ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/OpenCL/OpenCL.pm
(Generate patch)

Comparing OpenCL/OpenCL.pm (file contents):
Revision 1.5 by root, Wed Nov 16 00:35:30 2011 UTC vs.
Revision 1.70 by root, Thu May 3 23:32:47 2012 UTC

6 6
7 use OpenCL; 7 use OpenCL;
8 8
9=head1 DESCRIPTION 9=head1 DESCRIPTION
10 10
11This is an early release which might be useful, but hasn't seen any testing. 11This is an early release which might be useful, but hasn't seen much testing.
12 12
13=head2 OpenCL FROM 10000 FEET HEIGHT
14
15Here is a high level overview of OpenCL:
16
17First you need to find one or more OpenCL::Platforms (kind of like
18vendors) - usually there is only one.
19
20Each platform gives you access to a number of OpenCL::Device objects, e.g.
21your graphics card.
22
23From a platform and some device(s), you create an OpenCL::Context, which is
24a very central object in OpenCL: Once you have a context you can create
25most other objects:
26
27OpenCL::Program objects, which store source code and, after building for a
28specific device ("compiling and linking"), also binary programs. For each
29kernel function in a program you can then create an OpenCL::Kernel object
30which represents basically a function call with argument values.
31
32OpenCL::Memory objects of various flavours: OpenCL::Buffer objects (flat
33memory areas, think arrays or structs) and OpenCL::Image objects (think 2d
34or 3d array) for bulk data and input and output for kernels.
35
36OpenCL::Sampler objects, which are kind of like texture filter modes in
37OpenGL.
38
39OpenCL::Queue objects - command queues, which allow you to submit memory
40reads, writes and copies, as well as kernel calls to your devices. They
41also offer a variety of methods to synchronise request execution, for
42example with barriers or OpenCL::Event objects.
43
44OpenCL::Event objects are used to signal when something is complete.
45
13=head1 HELPFUL RESOURCES 46=head2 HELPFUL RESOURCES
14 47
15The OpenCL spec used to develop this module (1.2 spec was available, but 48The OpenCL specs used to develop this module:
16no implementation was available to me :).
17 49
18 http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf 50 http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf
51 http://www.khronos.org/registry/cl/specs/opencl-1.2.pdf
52 http://www.khronos.org/registry/cl/specs/opencl-1.2-extensions.pdf
19 53
20OpenCL manpages: 54OpenCL manpages:
21 55
22 http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/ 56 http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/
57 http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/
58
59If you are into UML class diagrams, the following diagram might help - if
60not, it will be mildly confusing (also, the class hierarchy of this module
61is much more fine-grained):
62
63 http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/classDiagram.html
64
65Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it
66is, but at least it's free of charge:
67
68 http://developer.amd.com/zones/OpenCLZone/courses/Documents/Introduction_to_OpenCL_Programming%20Training_Guide%20%28201005%29.pdf
69
70And here's NVIDIA's OpenCL Best Practises Guide:
71
72 http://developer.download.nvidia.com/compute/cuda/3_2/toolkit/docs/OpenCL_Best_Practices_Guide.pdf
73
74=head1 BASIC WORKFLOW
75
76To get something done, you basically have to do this once (refer to the
77examples below for actual code, this is just a high-level description):
78
79Find some platform (e.g. the first one) and some device(s) (e.g. the first
80device of the platform), and create a context from those.
81
82Create program objects from your OpenCL source code, then build (compile)
83the programs for each device you want to run them on.
84
85Create kernel objects for all kernels you want to use (surprisingly, these
86are not device-specific).
87
88Then, to execute stuff, you repeat these steps, possibly resuing or
89sharing some buffers:
90
91Create some input and output buffers from your context. Set these as
92arguments to your kernel.
93
94Enqueue buffer writes to initialise your input buffers (when not
95initialised at creation time).
96
97Enqueue the kernel execution.
98
99Enqueue buffer reads for your output buffer to read results.
23 100
24=head1 EXAMPLES 101=head1 EXAMPLES
25 102
26=head2 Enumerate all devices and get contexts for them. 103=head2 Enumerate all devices and get contexts for them.
27 104
105Best run this once to get a feel for the platforms and devices in your
106system.
107
28 for my $platform (OpenCL::platforms) { 108 for my $platform (OpenCL::platforms) {
29 warn $platform->info (OpenCL::PLATFORM_NAME); 109 printf "platform: %s\n", $platform->name;
30 warn $platform->info (OpenCL::PLATFORM_EXTENSIONS); 110 printf "extensions: %s\n", $platform->extensions;
31 for my $device ($platform->devices) { 111 for my $device ($platform->devices) {
32 warn $device->info (OpenCL::DEVICE_NAME); 112 printf "+ device: %s\n", $device->name;
33 my $ctx = $device->context_simple; 113 my $ctx = $platform->context (undef, [$device]);
34 # do stuff 114 # do stuff
35 } 115 }
36 } 116 }
37 117
38=head2 Get a useful context and a command queue. 118=head2 Get a useful context and a command queue.
39 119
40 my $dev = ((OpenCL::platforms)[0]->devices)[0]; 120This is a useful boilerplate for any OpenCL program that only wants to use
41 my $ctx = $dev->context_simple; 121one device,
42 my $queue = $ctx->command_queue_simple ($dev); 122
123 my ($platform) = OpenCL::platforms; # find first platform
124 my ($dev) = $platform->devices; # find first device of platform
125 my $ctx = $platform->context (undef, [$dev]); # create context out of those
126 my $queue = $ctx->queue ($dev); # create a command queue for the device
43 127
44=head2 Print all supported image formats of a context. 128=head2 Print all supported image formats of a context.
45 129
130Best run this once for your context, to see whats available and how to
131gather information.
132
46 for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) { 133 for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) {
47 say "supported image formats for ", OpenCL::enum2str $type; 134 print "supported image formats for ", OpenCL::enum2str $type, "\n";
48 135
49 for my $f ($ctx->supported_image_formats (0, $type)) { 136 for my $f ($ctx->supported_image_formats (0, $type)) {
50 printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1]; 137 printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1];
51 } 138 }
52 } 139 }
54=head2 Create a buffer with some predefined data, read it back synchronously, 141=head2 Create a buffer with some predefined data, read it back synchronously,
55then asynchronously. 142then asynchronously.
56 143
57 my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut"); 144 my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut");
58 145
59 $queue->enqueue_read_buffer ($buf, 1, 1, 3, my $data); 146 $queue->read_buffer ($buf, 1, 1, 3, my $data);
60 warn $data; 147 print "$data\n";
61 148
62 my $ev = $queue->enqueue_read_buffer ($buf, 0, 1, 3, my $data); 149 my $ev = $queue->read_buffer ($buf, 0, 1, 3, my $data);
63 $ev->wait; 150 $ev->wait;
64 warn $data; 151 print "$data\n"; # prints "elm"
65 152
66=head2 Create and build a program, then create a kernel out of one of its 153=head2 Create and build a program, then create a kernel out of one of its
67functions. 154functions.
68 155
69 my $src = ' 156 my $src = '
70 __kernel void 157 kernel void
71 squareit (__global float *input, __global float *output) 158 squareit (global float *input, global float *output)
72 { 159 {
73 size_t id = get_global_id (0); 160 $id = get_global_id (0);
74 output [id] = input [id] * input [id]; 161 output [id] = input [id] * input [id];
75 } 162 }
76 '; 163 ';
77 164
78 my $prog = $ctx->program_with_source ($src); 165 my $prog = $ctx->build_program ($src);
79
80 eval { $prog->build ($dev); 1 }
81 or die $prog->build_info ($dev, OpenCL::PROGRAM_BUILD_LOG);
82
83 my $kernel = $prog->kernel ("squareit"); 166 my $kernel = $prog->kernel ("squareit");
84 167
85=head2 Create some input and output float buffers, then call squareit on them. 168=head2 Create some input and output float buffers, then call the
169'squareit' kernel on them.
