is in python-pyopencl-doc 2017.2.2-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  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
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
    <title>Installation &#8212; PyOpenCL 2017.2.2 documentation</title>
    <link rel="stylesheet" href="_static/alabaster.css" type="text/css" />
    <link rel="stylesheet" href="_static/pygments.css" type="text/css" />
    <script type="text/javascript">
      var DOCUMENTATION_OPTIONS = {
        URL_ROOT:    './',
        VERSION:     '2017.2.2',
        COLLAPSE_INDEX: false,
        FILE_SUFFIX: '.html',
        HAS_SOURCE:  false,
        SOURCELINK_SUFFIX: '.txt'
      };
    </script>
    <script type="text/javascript" src="_static/jquery.js"></script>
    <script type="text/javascript" src="_static/underscore.js"></script>
    <script type="text/javascript" src="_static/doctools.js"></script>
    <link rel="index" title="Index" href="genindex.html" />
    <link rel="search" title="Search" href="search.html" />
    <link rel="prev" title="Built-in Utilities" href="tools.html" />
   
  <link rel="stylesheet" href="_static/custom.css" type="text/css" />
  
  <meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />

  </head>
  <body>
  

    <div class="document">
      <div class="documentwrapper">
        <div class="bodywrapper">
          <div class="body" role="main">
            
  <div class="section" id="installation">
<h1>Installation<a class="headerlink" href="#installation" title="Permalink to this headline"></a></h1>
<p>By far the easiest way to install PyOpenCL is to use the packages available in
<a class="reference external" href="https://conda-forge.org/">Conda Forge</a>. Conda Forge is a repository of
community-maintained packages for the <a class="reference external" href="https://conda.io/docs/">Conda</a>
package manager.</p>
<p>On Linux and OS X, the following set of instructions should work:</p>
<ol class="arabic">
<li><p class="first">Install a version of <a class="reference external" href="https://conda.io/miniconda.html">miniconda</a>
that fits your system. Both Python 2 and Python 3 work.
You can install these pieces of software in your user account and
do not need root/administrator privileges.</p>
<p>Note that if you already have Continuum Anaconda installed on your system,
you may just use that and do <em>not</em> need to install Miniconda.</p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">source</span> <span class="pre">/WHERE/YOU/INSTALLED/MINICONDA/bin/activate</span> <span class="pre">root</span></code></p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">conda</span> <span class="pre">config</span> <span class="pre">--add</span> <span class="pre">channels</span> <span class="pre">conda-forge</span></code></p>
</li>
<li><p class="first"><code class="docutils literal"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">pyopencl</span></code></p>
</li>
</ol>
<p>The analogous steps for Windows should also work.</p>
<p>Note that PyOpenCL is no fun (i.e. cannot run code) without an OpenCL device
driver (a so-called “ICD”, for “installable client driver”) that provides access
to hardware through OpenCL.  If you get an error message like
<code class="docutils literal"><span class="pre">pyopencl.cffi_cl.LogicError:</span> <span class="pre">clGetPlatformIDs</span> <span class="pre">failed:</span> <span class="pre">&lt;unknown</span> <span class="pre">error</span>
<span class="pre">-1001&gt;</span></code>, that means you have no OpenCL drivers installed.</p>
<p>Note that drivers (ICDs) are separate pieces of software from PyOpenCL.  They
might be provided by your hardware vendor (e.g. for Nvidia or AMD GPUs). If
you have such hardware, see below for instructions on how to make those work
with PyOpenCL from Conda Forge.</p>
<p>It is important to note that OpenCL is not restricted to GPUs. In fact, no special
hardware is required to use OpenCL for computation–your existing CPU is enough.
On Linux, type:</p>
<ol class="arabic simple">
<li><code class="docutils literal"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">pocl</span></code></li>
</ol>
<p>to install a CPU-based OpenCL driver. On Windows, you may install e.g.
the <a class="reference external" href="https://software.intel.com/en-us/articles/opencl-drivers#latest_CPU_runtime">CPU OpenCL driver from Intel</a>.
OS X has support for OpenCL built into the operating system and does not need
additional software to run code based on PyOpenCL (but see below).</p>
<p>You are now ready to run code based on PyOpenCL, such as the <a class="reference external" href="https://github.com/inducer/pyopencl/tree/master/examples">code
examples</a>.</p>
<div class="section" id="using-vendor-supplied-opencl-drivers-linux">
<h2>Using vendor-supplied OpenCL drivers (Linux)<a class="headerlink" href="#using-vendor-supplied-opencl-drivers-linux" title="Permalink to this headline"></a></h2>
<p>The instructions above help you get a basic OpenCL environment going that
will work independently of whether you have specialized hardware (such as GPUs
or FPGAs) available. If you <em>do</em> have such hardware, read on for how to make
it work.</p>
<p>On Linux, PyOpenCL finds which drivers are installed by looking for files with
the extension <code class="docutils literal"><span class="pre">.icd</span></code> in a directory. PyOpenCL as installed from Conda will
look for these files in
<code class="file docutils literal"><span class="pre">/WHERE/YOU/INSTALLED/MINICONDA/etc/OpenCL/vendors</span></code>.  They are just
simple text files containing either just  the file names or the fully
qualified path names of the shared library providing the OpenCL driver.</p>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p>If you ran the commands above in a
<a class="reference external" href="https://conda.io/docs/user-guide/tasks/manage-environments.html">Conda environment</a>
(i.e. if the environment indicator on your command line prompt says anything other
than <code class="docutils literal"><span class="pre">(root)</span></code>), then you may need to use a path like the following instead:</p>
<p><code class="file docutils literal"><span class="pre">/WHERE/YOU/INSTALLED/MINICONDA/envs/ENVIRONMENTNAME/etc/OpenCL/vendors</span></code></p>
<p class="last">Note that you should replace <code class="docutils literal"><span class="pre">ENVIRONMENTNAME</span></code> with the name of your environment,
shown between parentheses on your command line prompt.</p>
</div>
<p>If you have other OpenCL drivers installed (such as for your GPU), those will be
in <code class="file docutils literal"><span class="pre">/etc/OpenCL/vendors</span></code>. You can make them work with PyOpenCL from Conda Forge
by simply copying them to the above folder.</p>
<p>If you are looking for more information, see <a class="reference external" href="https://github.com/OCL-dev/ocl-icd">ocl-icd</a> and its documentation. Ocl-icd is the
“ICD loader” used by PyOpenCL when installed from Conda Forge. It represents the
code behind <code class="file docutils literal"><span class="pre">libOpenCL.so</span></code>.</p>
</div>
<div class="section" id="getting-a-better-cpu-based-opencl-driver-os-x">
<h2>Getting a better CPU-based OpenCL driver (OS X)<a class="headerlink" href="#getting-a-better-cpu-based-opencl-driver-os-x" title="Permalink to this headline"></a></h2>
<p>OS X has support for both CPU- and GPU-based OpenCL built in. Unfortunately,
the built-in drivers can be temperamental, and they have not advanced as quickly
as one might like. To make PyOpenCL use a more up-to-date (and open-source)
CPU-based OpenCL driver, type the following:</p>
<p><code class="docutils literal"><span class="pre">conda</span> <span class="pre">install</span> <span class="pre">osx-pocl-opencl</span> <span class="pre">pocl</span></code> (OS X)</p>
<p>Note that, by installing <code class="docutils literal"><span class="pre">osx-pocl-opencl</span></code>, you will no longer be able to
use PyOpenCL to talk to the system-wide Apple OpenCL drivers. To regain access
to those drivers, simply uninstall <code class="docutils literal"><span class="pre">osx-pocl-opencl</span></code> and reinstall <code class="docutils literal"><span class="pre">pyopencl</span></code>
afterwards.</p>
</div>
<div class="section" id="installing-from-source">
<h2>Installing from source<a class="headerlink" href="#installing-from-source" title="Permalink to this headline"></a></h2>
<p>Information on how to install PyOpenCL <em>from source</em> is maintained collaboratively on the
<a class="reference external" href="http://wiki.tiker.net/PyOpenCL/Installation">PyOpenCL Wiki</a>, but that should
mostly not be necessary unless you have very specific needs or would like to modify
PyOpenCL yourself.</p>
</div>
</div>
<div class="section" id="tips">
<h1>Tips<a class="headerlink" href="#tips" title="Permalink to this headline"></a></h1>
<div class="section" id="syntax-highlighting">
<h2>Syntax highlighting<a class="headerlink" href="#syntax-highlighting" title="Permalink to this headline"></a></h2>
<p>You can obtain Vim syntax highlighting for OpenCL C inlined in Python by
checking <a class="reference external" href="https://github.com/inducer/pyopencl/blob/master/contrib/pyopencl.vim">this file</a>.</p>
<p>Note that the triple-quoted strings containing the source must start with
<cite>“”“//CL// …”“”</cite>.</p>
</div>
<div class="section" id="ipython-integration">
<h2>IPython integration<a class="headerlink" href="#ipython-integration" title="Permalink to this headline"></a></h2>
<p>PyOpenCL comes with IPython integration, which lets you seamlessly integrate
PyOpenCL kernels into your IPython notebooks. Simply load the PyOpenCL
IPython extension using:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="o">%</span><span class="n">load_ext</span> <span class="n">pyopencl</span><span class="o">.</span><span class="n">ipython_ext</span>
</pre></div>
</div>
<p>and then use the <code class="docutils literal"><span class="pre">%%cl_kernel</span></code> ‘cell-magic’ command. See <a class="reference external" href="http://nbviewer.ipython.org/urls/raw.githubusercontent.com/inducer/pyopencl/master/examples/ipython-demo.ipynb">this notebook</a>
(which ships with PyOpenCL) for a demonstration.</p>
<p>You can pass build options to be used for building the program executable by using the <code class="docutils literal"><span class="pre">-o</span></code> flag on the first line of the cell (next to the <code class="docutils literal"><span class="pre">%%cl_kernel</span></code> directive). For example: <cite>%%cl_kernel -o “-cl-fast-relaxed-math”`</cite>.</p>