86 170
87 my $input = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, pack "f*", 1, 2, 3, 4.5); 171 my $input = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, pack "f*", 1, 2, 3, 4.5);
88 my $output = $ctx->buffer (0, OpenCL::SIZEOF_FLOAT * 5); 172 my $output = $ctx->buffer (0, OpenCL::SIZEOF_FLOAT * 5);
89 173
90 # set buffer 174 # set buffer
91 $kernel->set_buffer (0, $input); 175 $kernel->set_buffer (0, $input);
92 $kernel->set_buffer (1, $output); 176 $kernel->set_buffer (1, $output);
93 177
94 # execute it for all 4 numbers 178 # execute it for all 4 numbers
95 $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 179 $queue->nd_range_kernel ($kernel, undef, [4], undef);
96 180
97 # enqueue a synchronous read 181 # enqueue a synchronous read
98 $queue->enqueue_read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); 182 $queue->read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
99 183
100 # print the results: 184 # print the results:
101 say join ", ", unpack "f*", $data; 185 printf "%s\n", join ", ", unpack "f*", $data;
102 186
103=head2 The same enqueue operations as before, but assuming an out-of-order queue, 187=head2 The same enqueue operations as before, but assuming an out-of-order queue,
104showing off barriers. 188showing off barriers.
105 189
106 # execute it for all 4 numbers 190 # execute it for all 4 numbers
107 $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 191 $queue->nd_range_kernel ($kernel, undef, [4], undef);
108 192
109 # enqueue a barrier to ensure in-order execution 193 # enqueue a barrier to ensure in-order execution
110 $queue->enqueue_barrier; 194 $queue->barrier;
111 195
112 # enqueue an async read 196 # enqueue an async read
113 $queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); 197 $queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
114 198
115 # wait for all requests to finish 199 # wait for all requests to finish
116 $queue->finish; 200 $queue->finish;
117 201
118=head2 The same enqueue operations as before, but assuming an out-of-order queue, 202=head2 The same enqueue operations as before, but assuming an out-of-order queue,
119showing off event objects and wait lists. 203showing off event objects and wait lists.
120 204
121 # execute it for all 4 numbers 205 # execute it for all 4 numbers
122 my $ev = $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 206 my $ev = $queue->nd_range_kernel ($kernel, undef, [4], undef);
123 207
124 # enqueue an async read 208 # enqueue an async read
125 $ev = $queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev); 209 $ev = $queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev);
126 210
127 # wait for the last event to complete 211 # wait for the last event to complete
128 $ev->wait; 212 $ev->wait;
129 213
214=head2 Use the OpenGL module to share a texture between OpenCL and OpenGL and draw some julia
215set tunnel effect.
216
217This is quite a long example to get you going - you can download it from
218L<http://cvs.schmorp.de/OpenCL/examples/juliaflight>.
219
220 use OpenGL ":all";
221 use OpenCL;
222
223 my $S = $ARGV[0] || 256; # window/texture size, smaller is faster
224
225 # open a window and create a gl texture
226 OpenGL::glpOpenWindow width => $S, height => $S;
227 my $texid = glGenTextures_p 1;
228 glBindTexture GL_TEXTURE_2D, $texid;
229 glTexImage2D_c GL_TEXTURE_2D, 0, GL_RGBA8, $S, $S, 0, GL_RGBA, GL_UNSIGNED_BYTE, 0;
230
231 # find and use the first opencl device that let's us get a shared opengl context
232 my $platform;
233 my $dev;
234 my $ctx;
235
236 for (OpenCL::platforms) {
237 $platform = $_;
238 for ($platform->devices) {
239 $dev = $_;
240 $ctx = $platform->context ([OpenCL::GLX_DISPLAY_KHR, undef, OpenCL::GL_CONTEXT_KHR, undef], [$dev])
241 and last;
242 }
243 }
244
245 $ctx
246 or die "cannot find suitable OpenCL device\n";
247
248 my $queue = $ctx->queue ($dev);
249
250 # now attach an opencl image2d object to the opengl texture
251 my $tex = $ctx->gl_texture2d (OpenCL::MEM_WRITE_ONLY, GL_TEXTURE_2D, 0, $texid);
252
253 # now the boring opencl code
254 my $src = <<EOF;
255 kernel void
256 juliatunnel (write_only image2d_t img, float time)
257 {
258 int2 xy = (int2)(get_global_id (0), get_global_id (1));
259 float2 p = convert_float2 (xy) / $S.f * 2.f - 1.f;
260
261 float2 m = (float2)(1.f, p.y) / fabs (p.x); // tunnel
262 m.x = fabs (fmod (m.x + time * 0.05f, 4.f) - 2.f);
263
264 float2 z = m;
265 float2 c = (float2)(sin (time * 0.01133f), cos (time * 0.02521f));
266
267 for (int i = 0; i < 25 && dot (z, z) < 4.f; ++i) // standard julia
268 z = (float2)(z.x * z.x - z.y * z.y, 2.f * z.x * z.y) + c;
269
270 float3 colour = (float3)(z.x, z.y, atan2 (z.y, z.x));
271 write_imagef (img, xy, (float4)(colour * p.x * p.x, 1.));
272 }
273 EOF
274
275 my $prog = $ctx->build_program ($src);
276 my $kernel = $prog->kernel ("juliatunnel");
277
278 # program compiled, kernel ready, now draw and loop
279
280 for (my $time; ; ++$time) {
281 # acquire objects from opengl
282 $queue->acquire_gl_objects ([$tex]);
283
284 # configure and run our kernel
285 $kernel->setf ("mf", $tex, $time*2); # mf = memory object, float
286 $queue->nd_range_kernel ($kernel, undef, [$S, $S], undef);
287
288 # release objects to opengl again
289 $queue->release_gl_objects ([$tex]);
290
291 # wait
292 $queue->finish;
293
294 # now draw the texture, the defaults should be all right
295 glTexParameterf GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST;
296
297 glEnable GL_TEXTURE_2D;
298 glBegin GL_QUADS;
299 glTexCoord2f 0, 1; glVertex3i -1, -1, -1;
300 glTexCoord2f 0, 0; glVertex3i 1, -1, -1;
301 glTexCoord2f 1, 0; glVertex3i 1, 1, -1;
302 glTexCoord2f 1, 1; glVertex3i -1, 1, -1;
303 glEnd;
304
305 glXSwapBuffers;
306
307 select undef, undef, undef, 1/60;
308 }
309
310=head2 How to modify the previous example to not rely on GL sharing.
311
312For those poor souls with only a sucky CPU OpenCL implementation, you
313currently have to read the image into some perl scalar, and then modify a
314texture or use glDrawPixels or so).
315
316First, when you don't need gl sharing, you can create the context much simpler:
317
318 $ctx = $platform->context (undef, [$dev])
319
320To use a texture, you would modify the above example by creating an
321OpenCL::Image manually instead of deriving it from a texture:
322
323 my $tex = $ctx->image2d (OpenCL::MEM_WRITE_ONLY, OpenCL::RGBA, OpenCL::UNORM_INT8, $S, $S);
324
325And in the darw loop, intead of acquire_gl_objects/release_gl_objects, you
326would read the image2d after the kernel has written it:
327
328 $queue->read_image ($tex, 0, 0, 0, 0, $S, $S, 1, 0, 0, my $data);
329
330And then you would upload the pixel data to the texture (or use glDrawPixels):
331
332 glTexSubImage2D_s GL_TEXTURE_2D, 0, 0, 0, $S, $S, GL_RGBA, GL_UNSIGNED_BYTE, $data;
333
334The fully modified example can be found at
335L<http://cvs.schmorp.de/OpenCL/examples/juliaflight-nosharing>.