<p>There are also line magics: <code class="docutils literal"><span class="pre">cl_load_edit_kernel</span></code> which will load a file into the next cell (adding <code class="docutils literal"><span class="pre">cl_kernel</span></code> to the first line) and <code class="docutils literal"><span class="pre">cl_kernel_from_file</span></code> which will compile kernels from a file (as if you copy-and-pasted the contents of the file to a cell with <code class="docutils literal"><span class="pre">cl_kernel</span></code>). Both of these magics take options <code class="docutils literal"><span class="pre">-f</span></code> to specify the file and optionally <code class="docutils literal"><span class="pre">-o</span></code> for build options.</p>
<div class="versionadded">
<p><span class="versionmodified">New in version 2014.1.</span></p>
</div>
</div>
</div>
<div class="section" id="guidelines">
<h1>Guidelines<a class="headerlink" href="#guidelines" title="Permalink to this headline"></a></h1>
<div class="section" id="api-stability">
<span id="api-compatibility"></span><h2>API Stability<a class="headerlink" href="#api-stability" title="Permalink to this headline"></a></h2>
<p>I consider PyOpenCL’s API “stable”.  That doesn’t mean it can’t
change. But if it does, your code will generally continue to run. It
may however start spewing warnings about things you need to change to
stay compatible with future versions.</p>
<p>Deprecation warnings will be around for a whole year, as identified by the
first number in the release name.  (the “2014” in “2014.1”) I.e. a function
that was deprecated in 2014.n will generally be removed in 2015.n (or perhaps
later). Further, the stability promise applies for any code that’s part of a
released version. It doesn’t apply to undocumented bits of the API, and it
doesn’t apply to unreleased code downloaded from git.</p>
</div>
<div class="section" id="relation-with-opencl-s-c-bindings">
<span id="versus-c"></span><h2>Relation with OpenCL’s C Bindings<a class="headerlink" href="#relation-with-opencl-s-c-bindings" title="Permalink to this headline"></a></h2>
<p>We’ve tried to follow these guidelines when binding the OpenCL’s
C interface to Python:</p>
<ul class="simple">
<li>Remove the <cite>cl_</cite>, <cite>CL_</cite> and <cite>cl</cite> prefix from data types, macros and
function names.</li>
<li>Follow <span class="target" id="index-0"></span><a class="pep reference external" href="https://www.python.org/dev/peps/pep-0008"><strong>PEP 8</strong></a>, i.e.<ul>
<li>Make function names lowercase.</li>
<li>If a data type or function name is composed of more than one word,
separate the words with a single underscore.</li>
</ul>
</li>
<li><cite>get_info</cite> functions become attributes.</li>
<li>Object creation is done by constructors, to the extent possible.
(i.e. minimize use of “factory functions”)</li>
<li>If an operation involves two or more “complex” objects (like e.g. a
kernel enqueue involves a kernel and a queue), refuse the temptation
to guess which one should get a method for the operation.
Instead, simply leave that command to be a function.</li>
</ul>
</div>
<div class="section" id="interoperability-with-other-opencl-software">
<span id="interoperability"></span><h2>Interoperability with other OpenCL software<a class="headerlink" href="#interoperability-with-other-opencl-software" title="Permalink to this headline"></a></h2>
<p>Just about every object in <a class="reference internal" href="runtime.html#module-pyopencl" title="pyopencl"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl</span></code></a> supports the following
interface (here shown as an example for <a class="reference internal" href="runtime_memory.html#pyopencl.MemoryObject" title="pyopencl.MemoryObject"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.MemoryObject</span></code></a>,
from which <a class="reference internal" href="runtime_memory.html#pyopencl.Buffer" title="pyopencl.Buffer"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Buffer</span></code></a> and <code class="xref py py-class docutils literal"><span class="pre">pyopencl.Image</span></code> inherit):</p>
<ul class="simple">
<li><a class="reference internal" href="runtime_memory.html#pyopencl.MemoryObject.from_int_ptr" title="pyopencl.MemoryObject.from_int_ptr"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.MemoryObject.from_int_ptr()</span></code></a></li>
<li><a class="reference internal" href="runtime_memory.html#pyopencl.MemoryObject.int_ptr" title="pyopencl.MemoryObject.int_ptr"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.MemoryObject.int_ptr</span></code></a></li>
</ul>
<p>This allows retrieving the C-level pointer to an OpenCL object as a Python
integer, which may then be passed to other C libraries whose interfaces expose
OpenCL objects. It also allows turning C-level OpenCL objects obtained from
other software to be turned into the corresponding <a class="reference internal" href="runtime.html#module-pyopencl" title="pyopencl"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl</span></code></a> objects.</p>
<div class="versionadded">
<p><span class="versionmodified">New in version 2013.2.</span></p>
</div>
</div>
</div>
<div class="section" id="user-visible-changes">
<h1>User-visible Changes<a class="headerlink" href="#user-visible-changes" title="Permalink to this headline"></a></h1>
<div class="section" id="version-2016-3">
<h2>Version 2016.3<a class="headerlink" href="#version-2016-3" title="Permalink to this headline"></a></h2>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p class="last">This version is currently under development. You can get snapshots from
PyOpenCL’s <a class="reference external" href="https://github.com/inducer/pyopencl">git repository</a></p>
</div>
</div>
<div class="section" id="version-2016-2">
<h2>Version 2016.2<a class="headerlink" href="#version-2016-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Deprecate RANLUXCL. It will be removed in the 2018.x series of PyOpenCL.</li>
<li>Introduce Random123 random number generators. See <a class="reference internal" href="array.html#module-pyopencl.clrandom" title="pyopencl.clrandom"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.clrandom</span></code></a>
for more information.</li>
<li>Add support for <strong>range</strong> and <strong>slice</strong> kwargs and data-less reductions
to <a class="reference internal" href="algorithm.html#pyopencl.reduction.ReductionKernel" title="pyopencl.reduction.ReductionKernel"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.reduction.ReductionKernel</span></code></a>.</li>
<li>Add support for SPIR-V. (See <a class="reference internal" href="runtime_program.html#pyopencl.Program" title="pyopencl.Program"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Program</span></code></a>.)</li>
<li>Add support for <a class="reference internal" href="runtime_memory.html#svm"><span class="std std-ref">Shared Virtual Memory (SVM)</span></a>.</li>
<li><a class="reference internal" href="runtime_memory.html#pyopencl.MemoryMap" title="pyopencl.MemoryMap"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.MemoryMap</span></code></a> is usable as a context manager.</li>
</ul>
</div>
<div class="section" id="version-2016-1">
<h2>Version 2016.1<a class="headerlink" href="#version-2016-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>The <code class="docutils literal"><span class="pre">from_int_ptr</span></code> methods now take a <em>retain</em> parameter for more convenient
ownership management.</li>
<li>Kernel build options (if passed as a list) are now properly quoted.
(This is a potentially compatibility-breaking change.)</li>
<li>Many bug fixes. (GL interop, Windows, event callbacks and more)</li>
</ul>
</div>
<div class="section" id="version-2015-2-4">
<h2>Version 2015.2.4<a class="headerlink" href="#version-2015-2-4" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix building on Windows, using mingwpy and VS 2015.</li>
</ul>
</div>
<div class="section" id="version-2015-2-3">
<h2>Version 2015.2.3<a class="headerlink" href="#version-2015-2-3" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix one more Ubuntu 14.x build issue.</li>
</ul>
</div>
<div class="section" id="version-2015-2-2">
<h2>Version 2015.2.2<a class="headerlink" href="#version-2015-2-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix compatibility with CL 1.1</li>
<li>Fix compatibility with Ubuntu 14.x.</li>
<li>Various bug fixes</li>
</ul>
</div>
<div class="section" id="version-2015-2-1">
<h2>Version 2015.2.1<a class="headerlink" href="#version-2015-2-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix global_offset kernel launch parameter</li>
</ul>
</div>
<div class="section" id="version-2015-2">
<h2>Version 2015.2<a class="headerlink" href="#version-2015-2" title="Permalink to this headline"></a></h2>
<ul>
<li><p class="first"><strong>[INCOMPATIBLE]</strong> Changed PyOpenCL’s complex numbers from <code class="docutils literal"><span class="pre">float2</span></code> and
<code class="docutils literal"><span class="pre">double2</span></code> OpenCL vector types to custom <code class="docutils literal"><span class="pre">struct</span></code>. This was changed
because it very easily introduced bugs where</p>
<ul class="simple">
<li>complex*complex</li>
<li>real+complex</li>
</ul>
<p><em>look</em> like they may do the right thing, but silently do the wrong thing.</p>
</li>
<li><p class="first">Rewrite of the wrapper layer to be based on CFFI</p>
</li>
<li><p class="first">Pypy compatibility</p>
</li>
<li><p class="first">Faster kernel invocation through Python launcher code generation</p>
</li>
<li><p class="first">POCL compatibility</p>
</li>
</ul>
</div>
<div class="section" id="version-2015-1">
<h2>Version 2015.1<a class="headerlink" href="#version-2015-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Support for new-style buffer protocol</li>
<li>Numerous fixes</li>
</ul>
</div>
<div class="section" id="version-2014-1">
<h2>Version 2014.1<a class="headerlink" href="#version-2014-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li><span class="xref std std-ref">ipython-integration</span></li>
<li>Bug fixes</li>
</ul>
</div>
<div class="section" id="version-2013-2">
<h2>Version 2013.2<a class="headerlink" href="#version-2013-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add <a class="reference internal" href="array.html#pyopencl.array.Array.map_to_host" title="pyopencl.array.Array.map_to_host"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array.map_to_host()</span></code></a>.</li>
<li>Support <em>strides</em> on <a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_map_buffer" title="pyopencl.enqueue_map_buffer"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_map_buffer()</span></code></a> and
<a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_map_image" title="pyopencl.enqueue_map_image"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_map_image()</span></code></a>.</li>
<li><a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat" title="pyopencl.ImageFormat"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.ImageFormat</span></code></a> was made comparable and hashable.</li>
<li><a class="reference internal" href="algorithm.html#module-pyopencl.reduction" title="pyopencl.reduction"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.reduction</span></code></a> supports slicing (contributed by Alex Nitz)</li>
<li>Added <a class="reference internal" href="#interoperability"><span class="std std-ref">Interoperability with other OpenCL software</span></a></li>
<li>Bug fixes</li>
</ul>
</div>
<div class="section" id="version-2013-1">
<h2>Version 2013.1<a class="headerlink" href="#version-2013-1" title="Permalink to this headline"></a></h2>
<ul>
<li><p class="first">Vastly improved <a class="reference internal" href="algorithm.html#custom-scan"><span class="std std-ref">Prefix Sums (“scan”)</span></a>.</p>
</li>
<li><p class="first">Add <a class="reference internal" href="array.html#pyopencl.tools.match_dtype_to_c_struct" title="pyopencl.tools.match_dtype_to_c_struct"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.tools.match_dtype_to_c_struct()</span></code></a>,
for better integration of the CL and <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a> type systems.</p>
</li>
<li><p class="first">More/improved Bessel functions.
See <a class="reference external" href="https://github.com/inducer/pyopencl/tree/master/src/cl">the source</a>.</p>
</li>
<li><p class="first">Add <span class="target" id="index-1"></span><code class="xref std std-envvar docutils literal"><span class="pre">PYOPENCL_NO_CACHE</span></code> environment variable to aid debugging.
(e.g. with AMD’s CPU implementation, see
<a class="reference external" href="http://developer.amd.com/sdks/AMDAPPSDK/assets/AMD_Accelerated_Parallel_Processing_OpenCL_Programming_Guide.pdf">their programming guide</a>)</p>
</li>
<li><p class="first">Deprecated <a class="reference internal" href="array.html#pyopencl.tools.register_dtype" title="pyopencl.tools.register_dtype"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.tools.register_dtype()</span></code></a> in favor of
<a class="reference internal" href="array.html#pyopencl.tools.get_or_register_dtype" title="pyopencl.tools.get_or_register_dtype"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.tools.get_or_register_dtype()</span></code></a>.</p>
</li>
<li><p class="first">Clean up the <a class="reference internal" href="array.html#pyopencl.array.Array" title="pyopencl.array.Array"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.Array</span></code></a> constructor interface.</p>
</li>
<li><p class="first">Deprecate <code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.DefaultAllocator</span></code>.</p>
</li>
<li><p class="first">Deprecate <code class="xref py py-class docutils literal"><span class="pre">pyopencl.tools.CLAllocator</span></code>.</p>
</li>
<li><p class="first">Introduce <a class="reference internal" href="tools.html#pyopencl.tools.DeferredAllocator" title="pyopencl.tools.DeferredAllocator"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.tools.DeferredAllocator</span></code></a>, <a class="reference internal" href="tools.html#pyopencl.tools.ImmediateAllocator" title="pyopencl.tools.ImmediateAllocator"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.tools.ImmediateAllocator</span></code></a>.</p>
</li>
<li><p class="first">Allow arrays whose beginning does not coincide with the beginning of their
<a class="reference internal" href="array.html#pyopencl.array.Array.data" title="pyopencl.array.Array.data"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.array.Array.data</span></code></a> <a class="reference internal" href="runtime_memory.html#pyopencl.Buffer" title="pyopencl.Buffer"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Buffer</span></code></a>.
See <a class="reference internal" href="array.html#pyopencl.array.Array.base_data" title="pyopencl.array.Array.base_data"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.array.Array.base_data</span></code></a> and <a class="reference internal" href="array.html#pyopencl.array.Array.offset" title="pyopencl.array.Array.offset"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.array.Array.offset</span></code></a>.
Note that not all functions in PyOpenCL support such arrays just yet. These
will fail with <a class="reference internal" href="array.html#pyopencl.array.ArrayHasOffsetError" title="pyopencl.array.ArrayHasOffsetError"><code class="xref py py-exc docutils literal"><span class="pre">pyopencl.array.ArrayHasOffsetError</span></code></a>.</p>
</li>
<li><p class="first">Add <a class="reference internal" href="array.html#pyopencl.array.Array.__getitem__" title="pyopencl.array.Array.__getitem__"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array.__getitem__()</span></code></a> and <a class="reference internal" href="array.html#pyopencl.array.Array.__setitem__" title="pyopencl.array.Array.__setitem__"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array.__setitem__()</span></code></a>,
supporting generic slicing.</p>
<p>It is <em>possible</em> to create non-contiguous arrays using this functionality.
Most operations (elementwise etc.) will not work on such arrays.</p>
<p>Note also that some operations (specifically, reductions and scans) on sliced
arrays that start past the beginning of the original array will fail for now.
This will be fixed in a future release.</p>
</li>
<li><p class="first"><a class="reference internal" href="runtime_queue.html#pyopencl.CommandQueue" title="pyopencl.CommandQueue"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.CommandQueue</span></code></a> may be used as a context manager (in a <code class="docutils literal"><span class="pre">with</span></code> statement)</p>
</li>
<li><p class="first">Add <a class="reference internal" href="array.html#pyopencl.clmath.atan2" title="pyopencl.clmath.atan2"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.clmath.atan2()</span></code></a>, <a class="reference internal" href="array.html#pyopencl.clmath.atan2pi" title="pyopencl.clmath.atan2pi"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.clmath.atan2pi()</span></code></a>.</p>
</li>
<li><p class="first">Add <a class="reference internal" href="array.html#pyopencl.array.concatenate" title="pyopencl.array.concatenate"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.array.concatenate()</span></code></a>.</p>
</li>
<li><p class="first">Add <a class="reference internal" href="runtime_program.html#pyopencl.Kernel.capture_call" title="pyopencl.Kernel.capture_call"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Kernel.capture_call()</span></code></a>.</p>
</li>
</ul>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p>The addition of <a class="reference internal" href="array.html#pyopencl.array.Array.__getitem__" title="pyopencl.array.Array.__getitem__"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array.__getitem__()</span></code></a> has an unintended
consequence due to <a class="reference external" href="https://github.com/numpy/numpy/issues/3375">numpy bug 3375</a>.  For instance, this
expression:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</span><span class="p">(</span><span class="mi">5</span><span class="p">)</span> <span class="o">*</span> <span class="n">some_pyopencl_array</span>
</pre></div>
</div>
<p>may take a very long time to execute. This is because <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a> first
builds an object array of (compute-device) scalars (!) before it decides that
that’s probably not such a bright idea and finally calls
<code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array.__rmul__()</span></code>.</p>
<p>Note that only left arithmetic operations of <a class="reference internal" href="array.html#pyopencl.array.Array" title="pyopencl.array.Array"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.Array</span></code></a>