336
130=head1 DOCUMENTATION 337=head1 DOCUMENTATION
131 338
132=head2 BASIC CONVENTIONS 339=head2 BASIC CONVENTIONS
133 340
134This is not a 1:1 C-style translation of OpenCL to Perl - instead I 341This is not a one-to-one C-style translation of OpenCL to Perl - instead
135attempted to make the interface as type-safe as possible and introducing 342I attempted to make the interface as type-safe as possible by introducing
136object syntax where it makes sense. There are a number of important 343object syntax where it makes sense. There are a number of important
137differences between the OpenCL C API and this module: 344differences between the OpenCL C API and this module:
138 345
139=over 4 346=over 4
140 347
141=item * Object lifetime managament is automatic - there is no need 348=item * Object lifetime managament is automatic - there is no need
142to free objects explicitly (C<clReleaseXXX>), the release function 349to free objects explicitly (C<clReleaseXXX>), the release function
143is called automatically once all Perl references to it go away. 350is called automatically once all Perl references to it go away.
144 351
145=item * OpenCL uses CamelCase for function names (C<clGetPlatformInfo>), 352=item * OpenCL uses CamelCase for function names
353(e.g. C<clGetPlatformIDs>, C<clGetPlatformInfo>), while this module
146while this module uses underscores as word separator and often leaves out 354uses underscores as word separator and often leaves out prefixes
147prefixes (C<< $platform->info >>). 355(C<OpenCL::platforms>, C<< $platform->info >>).
148 356
149=item * OpenCL often specifies fixed vector function arguments as short 357=item * OpenCL often specifies fixed vector function arguments as short
150arrays (C<size_t origin[3]>), while this module explicitly expects the 358arrays (C<size_t origin[3]>), while this module explicitly expects the
151components as separate arguments- 359components as separate arguments (C<$orig_x, $orig_y, $orig_z>) in
360function calls.
152 361
153=item * Where possible, the row_pitch value is calculated from the perl 362=item * Structures are often specified by flattening out their components
154scalar length and need not be specified. 363as with short vectors, and returned as arrayrefs.
155 364
156=item * When enqueuing commands, the wait list is specified by adding 365=item * When enqueuing commands, the wait list is specified by adding
157extra arguments to the function - everywhere a C<$wait_events...> argument 366extra arguments to the function - anywhere a C<$wait_events...> argument
158is documented this can be any number of event objects. 367is documented this can be any number of event objects. As an extsnion
368implemented by this module, C<undef> values will be ignored in the event
369list.
159 370
160=item * When enqueuing commands, if the enqueue method is called in void 371=item * When enqueuing commands, if the enqueue method is called in void
161context, no event is created. In all other contexts an event is returned 372context, no event is created. In all other contexts an event is returned
162by the method. 373by the method.
163 374
165other status is returned the function will throw an exception, so you 376other status is returned the function will throw an exception, so you
166don't normally have to to any error checking. 377don't normally have to to any error checking.
167 378
168=back 379=back
169 380
381=head2 PERL AND OPENCL TYPES
382
383This handy(?) table lists OpenCL types and their perl, PDL and pack/unpack
384format equivalents:
385
386 OpenCL perl PDL pack/unpack
387 char IV - c
388 uchar IV byte C
389 short IV short s
390 ushort IV ushort S
391 int IV long? l
392 uint IV - L
393 long IV longlong q
394 ulong IV - Q
395 float NV float f
396 half IV ushort S
397 double NV double d
398
399=head2 GLX SUPPORT
400
401Due to the sad state that OpenGL support is in in Perl (mostly the OpenGL
402module, which has little to no documentation and has little to no support
403for glX), this module, as a special extension, treats context creation
404properties C<OpenCL::GLX_DISPLAY_KHR> and C<OpenCL::GL_CONTEXT_KHR>
405specially: If either or both of these are C<undef>, then the OpenCL
406module tries to dynamically resolve C<glXGetCurrentDisplay> and
407C<glXGetCurrentContext>, call these functions and use their return values
408instead.
409
410For this to work, the OpenGL library must be loaded, a GLX context must
411have been created and be made current, and C<dlsym> must be available and
412capable of finding the function via C<RTLD_DEFAULT>.
413
414=head2 EVENT SYSTEM
415
416OpenCL can generate a number of (potentially) asynchronous events, for
417example, after compiling a program, to signal a context-related error or,
418perhaps most important, to signal completion of queued jobs (by setting
419callbacks on OpenCL::Event objects).
420
421To facilitate this, this module maintains an event queue - each
422time an asynchronous event happens, it is queued, and perl will be
423interrupted. This is implemented via the L<Async::Interrupt> module. In
424addition, this module has L<AnyEvent> support, so it can seamlessly
425integrate itself into many event loops.
426
427Since this module is a bit hard to understand, here are some case examples:
428
429=head3 Don't use callbacks.
430
431When your program never uses any callbacks, then there will never be any
432notifications you need to take care of, and therefore no need to worry
433about all this.
434
435You can achieve a great deal by explicitly waiting for events, or using
436barriers and flush calls. In many programs, there is no need at all to
437tinker with asynchronous events.
438
439=head3 Use AnyEvent
440
441This module automatically registers a watcher that invokes all outstanding
442event callbacks when AnyEvent is initialised (and block asynchronous
443interruptions). Using this mode of operations is the safest and most
444recommended one.
445
446To use this, simply use AnyEvent and this module normally, make sure you
447have an event loop running:
448
449 use Gtk2 -init;
450 use AnyEvent;
451
452 # initialise AnyEvent, by creating a watcher, or:
453 AnyEvent::detect;
454
455 my $e = $queue->marker;
456 $e->cb (sub {
457 warn "opencl is finished\n";
458 })
459
460 main Gtk2;
461
462Note that this module will not initialise AnyEvent for you. Before
463AnyEvent is initialised, the module will asynchronously interrupt perl
464instead. To avoid any surprises, it's best to explicitly initialise
465AnyEvent.
466
467You can temporarily enable asynchronous interruptions (see next paragraph)
468by calling C<$OpenCL::INTERRUPT->unblock> and disable them again by
469calling C<$OpenCL::INTERRUPT->block>.
470
471=head3 Let yourself be interrupted at any time
472
473This mode is the default unless AnyEvent is loaded and initialised. In
474this mode, OpenCL asynchronously interrupts a running perl program. The
475emphasis is on both I<asynchronously> and I<running> here.
476
477Asynchronously means that perl might execute your callbacks at any
478time. For example, in the following code (I<THAT YOU SHOULD NOT COPY>),
479the C<until> loop following the marker call will be interrupted by the
480callback:
481
482 my $e = $queue->marker;
483 my $flag;
484 $e->cb (sub { $flag = 1 });
485 1 until $flag;
486 # $flag is now 1
487
488The reason why you shouldn't blindly copy the above code is that
489busy waiting is a really really bad thing, and really really bad for
490performance.
491
492While at first this asynchronous business might look exciting, it can be
493really hard, because you need to be prepared for the callback code to be
494executed at any time, which limits the amount of things the callback code
495can do safely.
496
497This can be mitigated somewhat by using C<<
498$OpenCL::INTERRUPT->scope_block >> (see the L<Async::Interrupt>
499documentation for details).
500
501The other problem is that your program must be actively I<running> to be
502interrupted. When you calculate stuff, your program is running. When you
503hang in some C functions or other block execution (by calling C<sleep>,
504C<select>, running an event loop and so on), your program is waiting, not
505running.