by <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a> scalars are affected. Python’s number types (<a class="reference external" href="file:///usr/share/doc/python3-doc/html/library/functions.html#float" title="(in Python v3.6)"><code class="xref py py-class docutils literal"><span class="pre">float</span></code></a> etc.)
are unaffected, as are right multiplications.</p>
<p>If a program that used to run fast suddenly runs extremely slowly, it is
likely that this bug is to blame.</p>
<p>Here’s what you can do:</p>
<ul class="last simple">
<li>Use Python scalars instead of <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a> scalars.</li>
<li>Switch to right multiplications if possible.</li>
<li>Use a patched <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a>. See the bug report linked above for a pull
request with a fix.</li>
<li>Switch to a fixed version of <a class="reference external" href="file:///usr/share/doc/python-numpy-doc/html/reference/index.html#module-numpy" title="(in NumPy v1.13)"><code class="xref py py-mod docutils literal"><span class="pre">numpy</span></code></a> when available.</li>
</ul>
</div>
</div>
<div class="section" id="version-2012-1">
<h2>Version 2012.1<a class="headerlink" href="#version-2012-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Support for complex numbers.</li>
<li>Support for Bessel functions. (experimental)</li>
<li>Numerous fixes.</li>
</ul>
</div>
<div class="section" id="version-2011-2">
<h2>Version 2011.2<a class="headerlink" href="#version-2011-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add <code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_migrate_mem_object()</span></code>.</li>
<li>Add <a class="reference internal" href="runtime_memory.html#pyopencl.image_from_array" title="pyopencl.image_from_array"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.image_from_array()</span></code></a>.</li>
<li>IMPORTANT BUGFIX: Kernel caching was broken for all the 2011.1.x releases, with
severe consequences on the execution time of <a class="reference internal" href="array.html#pyopencl.array.Array" title="pyopencl.array.Array"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.Array</span></code></a>
operations.
Henrik Andresen at a <a class="reference external" href="http://gpulab.imm.dtu.dk/courses.html">PyOpenCL workshop at DTU</a>
first noticed the strange timings.</li>
<li>All comparable PyOpenCL objects are now also hashable.</li>
<li>Add <code class="xref py py-func docutils literal"><span class="pre">pyopencl.tools.context_dependent_memoize()</span></code> to the documented
functionality.</li>
<li>Base <a class="reference internal" href="array.html#module-pyopencl.clrandom" title="pyopencl.clrandom"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.clrandom</span></code></a> on <a class="reference external" href="https://bitbucket.org/ivarun/ranluxcl">RANLUXCL</a>,
add functionality.</li>
<li>Add <a class="reference internal" href="runtime_queue.html#pyopencl.NannyEvent" title="pyopencl.NannyEvent"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.NannyEvent</span></code></a> objects.</li>
<li>Add <a class="reference internal" href="tools.html#module-pyopencl.characterize" title="pyopencl.characterize"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.characterize</span></code></a>.</li>
<li>Ensure compatibility with OS X Lion.</li>
<li>Add <a class="reference internal" href="array.html#pyopencl.tools.register_dtype" title="pyopencl.tools.register_dtype"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.tools.register_dtype()</span></code></a> to enable scan/reduction on struct types.</li>
<li><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_migrate_mem_object()</span></code> was renamed
<a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_migrate_mem_object_ext" title="pyopencl.enqueue_migrate_mem_object_ext"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_migrate_mem_object_ext()</span></code></a>.
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_migrate_mem_object()</span></code> now refers to the OpenCL 1.2 function
of this name, if available.</li>
<li><code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_sub_devices()</span></code> was renamed
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_sub_devices_ext()</span></code>.
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_sub_devices()</span></code> now refers to the OpenCL 1.2 function
of this name, if available.</li>
<li>Alpha support for OpenCL 1.2.</li>
</ul>
</div>
<div class="section" id="version-2011-1-2">
<h2>Version 2011.1.2<a class="headerlink" href="#version-2011-1-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>More bug fixes.</li>
</ul>
</div>
<div class="section" id="version-2011-1-1">
<h2>Version 2011.1.1<a class="headerlink" href="#version-2011-1-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fixes for Python 3 compatibility. (with work by Christoph Gohlke)</li>
</ul>
</div>
<div class="section" id="version-2011-1">
<h2>Version 2011.1<a class="headerlink" href="#version-2011-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>All <em>is_blocking</em> parameters now default to <em>True</em> to avoid
crashy-by-default behavior. (suggested by Jan Meinke)
In particular, this change affects
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_read_buffer()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_write_buffer()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_read_buffer_rect()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_write_buffer_rect()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_read_image()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_write_image()</span></code>,
<a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_map_buffer" title="pyopencl.enqueue_map_buffer"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_map_buffer()</span></code></a>,
<a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_map_image" title="pyopencl.enqueue_map_image"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_map_image()</span></code></a>.</li>
<li>Add <a class="reference internal" href="algorithm.html#module-pyopencl.reduction" title="pyopencl.reduction"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.reduction</span></code></a>.</li>
<li>Add <a class="reference internal" href="array.html#reductions"><span class="std std-ref">Reductions</span></a>.</li>
<li>Add <a class="reference internal" href="algorithm.html#module-pyopencl.scan" title="pyopencl.scan"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl.scan</span></code></a>.</li>
<li>Add <a class="reference internal" href="runtime_memory.html#pyopencl.MemoryObject.get_host_array" title="pyopencl.MemoryObject.get_host_array"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.MemoryObject.get_host_array()</span></code></a>.</li>
<li>Deprecate context arguments of
<a class="reference internal" href="array.html#pyopencl.array.to_device" title="pyopencl.array.to_device"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.array.to_device()</span></code></a>,
<a class="reference internal" href="array.html#pyopencl.array.zeros" title="pyopencl.array.zeros"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.array.zeros()</span></code></a>,
<a class="reference internal" href="array.html#pyopencl.array.arange" title="pyopencl.array.arange"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.array.arange()</span></code></a>.</li>
<li>Make construction of <a class="reference internal" href="array.html#pyopencl.array.Array" title="pyopencl.array.Array"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.Array</span></code></a> more flexible (<em>cqa</em> argument.)</li>
<li>Add <a class="reference internal" href="tools.html#memory-pools"><span class="std std-ref">Memory Pools</span></a>.</li>
<li>Add vector types, see <a class="reference internal" href="array.html#pyopencl.array.vec" title="pyopencl.array.vec"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.array.vec</span></code></a>.</li>
<li>Add <a class="reference internal" href="array.html#pyopencl.array.Array.strides" title="pyopencl.array.Array.strides"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.array.Array.strides</span></code></a>, <a class="reference internal" href="array.html#pyopencl.array.Array.flags" title="pyopencl.array.Array.flags"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.array.Array.flags</span></code></a>.
Allow the creation of arrays in C and Fortran order.</li>
<li>Add <a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_copy" title="pyopencl.enqueue_copy"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_copy()</span></code></a>. Deprecate all other transfer functions.</li>
<li>Add support for numerous extensions, among them device fission.</li>
<li>Add a compiler cache.</li>
<li>Add the ‘g_times_l’ keyword arg to kernel execution.</li>
</ul>
</div>
<div class="section" id="version-0-92">
<h2>Version 0.92<a class="headerlink" href="#version-0-92" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add support for OpenCL 1.1.</li>
<li>Add support for the
<a class="reference external" href="ghttp://www.khronos.org/registry/cl/extensions/khr/cl_khr_gl_sharing.txt">cl_khr_gl_sharing</a>
extension, leading to working GL interoperability.</li>
<li>Add <a class="reference internal" href="runtime_program.html#pyopencl.Kernel.set_args" title="pyopencl.Kernel.set_args"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Kernel.set_args()</span></code></a>.</li>
<li>The call signature of <a class="reference internal" href="runtime_program.html#pyopencl.Kernel.__call__" title="pyopencl.Kernel.__call__"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Kernel.__call__()</span></code></a> changed to
emphasize the importance of <em>local_size</em>.</li>
<li>Add <a class="reference internal" href="runtime_program.html#pyopencl.Kernel.set_scalar_arg_dtypes" title="pyopencl.Kernel.set_scalar_arg_dtypes"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Kernel.set_scalar_arg_dtypes()</span></code></a>.</li>
<li>Add support for the
<a class="reference external" href="http://www.khronos.org/registry/cl/extensions/khr/cl_nv_device_attribute_query.txt">cl_nv_device_attribute_query</a>
extension.</li>
<li>Add <a class="reference internal" href="array.html#pyopencl.array.Array" title="pyopencl.array.Array"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.array.Array()</span></code></a> and related functionality.</li>
<li>Make build not depend on Boost C++.</li>
</ul>
</div>
<div class="section" id="version-0-91-5">
<h2>Version 0.91.5<a class="headerlink" href="#version-0-91-5" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add <a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat.channel_count" title="pyopencl.ImageFormat.channel_count"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.ImageFormat.channel_count</span></code></a>,
<a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat.dtype_size" title="pyopencl.ImageFormat.dtype_size"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.ImageFormat.dtype_size</span></code></a>,
<a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat.itemsize" title="pyopencl.ImageFormat.itemsize"><code class="xref py py-attr docutils literal"><span class="pre">pyopencl.ImageFormat.itemsize</span></code></a>.</li>
<li>Add missing <a class="reference internal" href="runtime_memory.html#pyopencl.enqueue_copy_buffer" title="pyopencl.enqueue_copy_buffer"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_copy_buffer()</span></code></a>.</li>
<li>Add <a class="reference internal" href="runtime_platform.html#pyopencl.create_some_context" title="pyopencl.create_some_context"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_some_context()</span></code></a>.</li>
<li>Add <a class="reference internal" href="runtime_queue.html#pyopencl.enqueue_barrier" title="pyopencl.enqueue_barrier"><code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_barrier()</span></code></a>, which was previously missing.</li>
</ul>
</div>
<div class="section" id="version-0-91-4">
<h2>Version 0.91.4<a class="headerlink" href="#version-0-91-4" title="Permalink to this headline"></a></h2>
<p>A bugfix release. No user-visible changes.</p>
</div>
<div class="section" id="version-0-91-3">
<h2>Version 0.91.3<a class="headerlink" href="#version-0-91-3" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>All parameters named <em>host_buffer</em> were renamed <em>hostbuf</em> for consistency
with the <a class="reference internal" href="runtime_memory.html#pyopencl.Buffer" title="pyopencl.Buffer"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Buffer</span></code></a> constructor introduced in 0.91.
Compatibility code is in place.</li>
<li>The <code class="xref py py-class docutils literal"><span class="pre">pyopencl.Image</span></code> constructor does not need a <em>shape</em> parameter if the
given <em>hostbuf</em> has <em>hostbuf.shape</em>.</li>
<li>The <a class="reference internal" href="runtime_platform.html#pyopencl.Context" title="pyopencl.Context"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Context</span></code></a> constructor can now be called without parameters.</li>
</ul>
</div>
<div class="section" id="version-0-91-2">
<h2>Version 0.91.2<a class="headerlink" href="#version-0-91-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li><a class="reference internal" href="runtime_program.html#pyopencl.Program.build" title="pyopencl.Program.build"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Program.build()</span></code></a> now captures build logs and adds them
to the exception text.</li>
<li>Deprecate <code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_context_from_type()</span></code> in favor of second
form of <a class="reference internal" href="runtime_platform.html#pyopencl.Context" title="pyopencl.Context"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Context</span></code></a> constructor</li>
<li>Introduce <a class="reference internal" href="runtime_program.html#pyopencl.LocalMemory" title="pyopencl.LocalMemory"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.LocalMemory</span></code></a>.</li>
<li>Document kernel invocation and <a class="reference internal" href="runtime_program.html#pyopencl.Kernel.set_arg" title="pyopencl.Kernel.set_arg"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Kernel.set_arg()</span></code></a>.</li>
</ul>
</div>
<div class="section" id="version-0-91-1">
<h2>Version 0.91.1<a class="headerlink" href="#version-0-91-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fixed a number of bugs, notably involving <a class="reference internal" href="runtime_memory.html#pyopencl.Sampler" title="pyopencl.Sampler"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Sampler</span></code></a>.</li>
<li><a class="reference internal" href="runtime_platform.html#pyopencl.Device" title="pyopencl.Device"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Device</span></code></a>, <a class="reference internal" href="runtime_platform.html#pyopencl.Platform" title="pyopencl.Platform"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Platform</span></code></a>,
<a class="reference internal" href="runtime_platform.html#pyopencl.Context" title="pyopencl.Context"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Context</span></code></a> now have nicer string representations.</li>
<li>Add <code class="xref py py-attr docutils literal"><span class="pre">Image.shape</span></code>. (suggested by David Garcia)</li>
</ul>
</div>
<div class="section" id="version-0-91">
<h2>Version 0.91<a class="headerlink" href="#version-0-91" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add <a class="reference internal" href="runtime_gl.html#gl-interop"><span class="std std-ref">OpenCL Runtime: OpenGL Interoperability</span></a>.</li>
<li>Add a test suite.</li>
<li>Fix numerous <cite>get_info</cite> bugs. (reports by David Garcia and the test suite)</li>
<li>Add <a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat.__repr__" title="pyopencl.ImageFormat.__repr__"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.ImageFormat.__repr__()</span></code></a>.</li>
<li>Add <code class="xref py py-meth docutils literal"><span class="pre">pyopencl.addressing_mode.to_string()</span></code> and colleagues.</li>
<li>The <cite>pitch</cite> arguments to
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_image_2d()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_image_3d()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_read_image()</span></code>, and
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.enqueue_write_image()</span></code>
are now defaulted to zero. The argument order of <cite>enqueue_{read,write}_image</cite>
has changed for this reason.</li>
<li>Deprecate
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_image_2d()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_image_3d()</span></code>
in favor of the <code class="xref py py-class docutils literal"><span class="pre">pyopencl.Image</span></code> constructor.</li>
<li>Deprecate
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_program_with_source()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_program_with_binary()</span></code>
in favor of the <a class="reference internal" href="runtime_program.html#pyopencl.Program" title="pyopencl.Program"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Program</span></code></a> constructor.</li>
<li>Deprecate
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_buffer()</span></code>,
<code class="xref py py-func docutils literal"><span class="pre">pyopencl.create_host_buffer()</span></code>
in favor of the <a class="reference internal" href="runtime_memory.html#pyopencl.Buffer" title="pyopencl.Buffer"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.Buffer</span></code></a> constructor.</li>
<li><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.MemoryObject.get_image_info()</span></code> now actually exists.</li>
<li>Add <code class="xref py py-attr docutils literal"><span class="pre">pyopencl.MemoryObject.image.info</span></code>.</li>
<li>Fix API tracing.</li>
<li>Add constructor arguments to <a class="reference internal" href="runtime_memory.html#pyopencl.ImageFormat" title="pyopencl.ImageFormat"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.ImageFormat</span></code></a>.  (suggested by David Garcia)</li>
</ul>
</div>
<div class="section" id="version-0-90-4">
<h2>Version 0.90.4<a class="headerlink" href="#version-0-90-4" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Add build fixes for Windows and OS X.</li>
</ul>
</div>
<div class="section" id="version-0-90-3">