506
507One way around that would be to attach a read watcher to your event loop,
508listening for events on C<< $OpenCL::INTERRUPT->pipe_fileno >>, using a
509dummy callback (C<sub { }>) to temporarily execute some perl code.
510
511That is then awfully close to using the built-in AnyEvent support above,
512though, so consider that one instead.
513
514=head3 Be creative
515
516OpenCL exports the L<Async::Interrupt> object it uses in the global
517variable C<$OpenCL::INTERRUPT>. You can configure it in any way you like.
518
519So if you want to feel like a real pro, err, wait, if you feel no risk
520menas no fun, you can experiment by implementing your own mode of
521operations.
522
523=cut
524
525package OpenCL;
526
527use common::sense;
528use Carp ();
529use Async::Interrupt ();
530
531our $POLL_FUNC; # set by XS
532
533BEGIN {
534 our $VERSION = '0.99';
535
536 require XSLoader;
537 XSLoader::load (__PACKAGE__, $VERSION);
538
539 @OpenCL::Platform::ISA =
540 @OpenCL::Device::ISA =
541 @OpenCL::Context::ISA =
542 @OpenCL::Queue::ISA =
543 @OpenCL::Memory::ISA =
544 @OpenCL::Sampler::ISA =
545 @OpenCL::Program::ISA =
546 @OpenCL::Kernel::ISA =
547 @OpenCL::Event::ISA = OpenCL::Object::;
548
549 @OpenCL::Buffer::ISA =
550 @OpenCL::Image::ISA = OpenCL::Memory::;
551
552 @OpenCL::BufferObj::ISA = OpenCL::Buffer::;
553
554 @OpenCL::Image2D::ISA =
555 @OpenCL::Image3D::ISA =
556 @OpenCL::Image2DArray::ISA =
557 @OpenCL::Image1D::ISA =
558 @OpenCL::Image1DArray::ISA =
559 @OpenCL::Image1DBuffer::ISA = OpenCL::Image::;
560
561 @OpenCL::UserEvent::ISA = OpenCL::Event::;
562
563 @OpenCL::MappedBuffer::ISA =
564 @OpenCL::MappedImage::ISA = OpenCL::Mapped::;
565}
566
170=head2 THE OpenCL PACKAGE 567=head2 THE OpenCL PACKAGE
171 568
172=over 4 569=over 4
173 570
174=item $int = OpenCL::errno 571=item $int = OpenCL::errno
175 572
176The last error returned by a function - it's only changed on errors. 573The last error returned by a function - it's only valid after an error occured
574and before calling another OpenCL function.
177 575
178=item $str = OpenCL::err2str $errval 576=item $str = OpenCL::err2str [$errval]
179 577
180Comverts an error value into a human readable string. 578Converts an error value into a human readable string. IF no error value is
579given, then the last error will be used (as returned by OpenCL::errno).
181 580
182=item $str = OpenCL::err2str $enum 581=item $str = OpenCL::enum2str $enum
183 582
184Converts most enum values (inof parameter names, image format constants, 583Converts most enum values (of parameter names, image format constants,
185object types, addressing and filter modes, command types etc.) into a 584object types, addressing and filter modes, command types etc.) into a
186human readbale string. When confronted with some random integer it can be 585human readable string. When confronted with some random integer it can be
187very helpful to pass it through this function to maybe get some readable 586very helpful to pass it through this function to maybe get some readable
188string out of it. 587string out of it.
189 588
190=item @platforms = OpenCL::platforms 589=item @platforms = OpenCL::platforms
191 590
192Returns all available OpenCL::Platform objects. 591Returns all available OpenCL::Platform objects.
193 592
194L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html> 593L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html>
195 594
196=item $ctx = OpenCL::context_from_type_simple $type = OpenCL::DEVICE_TYPE_DEFAULT 595=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr
197 596
198Tries to create a context from a default device and platform - never worked for me. 597Tries to create a context from a default device and platform type - never worked for me.
199 598
200L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> 599L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
201 600
601=item $ctx = OpenCL::context $properties, \@devices, $callback->($err, $pvt) = $print_stderr)
602
603Create a new OpenCL::Context object using the given device object(s). This
604function isn't implemented yet, use C<< $platform->context >> instead.
605
606L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
607
202=item OpenCL::wait_for_events $wait_events... 608=item OpenCL::wait_for_events $wait_events...
203 609
204Waits for all events to complete. 610Waits for all events to complete.
205 611
206L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> 612L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
207 613
614=item OpenCL::poll
615
616Checks if there are any outstanding events (see L<EVENT SYSTEM>) and
617invokes their callbacks.
618
619=item $OpenCL::INTERRUPT
620
621The L<Async::Interrupt> object used to signal asynchronous events (see
622L<EVENT SYSTEM>).
623
624=cut
625
626our $INTERRUPT = new Async::Interrupt c_cb => [$POLL_FUNC, 0];
627
628&_eq_initialise ($INTERRUPT->signal_func);
629
630=item $OpenCL::WATCHER
631
632The L<AnyEvent> watcher object used to watch for asynchronous events (see
633L<EVENT SYSTEM>). This variable is C<undef> until L<AnyEvent> has been
634loaded I<and> initialised (e.g. by calling C<AnyEvent::detect>).
635
636=cut
637
638our $WATCHER;
639
640sub _init_anyevent {
641 $INTERRUPT->block;
642 $WATCHER = AE::io ($INTERRUPT->pipe_fileno, 0, sub { $INTERRUPT->handle });
643}
644
645if (defined $AnyEvent::MODEL) {
646 _init_anyevent;
647} else {
648 push @AnyEvent::post_detect, \&_init_anyevent;
649}
650
208=back 651=back
209 652
653=head2 THE OpenCL::Object CLASS
654
655This is the base class for all objects in the OpenCL module. The only
656method it implements is the C<id> method, which is only useful if you want
657to interface to OpenCL on the C level.
658
659=over 4
660
661=item $iv = $obj->id
662
663OpenCL objects are represented by pointers or integers on the C level. If
664you want to interface to an OpenCL object directly on the C level, then
665you need this value, which is returned by this method. You should use an
666C<IV> type in your code and cast that to the correct type.
667
668=cut
669
670sub OpenCL::Object::id {
671 ref $_[0] eq "SCALAR"
672 ? ${ $_[0] }
673 : $_[0][0]
674}
675
676=back
677
210=head2 THE OpenCL::Platform CLASS 678=head2 THE OpenCL::Platform CLASS
211 679
212=over 4 680=over 4
213 681
682=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL)
683
684Returns a list of matching OpenCL::Device objects.
685
686=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr)
687
688Tries to create a context. Never worked for me, and you need devices explicitly anyway.
689
690L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
691
692=item $ctx = $platform->context ($properties, \@devices, $callback->($err, $pvt) = $print_stderr)
693
694Create a new OpenCL::Context object using the given device object(s)- a
695CL_CONTEXT_PLATFORM property is supplied automatically.
696
697L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
698
214=item $packed_value = $platform->info ($name) 699=item $packed_value = $platform->info ($name)
215 700
216Calls C<clGetPlatformInfo> and returns the packed, raw value - for 701Calls C<clGetPlatformInfo> and returns the packed, raw value - for
217strings, this will be the string, for other values you probably need to 702strings, this will be the string (possibly including terminating \0), for
218use the correct C<unpack>. This might get improved in the future. Hopefully. 703other values you probably need to use the correct C<unpack>.
704
705It's best to avoid this method and use one of the following convenience
706wrappers.
219 707
220L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html> 708L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html>
221 709
222=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL) 710=item $platform->unload_compiler
223 711
224Returns a list of matching OpenCL::Device objects. 712Attempts to unload the compiler for this platform, for endless
713profit. Does nothing on OpenCL 1.1.
225 714
226=item $ctx = $platform->context_from_type_simple ($type = OpenCL::DEVICE_TYPE_DEFAULT)
227
228Tries to create a context. Never worked for me.
229
230L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> 715L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clUnloadPlatformCompiler.html>
716
717=for gengetinfo begin platform
718
719=item $string = $platform->profile
720
721Calls C<clGetPlatformInfo> with C<CL_PLATFORM_PROFILE> and returns the result.
722
723=item $string = $platform->version
724
725Calls C<clGetPlatformInfo> with C<CL_PLATFORM_VERSION> and returns the result.
726
727=item $string = $platform->name
728
729Calls C<clGetPlatformInfo> with C<CL_PLATFORM_NAME> and returns the result.
730
731=item $string = $platform->vendor
732
733Calls C<clGetPlatformInfo> with C<CL_PLATFORM_VENDOR> and returns the result.
734
735=item $string = $platform->extensions
736
737Calls C<clGetPlatformInfo> with C<CL_PLATFORM_EXTENSIONS> and returns the result.
738
739=for gengetinfo end platform
231 740
232=back 741=back
233 742
234=head2 THE OpenCL::Device CLASS 743=head2 THE OpenCL::Device CLASS
235 744
239 748
240See C<< $platform->info >> for details. 749See C<< $platform->info >> for details.
241 750
242L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html> 751L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html>
243 752
753=for gengetinfo begin device
754
755=item $device_type = $device->type
756
757Calls C<clGetDeviceInfo> with C<CL_DEVICE_TYPE> and returns the result.
758
759=item $uint = $device->vendor_id
760
761Calls C<clGetDeviceInfo> with C<CL_DEVICE_VENDOR_ID> and returns the result.
762
763=item $uint = $device->max_compute_units
764
765Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_COMPUTE_UNITS> and returns the result.
766
767=item $uint = $device->max_work_item_dimensions
768
769Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS> and returns the result.
770
771=item $int = $device->max_work_group_size
772
773Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_WORK_GROUP_SIZE> and returns the result.
774
775=item @ints = $device->max_work_item_sizes
776
777Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_WORK_ITEM_SIZES> and returns the result.
778
779=item $uint = $device->preferred_vector_width_char
780
781Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_CHAR> and returns the result.
782
783=item $uint = $device->preferred_vector_width_short
784
785Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_SHORT> and returns the result.
786
787=item $uint = $device->preferred_vector_width_int
788
789Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_INT> and returns the result.
790
791=item $uint = $device->preferred_vector_width_long
792
793Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_LONG> and returns the result.
794
795=item $uint = $device->preferred_vector_width_float
796
797Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_FLOAT> and returns the result.
798
799=item $uint = $device->preferred_vector_width_double
800
801Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE> and returns the result.
802
803=item $uint = $device->max_clock_frequency
804
805Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_CLOCK_FREQUENCY> and returns the result.
806
807=item $bitfield = $device->address_bits
808
809Calls C<clGetDeviceInfo> with C<CL_DEVICE_ADDRESS_BITS> and returns the result.
810
811=item $uint = $device->max_read_image_args
812
813Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_READ_IMAGE_ARGS> and returns the result.
814
815=item $uint = $device->max_write_image_args
816
817Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_WRITE_IMAGE_ARGS> and returns the result.
818
819=item $ulong = $device->max_mem_alloc_size
820
821Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_MEM_ALLOC_SIZE> and returns the result.
822
823=item $int = $device->image2d_max_width
824
825Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE2D_MAX_WIDTH> and returns the result.
826
827=item $int = $device->image2d_max_height
828
829Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE2D_MAX_HEIGHT> and returns the result.
830
831=item $int = $device->image3d_max_width
832
833Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE3D_MAX_WIDTH> and returns the result.
834
835=item $int = $device->image3d_max_height
836
837Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE3D_MAX_HEIGHT> and returns the result.
838
839=item $int = $device->image3d_max_depth
840
841Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE3D_MAX_DEPTH> and returns the result.
842
843=item $uint = $device->image_support
844
845Calls C<clGetDeviceInfo> with C<CL_DEVICE_IMAGE_SUPPORT> and returns the result.
846
847=item $int = $device->max_parameter_size
848
849Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_PARAMETER_SIZE> and returns the result.
850
244=item $ctx = $device->context_simple 851=item $uint = $device->max_samplers
245 852
246Convenience function to create a new OpenCL::Context object. 853Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_SAMPLERS> and returns the result.
247 854
248L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html> 855=item $uint = $device->mem_base_addr_align
856
857Calls C<clGetDeviceInfo> with C<CL_DEVICE_MEM_BASE_ADDR_ALIGN> and returns the result.
858
859=item $uint = $device->min_data_type_align_size
860
861Calls C<clGetDeviceInfo> with C<CL_DEVICE_MIN_DATA_TYPE_ALIGN_SIZE> and returns the result.
862
863=item $device_fp_config = $device->single_fp_config
864
865Calls C<clGetDeviceInfo> with C<CL_DEVICE_SINGLE_FP_CONFIG> and returns the result.
866
867=item $device_mem_cache_type = $device->global_mem_cache_type
868
869Calls C<clGetDeviceInfo> with C<CL_DEVICE_GLOBAL_MEM_CACHE_TYPE> and returns the result.
870
871=item $uint = $device->global_mem_cacheline_size
872
873Calls C<clGetDeviceInfo> with C<CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE> and returns the result.
874
875=item $ulong = $device->global_mem_cache_size
876
877Calls C<clGetDeviceInfo> with C<CL_DEVICE_GLOBAL_MEM_CACHE_SIZE> and returns the result.
878
879=item $ulong = $device->global_mem_size
880
881Calls C<clGetDeviceInfo> with C<CL_DEVICE_GLOBAL_MEM_SIZE> and returns the result.
882
883=item $ulong = $device->max_constant_buffer_size
884
885Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE> and returns the result.
886
887=item $uint = $device->max_constant_args
888
889Calls C<clGetDeviceInfo> with C<CL_DEVICE_MAX_CONSTANT_ARGS> and returns the result.
890
891=item $device_local_mem_type = $device->local_mem_type
892
893Calls C<clGetDeviceInfo> with C<CL_DEVICE_LOCAL_MEM_TYPE> and returns the result.
894
895=item $ulong = $device->local_mem_size
896
897Calls C<clGetDeviceInfo> with C<CL_DEVICE_LOCAL_MEM_SIZE> and returns the result.
898
899=item $boolean = $device->error_correction_support
900
901Calls C<clGetDeviceInfo> with C<CL_DEVICE_ERROR_CORRECTION_SUPPORT> and returns the result.
902
903=item $int = $device->profiling_timer_resolution
904
905Calls C<clGetDeviceInfo> with C<CL_DEVICE_PROFILING_TIMER_RESOLUTION> and returns the result.
906
907=item $boolean = $device->endian_little
908
909Calls C<clGetDeviceInfo> with C<CL_DEVICE_ENDIAN_LITTLE> and returns the result.
910
911=item $boolean = $device->available
912
913Calls C<clGetDeviceInfo> with C<CL_DEVICE_AVAILABLE> and returns the result.
914
915=item $boolean = $device->compiler_available
916
917Calls C<clGetDeviceInfo> with C<CL_DEVICE_COMPILER_AVAILABLE> and returns the result.
918
919=item $device_exec_capabilities = $device->execution_capabilities
920
921Calls C<clGetDeviceInfo> with C<CL_DEVICE_EXECUTION_CAPABILITIES> and returns the result.