<h2>Version 0.90.3<a class="headerlink" href="#version-0-90-3" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix a GNU-ism in the C++ code of the wrapper.</li>
</ul>
</div>
<div class="section" id="version-0-90-2">
<h2>Version 0.90.2<a class="headerlink" href="#version-0-90-2" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix <a class="reference internal" href="runtime_platform.html#pyopencl.Platform.get_info" title="pyopencl.Platform.get_info"><code class="xref py py-meth docutils literal"><span class="pre">pyopencl.Platform.get_info()</span></code></a>.</li>
<li>Fix passing properties to <a class="reference internal" href="runtime_queue.html#pyopencl.CommandQueue" title="pyopencl.CommandQueue"><code class="xref py py-class docutils literal"><span class="pre">pyopencl.CommandQueue</span></code></a>.
Also fix related documentation.</li>
</ul>
</div>
<div class="section" id="version-0-90-1">
<h2>Version 0.90.1<a class="headerlink" href="#version-0-90-1" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Fix building on the Mac.</li>
</ul>
</div>
<div class="section" id="version-0-90">
<h2>Version 0.90<a class="headerlink" href="#version-0-90" title="Permalink to this headline"></a></h2>
<ul class="simple">
<li>Initial release.</li>
</ul>
</div>
</div>
<div class="section" id="license">
<span id="id1"></span><h1>License<a class="headerlink" href="#license" title="Permalink to this headline"></a></h1>
<p>PyOpenCL is licensed to you under the MIT/X Consortium license:</p>
<p>Copyright (c) 2009-13 Andreas Klöckner and Contributors.</p>
<p>Permission is hereby granted, free of charge, to any person
obtaining a copy of this software and associated documentation
files (the “Software”), to deal in the Software without
restriction, including without limitation the rights to use,
copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the
Software is furnished to do so, subject to the following
conditions:</p>
<p>The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.</p>
<p>THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
OTHER DEALINGS IN THE SOFTWARE.</p>
<p>PyOpenCL includes derivatives of parts of the <a class="reference external" href="https://github.com/thrust/thrust/">Thrust</a> computing package (in particular the scan
implementation). These parts are licensed as follows:</p>
<blockquote>
<div><p>Copyright 2008-2011 NVIDIA Corporation</p>
<p>Licensed under the Apache License, Version 2.0 (the “License”);
you may not use this file except in compliance with the License.
You may obtain a copy of the License at</p>
<blockquote>
<div>&lt;<a class="reference external" href="http://www.apache.org/licenses/LICENSE-2.0">http://www.apache.org/licenses/LICENSE-2.0</a>&gt;</div></blockquote>
<p>Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an “AS IS” BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.</p>
</div></blockquote>
<div class="admonition note">
<p class="first admonition-title">Note</p>
<p class="last">If you use Apache-licensed parts, be aware that these may be incompatible
with software licensed exclusively under GPL2.  (Most software is licensed
as GPL2 or later, in which case this is not an issue.)</p>
</div>
<p>PyOpenCL includes parts of the Random123 suite of random number generators:</p>
<blockquote>
<div><p>Copyright 2010-2012, D. E. Shaw Research.
All rights reserved.</p>
<p>Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:</p>
<ul class="simple">
<li>Redistributions of source code must retain the above copyright
notice, this list of conditions, and the following disclaimer.</li>
<li>Redistributions in binary form must reproduce the above copyright
notice, this list of conditions, and the following disclaimer in the
documentation and/or other materials provided with the distribution.</li>
<li>Neither the name of D. E. Shaw Research nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.</li>
</ul>
<p>THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
“AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.</p>
</div></blockquote>
<p>PyOpenCL includes the RANLUXCL random number generator:</p>
<blockquote>
<div><p>Copyright (c) 2011 Ivar Ursin Nikolaisen</p>
<p>Permission is hereby granted, free of charge, to any person obtaining a copy of this
software and associated documentation files (the “Software”), to deal in the Software
without restriction, including without limitation the rights to use, copy, modify,
merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to the following
conditions:</p>
<p>The above copyright notice and this permission notice shall be included in all copies
or substantial portions of the Software.</p>
<p>THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.</p>
</div></blockquote>
</div>
<div class="section" id="frequently-asked-questions">
<h1>Frequently Asked Questions<a class="headerlink" href="#frequently-asked-questions" title="Permalink to this headline"></a></h1>
<p>The FAQ is maintained collaboratively on the
<a class="reference external" href="http://wiki.tiker.net/PyOpenCL/FrequentlyAskedQuestions">Wiki FAQ page</a>.</p>
</div>
<div class="section" id="citing-pyopencl">
<h1>Citing PyOpenCL<a class="headerlink" href="#citing-pyopencl" title="Permalink to this headline"></a></h1>
<p>We are not asking you to gratuitously cite PyOpenCL in work that is otherwise
unrelated to software. That said, if you do discuss some of the development
aspects of your code and would like to highlight a few of the ideas behind
PyOpenCL, feel free to cite <a class="reference external" href="http://dx.doi.org/10.1016/j.parco.2011.09.001">this article</a>:</p>
<blockquote>
<div>Andreas Klöckner, Nicolas Pinto, Yunsup Lee, Bryan Catanzaro, Paul Ivanov,
Ahmed Fasih, PyCUDA and PyOpenCL: A scripting-based approach to GPU
run-time code generation, Parallel Computing, Volume 38, Issue 3, March
2012, Pages 157-174.</div></blockquote>
<p>Here’s a Bibtex entry for your convenience:</p>
<div class="highlight-default"><div class="highlight"><pre><span></span><span class="nd">@article</span><span class="p">{</span><span class="n">kloeckner_pycuda_2012</span><span class="p">,</span>
   <span class="n">author</span> <span class="o">=</span> <span class="p">{{</span><span class="n">Kl</span><span class="p">{</span>\<span class="s2">&quot;o}ckner}, Andreas</span>
        <span class="ow">and</span> <span class="p">{</span><span class="n">Pinto</span><span class="p">},</span> <span class="n">Nicolas</span>
        <span class="ow">and</span> <span class="p">{</span><span class="n">Lee</span><span class="p">},</span> <span class="n">Yunsup</span>
        <span class="ow">and</span> <span class="p">{</span><span class="n">Catanzaro</span><span class="p">},</span> <span class="n">B</span><span class="o">.</span>
        <span class="ow">and</span> <span class="p">{</span><span class="n">Ivanov</span><span class="p">},</span> <span class="n">Paul</span>
        <span class="ow">and</span> <span class="p">{</span><span class="n">Fasih</span><span class="p">},</span> <span class="n">Ahmed</span> <span class="p">},</span>
   <span class="n">title</span> <span class="o">=</span> <span class="s2">&quot;{PyCUDA and PyOpenCL: A Scripting-Based Approach to GPU Run-Time Code Generation}&quot;</span><span class="p">,</span>
   <span class="n">journal</span> <span class="o">=</span> <span class="s2">&quot;Parallel Computing&quot;</span><span class="p">,</span>
   <span class="n">volume</span> <span class="o">=</span> <span class="s2">&quot;38&quot;</span><span class="p">,</span>
   <span class="n">number</span> <span class="o">=</span> <span class="s2">&quot;3&quot;</span><span class="p">,</span>
   <span class="n">pages</span> <span class="o">=</span> <span class="s2">&quot;157--174&quot;</span><span class="p">,</span>
   <span class="n">year</span> <span class="o">=</span> <span class="s2">&quot;2012&quot;</span><span class="p">,</span>
   <span class="n">issn</span> <span class="o">=</span> <span class="s2">&quot;0167-8191&quot;</span><span class="p">,</span>
   <span class="n">doi</span> <span class="o">=</span> <span class="s2">&quot;10.1016/j.parco.2011.09.001&quot;</span><span class="p">,</span>
<span class="p">}</span>
</pre></div>
</div>
</div>
<div class="section" id="acknowledgments">
<h1>Acknowledgments<a class="headerlink" href="#acknowledgments" title="Permalink to this headline"></a></h1>
<div class="section" id="contributors">
<h2>Contributors<a class="headerlink" href="#contributors" title="Permalink to this headline"></a></h2>
<p>Too many to list. Please see the
<a class="reference external" href="https://github.com/inducer/pyopencl/commits/master">commit log</a>
for detailed acknowledgments.</p>
</div>
<div class="section" id="funding">
<h2>Funding<a class="headerlink" href="#funding" title="Permalink to this headline"></a></h2>
<p>Andreas Klöckner’s work on <a class="reference internal" href="runtime.html#module-pyopencl" title="pyopencl"><code class="xref py py-mod docutils literal"><span class="pre">pyopencl</span></code></a> was supported in part by</p>
<ul class="simple">
<li>US Navy ONR grant number N00014-14-1-0117</li>
<li>the US National Science Foundation under grant numbers DMS-1418961 and CCF-1524433.</li>
</ul>
<p>AK also gratefully acknowledges a hardware gift from Nvidia Corporation.  The
views and opinions expressed herein do not necessarily reflect those of the
funding agencies.</p>
</div>
</div>