922
923=item $command_queue_properties = $device->properties
924
925Calls C<clGetDeviceInfo> with C<CL_DEVICE_QUEUE_PROPERTIES> and returns the result.
926
927=item $ = $device->platform
928
929Calls C<clGetDeviceInfo> with C<CL_DEVICE_PLATFORM> and returns the result.
930
931=item $string = $device->name
932
933Calls C<clGetDeviceInfo> with C<CL_DEVICE_NAME> and returns the result.
934
935=item $string = $device->vendor
936
937Calls C<clGetDeviceInfo> with C<CL_DEVICE_VENDOR> and returns the result.
938
939=item $string = $device->driver_version
940
941Calls C<clGetDeviceInfo> with C<CL_DRIVER_VERSION> and returns the result.
942
943=item $string = $device->profile
944
945Calls C<clGetDeviceInfo> with C<CL_DEVICE_PROFILE> and returns the result.
946
947=item $string = $device->version
948
949Calls C<clGetDeviceInfo> with C<CL_DEVICE_VERSION> and returns the result.
950
951=item $string = $device->extensions
952
953Calls C<clGetDeviceInfo> with C<CL_DEVICE_EXTENSIONS> and returns the result.
954
955=item $uint = $device->preferred_vector_width_half
956
957Calls C<clGetDeviceInfo> with C<CL_DEVICE_PREFERRED_VECTOR_WIDTH_HALF> and returns the result.
958
959=item $uint = $device->native_vector_width_char
960
961Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_CHAR> and returns the result.
962
963=item $uint = $device->native_vector_width_short
964
965Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_SHORT> and returns the result.
966
967=item $uint = $device->native_vector_width_int
968
969Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_INT> and returns the result.
970
971=item $uint = $device->native_vector_width_long
972
973Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_LONG> and returns the result.
974
975=item $uint = $device->native_vector_width_float
976
977Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_FLOAT> and returns the result.
978
979=item $uint = $device->native_vector_width_double
980
981Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE> and returns the result.
982
983=item $uint = $device->native_vector_width_half
984
985Calls C<clGetDeviceInfo> with C<CL_DEVICE_NATIVE_VECTOR_WIDTH_HALF> and returns the result.
986
987=item $device_fp_config = $device->double_fp_config
988
989Calls C<clGetDeviceInfo> with C<CL_DEVICE_DOUBLE_FP_CONFIG> and returns the result.
990
991=item $device_fp_config = $device->half_fp_config
992
993Calls C<clGetDeviceInfo> with C<CL_DEVICE_HALF_FP_CONFIG> and returns the result.
994
995=item $boolean = $device->host_unified_memory
996
997Calls C<clGetDeviceInfo> with C<CL_DEVICE_HOST_UNIFIED_MEMORY> and returns the result.
998
999=item $device = $device->parent_device_ext
1000
1001Calls C<clGetDeviceInfo> with C<CL_DEVICE_PARENT_DEVICE_EXT> and returns the result.
1002
1003=item @device_partition_property_exts = $device->partition_types_ext
1004
1005Calls C<clGetDeviceInfo> with C<CL_DEVICE_PARTITION_TYPES_EXT> and returns the result.
1006
1007=item @device_partition_property_exts = $device->affinity_domains_ext
1008
1009Calls C<clGetDeviceInfo> with C<CL_DEVICE_AFFINITY_DOMAINS_EXT> and returns the result.
1010
1011=item $uint = $device->reference_count_ext
1012
1013Calls C<clGetDeviceInfo> with C<CL_DEVICE_REFERENCE_COUNT_EXT> and returns the result.
1014
1015=item @device_partition_property_exts = $device->partition_style_ext
1016
1017Calls C<clGetDeviceInfo> with C<CL_DEVICE_PARTITION_STYLE_EXT> and returns the result.
1018
1019=for gengetinfo end device
249 1020
250=back 1021=back
251 1022
252=head2 THE OpenCL::Context CLASS 1023=head2 THE OpenCL::Context CLASS
253 1024
254=over 4 1025=over 4
255 1026
256=item $packed_value = $ctx->info ($name) 1027=item $prog = $ctx->build_program ($program, $options = "")
257 1028
258See C<< $platform->info >> for details. 1029This convenience function tries to build the program on all devices in
1030the context. If the build fails, then the function will C<croak> with the
1031build log. Otherwise ti returns the program object.
259 1032
260L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html> 1033The C<$program> can either be a C<OpenCL::Program> object or a string
1034containing the program. In the latter case, a program objetc will be
1035created automatically.
261 1036
262=item $queue = $ctx->command_queue_simple ($device) 1037=cut
263 1038
1039sub OpenCL::Context::build_program {
1040 my ($self, $prog, $options) = @_;
1041
1042 $prog = $self->program_with_source ($prog)
1043 unless ref $prog;
1044
1045 eval { $prog->build (undef, $options); 1 }
1046 or errno == BUILD_PROGRAM_FAILURE
1047 or errno == INVALID_BINARY # workaround nvidia bug
1048 or Carp::croak "OpenCL::Context->build_program: " . err2str;
1049
1050 # we check status for all devices
1051 for my $dev ($self->devices) {
1052 $prog->build_status ($dev) == BUILD_SUCCESS
1053 or Carp::croak "Building OpenCL program for device '" . $dev->name . "' failed:\n"
1054 . $prog->build_log ($dev);
1055 }
1056
1057 $prog
1058}
1059
1060=item $queue = $ctx->queue ($device, $properties)
1061
264Convenience function to create a new OpenCL::Queue object from the context and the given device. 1062Create a new OpenCL::Queue object from the context and the given device.
265 1063
266L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html> 1064L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html>
267 1065
1066Example: create an out-of-order queue.
1067
1068 $queue = $ctx->queue ($device, OpenCL::QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE);
1069
268=item $ev = $ctx->user_event 1070=item $ev = $ctx->user_event
269 1071
270Creates a new OpenCL::UserEvent object. 1072Creates a new OpenCL::UserEvent object.
271 1073
272L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html> 1074L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html>
273 1075
274=item $buf = $ctx->buffer ($flags, $len) 1076=item $buf = $ctx->buffer ($flags, $len)
275 1077
276Creates a new OpenCL::Buffer object with the given flags and octet-size. 1078Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object with the
1079given flags and octet-size.
277 1080
278L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html> 1081L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html>
279 1082
280=item $buf = $ctx->buffer_sv ($flags, $data) 1083=item $buf = $ctx->buffer_sv ($flags, $data)
281 1084
282Creates a new OpenCL::Buffer object and initialise it with the given data values. 1085Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object and
1086initialise it with the given data values.
283 1087
1088=item $img = $ctx->image ($self, $flags, $channel_order, $channel_type, $type, $width, $height, $depth = 0, $array_size = 0, $row_pitch = 0, $slice_pitch = 0, $num_mip_level = 0, $num_samples = 0, $*data = &PL_sv_undef)
1089
1090Creates a new OpenCL::Image object and optionally initialises it with
1091the given data values.
1092
1093L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage.html>
1094
284=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $data) 1095=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $row_pitch = 0, $data = undef)
285 1096
286Creates a new OpenCL::Image2D object and optionally initialises it with the given data values. 1097Creates a new OpenCL::Image2D object and optionally initialises it with
1098the given data values.
287 1099
288L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html> 1100L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html>
289 1101
290=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $slice_pitch, $data) 1102=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $row_pitch = 0, $slice_pitch = 0, $data = undef)
291 1103
292Creates a new OpenCL::Image3D object and optionally initialises it with the given data values. 1104Creates a new OpenCL::Image3D object and optionally initialises it with
1105the given data values.
293 1106
294L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html> 1107L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html>
1108
1109=item $buffer = $ctx->gl_buffer ($flags, $bufobj)
1110
1111Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object that refers to the given
1112OpenGL buffer object.
1113
1114http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLBuffer.html
1115
1116=item $img = $ctx->gl_texture ($flags, $target, $miplevel, $texture)
1117
1118Creates a new OpenCL::Image object that refers to the given OpenGL
1119texture object or buffer.
1120
1121http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture.html
1122
1123=item $img = $ctx->gl_texture2d ($flags, $target, $miplevel, $texture)
1124
1125Creates a new OpenCL::Image2D object that refers to the given OpenGL
11262D texture object.
1127
1128http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture2D.html
1129
1130=item $img = $ctx->gl_texture3d ($flags, $target, $miplevel, $texture)
1131
1132Creates a new OpenCL::Image3D object that refers to the given OpenGL
11333D texture object.
1134
1135http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture3D.html
1136
1137=item $ctx->gl_renderbuffer ($flags, $renderbuffer)
1138
1139Creates a new OpenCL::Image2D object that refers to the given OpenGL
1140render buffer.
1141
1142http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLRenderbuffer.html
295 1143
296=item @formats = $ctx->supported_image_formats ($flags, $image_type) 1144=item @formats = $ctx->supported_image_formats ($flags, $image_type)
297 1145
298Returns a list of matching image formats - each format is an arrayref with 1146Returns a list of matching image formats - each format is an arrayref with
299two values, $channel_order and $channel_type, in it. 1147two values, $channel_order and $channel_type, in it.
310 1158
311Creates a new OpenCL::Program object from the given source code. 1159Creates a new OpenCL::Program object from the given source code.
312 1160
313L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html> 1161L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html>
314 1162
1163=item ($program, \@status) = $ctx->program_with_binary (\@devices, \@binaries)
1164
1165Creates a new OpenCL::Program object from the given binaries.
1166
1167L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithBinary.html>
1168
1169Example: clone an existing program object that contains a successfully
1170compiled program, no matter how useless this is.
1171
1172 my $clone = $ctx->program_with_binary ([$prog->devices], [$prog->binaries]);
1173
1174=item $packed_value = $ctx->info ($name)
1175
1176See C<< $platform->info >> for details.
1177
1178L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html>
1179
1180=for gengetinfo begin context
1181
1182=item $uint = $context->reference_count
1183
1184Calls C<clGetContextInfo> with C<CL_CONTEXT_REFERENCE_COUNT> and returns the result.
1185
1186=item @devices = $context->devices
1187
1188Calls C<clGetContextInfo> with C<CL_CONTEXT_DEVICES> and returns the result.
1189
1190=item @property_ints = $context->properties
1191
1192Calls C<clGetContextInfo> with C<CL_CONTEXT_PROPERTIES> and returns the result.
1193
1194=item $uint = $context->num_devices
1195
1196Calls C<clGetContextInfo> with C<CL_CONTEXT_NUM_DEVICES> and returns the result.
1197
1198=for gengetinfo end context
1199
315=back 1200=back
316 1201
317=head2 THE OpenCL::Queue CLASS 1202=head2 THE OpenCL::Queue CLASS
318 1203
319An OpenCL::Queue represents an execution queue for OpenCL. You execute 1204An OpenCL::Queue represents an execution queue for OpenCL. You execute
320requests by calling their respective C<enqueue_xxx> method and waitinf for 1205requests by calling their respective method and waiting for it to complete
321it to complete in some way. 1206in some way.
322 1207
323All the enqueue methods return an event object that can be used to wait 1208Most methods that enqueue some request return an event object that can
324for completion, unless the method is called in void context, in which case 1209be used to wait for completion (optionally using a callback), unless
325no event object is created. 1210the method is called in void context, in which case no event object is
1211created.
326 1212
327They also allow you to specify any number of other event objects that this 1213They also allow you to specify any number of other event objects that this
328request has to wait for before it starts executing, by simply passing the 1214request has to wait for before it starts executing, by simply passing the
329event objects as extra parameters to the enqueue methods. 1215event objects as extra parameters to the enqueue methods. To simplify
1216program design, this module ignores any C<undef> values in the list of
1217events. This makes it possible to code operations such as this, without
1218having to put a valid event object into C<$event> first:
1219
1220 $event = $queue->xxx (..., $event);
330 1221
331Queues execute in-order by default, without any parallelism, so in most 1222Queues execute in-order by default, without any parallelism, so in most
332cases it's not necessary to wait for or create event objects. 1223cases (i.e. you use only one queue) it's not necessary to wait for or
1224create event objects, althoguh an our of order queue is often a bit
1225faster.
333 1226
334=over 4 1227=over 4
335 1228
336=item $packed_value = $ctx->info ($name)
337
338See C<< $platform->info >> for details.
339
340L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetCommandQueueInfo.html>
341
342=item $ev = $queue->enqueue_read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...) 1229=item $ev = $queue->read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...)
343 1230
344Reads data from buffer into the given string. 1231Reads data from buffer into the given string.
345 1232
346L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html> 1233L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html>
347 1234
348=item $ev = $queue->enqueue_write_buffer ($buffer, $blocking, $offset, $data, $wait_events...) 1235=item $ev = $queue->write_buffer ($buffer, $blocking, $offset, $data, $wait_events...)
349 1236
350Writes data to buffer from the given string. 1237Writes data to buffer from the given string.
351 1238
352L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html> 1239L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html>
353 1240
354=item $ev = $queue->enqueue_copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...) 1241=item $ev = $queue->copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...)
355 1242
356L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html> 1243L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html>
357 1244
1245=item $ev = $queue->read_buffer_rect (OpenCL::Memory buf, cl_bool blocking, $buf_x, $buf_y, $buf_z, $host_x, $host_y, $host_z, $width, $height, $depth, $buf_row_pitch, $buf_slice_pitch, $host_row_pitch, $host_slice_pitch, $data, $wait_events...)
1246
1247http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBufferRect.html
1248
1249=item $ev = $queue->write_buffer_rect (OpenCL::Memory buf, cl_bool blocking, $buf_x, $buf_y, $buf_z, $host_x, $host_y, $host_z, $width, $height, $depth, $buf_row_pitch, $buf_slice_pitch, $host_row_pitch, $host_slice_pitch, $data, $wait_events...)
1250
1251http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBufferRect.html
1252
1253=item $ev = $queue->copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)
1254
1255L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>
1256
358=item $ev = $queue->enqueue_read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) 1257=item $ev = $queue->read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
1258
1259C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
1260module uses the image width (and height) to supply default values.
359 1261
360L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html> 1262L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html>
361 1263
362=item $ev = $queue->enqueue_write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $data, $wait_events...) 1264=item $ev = $queue->write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
363 1265
1266C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
1267module uses the image width (and height) to supply default values.
364L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html> 1268L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html>
365 1269
1270=item $ev = $queue->copy_image ($src_image, $dst_image, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)
1271
1272L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html>
1273
1274=item $ev = $queue->copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...)
1275
1276L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html>
1277
366=item $ev = $queue->enqueue_copy_buffer_rect ($src, $dst, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $src_row_pitch, $src_slice_pitch, 4dst_row_pitch, $dst_slice_pitch, $ait_event...) 1278=item $ev = $queue->copy_buffer_rect ($src, $dst, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $src_row_pitch, $src_slice_pitch, $dst_row_pitch, $dst_slice_pitch, $wait_event...)
367 1279
368Yeah. 1280Yeah.
369 1281
370L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferRect.html>
371
372=item $ev = $queue->enqueue_copy_buffer_to_image (OpenCL::Buffer src, OpenCL::Image dst, size_t src_offset, size_t dst_x, size_t dst_y, size_t dst_z, size_t width, size_t height, size_t depth, ...)