          </div>
        </div>
      </div>
      <div class="sphinxsidebar" role="navigation" aria-label="main navigation">
        <div class="sphinxsidebarwrapper">
<h1 class="logo"><a href="index.html">PyOpenCL</a></h1>








<h3>Navigation</h3>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="runtime.html">OpenCL Runtime: Basics</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_const.html">OpenCL Runtime: Constants</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_platform.html">OpenCL Runtime: Platforms, Devices and Contexts</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_queue.html">OpenCL Runtime: Command Queues and Events</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_memory.html">OpenCL Runtime: Memory</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_program.html">OpenCL Runtime: Programs and Kernels</a></li>
<li class="toctree-l1"><a class="reference internal" href="runtime_gl.html">OpenCL Runtime: OpenGL Interoperability</a></li>
<li class="toctree-l1"><a class="reference internal" href="array.html">Multi-dimensional arrays</a></li>
<li class="toctree-l1"><a class="reference internal" href="algorithm.html">Parallel Algorithms</a></li>
<li class="toctree-l1"><a class="reference internal" href="howto.html">How-tos</a></li>
<li class="toctree-l1"><a class="reference internal" href="tools.html">Built-in Utilities</a></li>
<li class="toctree-l1 current"><a class="current reference internal" href="#">Installation</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#using-vendor-supplied-opencl-drivers-linux">Using vendor-supplied OpenCL drivers (Linux)</a></li>
<li class="toctree-l2"><a class="reference internal" href="#getting-a-better-cpu-based-opencl-driver-os-x">Getting a better CPU-based OpenCL driver (OS X)</a></li>
<li class="toctree-l2"><a class="reference internal" href="#installing-from-source">Installing from source</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="#tips">Tips</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#syntax-highlighting">Syntax highlighting</a></li>
<li class="toctree-l2"><a class="reference internal" href="#ipython-integration">IPython integration</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="#guidelines">Guidelines</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#api-stability">API Stability</a></li>
<li class="toctree-l2"><a class="reference internal" href="#relation-with-opencl-s-c-bindings">Relation with OpenCL’s C Bindings</a></li>
<li class="toctree-l2"><a class="reference internal" href="#interoperability-with-other-opencl-software">Interoperability with other OpenCL software</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="#user-visible-changes">User-visible Changes</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#version-2016-3">Version 2016.3</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2016-2">Version 2016.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2016-1">Version 2016.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-2-4">Version 2015.2.4</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-2-3">Version 2015.2.3</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-2-2">Version 2015.2.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-2-1">Version 2015.2.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-2">Version 2015.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2015-1">Version 2015.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2014-1">Version 2014.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2013-2">Version 2013.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2013-1">Version 2013.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2012-1">Version 2012.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2011-2">Version 2011.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2011-1-2">Version 2011.1.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2011-1-1">Version 2011.1.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-2011-1">Version 2011.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-92">Version 0.92</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91-5">Version 0.91.5</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91-4">Version 0.91.4</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91-3">Version 0.91.3</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91-2">Version 0.91.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91-1">Version 0.91.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-91">Version 0.91</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-90-4">Version 0.90.4</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-90-3">Version 0.90.3</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-90-2">Version 0.90.2</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-90-1">Version 0.90.1</a></li>
<li class="toctree-l2"><a class="reference internal" href="#version-0-90">Version 0.90</a></li>
</ul>
</li>
<li class="toctree-l1"><a class="reference internal" href="#license">License</a></li>
<li class="toctree-l1"><a class="reference internal" href="#frequently-asked-questions">Frequently Asked Questions</a></li>
<li class="toctree-l1"><a class="reference internal" href="#citing-pyopencl">Citing PyOpenCL</a></li>
<li class="toctree-l1"><a class="reference internal" href="#acknowledgments">Acknowledgments</a><ul>
<li class="toctree-l2"><a class="reference internal" href="#contributors">Contributors</a></li>
<li class="toctree-l2"><a class="reference internal" href="#funding">Funding</a></li>
</ul>
</li>
</ul>


<hr />
<ul>
    
    <li class="toctree-l1"><a href="https://github.com/inducer/pyopencl">🚀 Github</a></li>
    
    <li class="toctree-l1"><a href="https://pypi.python.org/pypi/pyopencl">💾 Download Releases</a></li>
    
</ul>
<div class="relations">
<h3>Related Topics</h3>
<ul>
  <li><a href="index.html">Documentation overview</a><ul>
      <li>Previous: <a href="tools.html" title="previous chapter">Built-in Utilities</a></li>
  </ul></li>
</ul>
</div>
<div id="searchbox" style="display: none" role="search">
  <h3>Quick search</h3>
    <form class="search" action="search.html" method="get">
      <div><input type="text" name="q" /></div>
      <div><input type="submit" value="Go" /></div>
      <input type="hidden" name="check_keywords" value="yes" />
      <input type="hidden" name="area" value="default" />
    </form>
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
        </div>
      </div>
      <div class="clearer"></div>
    </div>
    <div class="footer">
      &copy;2017, Andreas Kloeckner.
      
      |
      Powered by <a href="http://sphinx-doc.org/">Sphinx 1.6.5</a>
      &amp; <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.8</a>
      
    </div>

    

    
  </body>
</html>