373
374L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>. 1282L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>.
375 1283
376=item $ev = $queue->enqueue_copy_image (OpenCL::Image src, OpenCL::Buffer dst, size_t src_x, size_t src_y, size_t src_z, size_t dst_x, size_t dst_y, size_t dst_z, size_t width, size_t height, size_t depth, ...) 1284=item $ev = $queue->fill_buffer ($mem, $pattern, $offset, $size, ...)
377 1285
1286Fills the given buffer object with repeated applications of C<$pattern>,
1287starting at C<$offset> for C<$size> octets.
1288
1289L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillBuffer.html>
1290
1291=item $ev = $queue->fill_image ($img, $r, $g, $b, $a, $x, $y, $z, $width, $height, $depth, ...)
1292
1293Fills the given image area with the given rgba colour components. The
1294components are normally floating point values between C<0> and C<1>,
1295except when the image channel data type is a signe dor unsigned
1296unnormalised format, in which case the range is determined by the format.
1297
378L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html> 1298L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillImage.html>
379 1299
380=item $ev = $queue->enqueue_copy_image_to_buffer (OpenCL::Image src, OpenCL::Buffer dst, size_t src_x, size_t src_y, size_t src_z, size_t width, size_t height, size_t depth, size_t dst_offset, ...)
381
382L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html>
383
384=item $ev = $queue->enqueue_task ($kernel, $wait_events...) 1300=item $ev = $queue->task ($kernel, $wait_events...)
385 1301
386L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html> 1302L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html>
387 1303
388=item $ev = $queue->enqueue_nd_range_kernel ($kernel, @$global_work_offset, @$global_work_size, @$local_work_size, $wait_events...) 1304=item $ev = $queue->nd_range_kernel ($kernel, \@global_work_offset, \@global_work_size, \@local_work_size, $wait_events...)
389 1305
390Enqueues a kernel execution. 1306Enqueues a kernel execution.
391 1307
392@$global_work_size must be specified as a reference to an array of 1308\@global_work_size must be specified as a reference to an array of
393integers specifying the work sizes (element counts). 1309integers specifying the work sizes (element counts).
394 1310
395@$global_work_offset must be either C<undef> (in which case all offsets 1311\@global_work_offset must be either C<undef> (in which case all offsets
396are C<0>), or a reference to an array of work offsets, with the same number 1312are C<0>), or a reference to an array of work offsets, with the same number
397of elements as @$global_work_size. 1313of elements as \@global_work_size.
398 1314
399@$local_work_size must be either C<undef> (in which case the 1315\@local_work_size must be either C<undef> (in which case the
400implementation is supposed to choose good local work sizes), or a 1316implementation is supposed to choose good local work sizes), or a
401reference to an array of local work sizes, with the same number of 1317reference to an array of local work sizes, with the same number of
402elements as @$global_work_size. 1318elements as \@global_work_size.
403 1319
404L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html> 1320L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html>
405 1321
406=item $ev = $queue->enqueue_marker 1322=item $ev = $queue->acquire_gl_objects ([object, ...], $wait_events...)
407 1323
1324Enqueues a list (an array-ref of OpenCL::Memory objects) to be acquired
1325for subsequent OpenCL usage.
1326
408L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMarker.html> 1327L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueAcquireGLObjects.html>
409 1328
1329=item $ev = $queue->release_gl_objects ([object, ...], $wait_events...)
1330
1331Enqueues a list (an array-ref of OpenCL::Memory objects) to be released
1332for subsequent OpenGL usage.
1333
1334L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReleaseGLObjects.html>
1335
410=item $ev = $queue->enqueue_wait_for_events ($wait_events...) 1336=item $ev = $queue->wait_for_events ($wait_events...)
411 1337
412L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html> 1338L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html>
413 1339
414=item $queue->enqueue_barrier 1340=item $ev = $queue->marker ($wait_events...)
415 1341
1342L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMarkerWithWaitList.html>
1343
1344=item $ev = $queue->barrier ($wait_events...)
1345
416L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueBarrier.html> 1346L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueBarrierWithWaitList.html>
417 1347
418=item $queue->flush 1348=item $queue->flush
419 1349
420L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html> 1350L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html>
421 1351
422=item $queue->finish 1352=item $queue->finish
423 1353
424L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html> 1354L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html>
425 1355
1356=item $packed_value = $queue->info ($name)
1357
1358See C<< $platform->info >> for details.
1359
1360L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetCommandQueueInfo.html>
1361
1362=for gengetinfo begin command_queue
1363
1364=item $ctx = $command_queue->context
1365
1366Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_CONTEXT> and returns the result.
1367
1368=item $device = $command_queue->device
1369
1370Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_DEVICE> and returns the result.
1371
1372=item $uint = $command_queue->reference_count
1373
1374Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_REFERENCE_COUNT> and returns the result.
1375
1376=item $command_queue_properties = $command_queue->properties
1377
1378Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_PROPERTIES> and returns the result.
1379
1380=for gengetinfo end command_queue
1381
426=back 1382=back
427 1383
1384=head3 MEMORY MAPPED BUFFERS
1385
1386OpenCL allows you to map buffers and images to host memory (read: perl
1387scalars). This is done much like reading or copying a buffer, by enqueuing
1388a map or unmap operation on the command queue.
1389
1390The map operations return an C<OpenCL::Mapped> object - see L<THE
1391OpenCL::Mapped CLASS> section for details on what to do with these
1392objects.
1393
1394The object will be unmapped automatically when the mapped object is
1395destroyed (you can use a barrier to make sure the unmap has finished,
1396before using the buffer in a kernel), but you can also enqueue an unmap
1397operation manually.
1398
1399=over 4
1400
1401=item $mapped_buffer = $queue->map_buffer ($buf, $blocking=1, $map_flags=OpenCL::MAP_READ|OpenCL::MAP_WRITE, $offset=0, $size=undef, $wait_events...)
1402
1403Maps the given buffer into host memory and returns an
1404C<OpenCL::MappedBuffer> object. If C<$size> is specified as undef, then
1405the map will extend to the end of the buffer.
1406
1407L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapBuffer.html>
1408
1409Example: map the buffer $buf fully and replace the first 4 bytes by "abcd", then unmap.
1410
1411 {
1412 my $mapped = $queue->map_buffer ($buf, 1, OpenCL::MAP_WRITE);
1413 substr $$mapped, 0, 4, "abcd";
1414 } # asynchronously unmap because $mapped is destroyed
1415
1416=item $mapped_image = $queue->map_image ($img, $blocking=1, $map_flags=OpenCL::MAP_READ|OpenCL::MAP_WRITE, $x=0, $y=0, $z=0, $width=undef, $height=undef, $depth=undef, $wait_events...)
1417
1418Maps the given image area into host memory and return an
1419C<OpenCL::MappedImage> object.
1420
1421If any of C<$width>, C<$height> and/or C<$depth> are C<undef> then they
1422will be replaced by the maximum possible value.
1423
1424L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapImage.html>
1425
1426Example: map an image (with OpenCL::UNSIGNED_INT8 channel type) and set
1427the first channel of the leftmost column to 5, then explicitly unmap
1428it. You are not necessarily meant to do it this way, this example just
1429shows you the accessors to use :)
1430
1431 my $mapped = $queue->map_image ($image, 1, OpenCL::MAP_WRITE);
1432
1433 $mapped->set ($_ * $mapped->row_pitch, pack "C", 5)
1434 for 0..$image->height;
1435
1436 $mapped->unmap;.
1437 $mapped->wait; # only needed for out of order queues normally
1438
1439=item $ev = $queue->unmap ($mapped, $wait_events...)
1440
1441Unmaps the data from host memory. You must not call any methods that
1442modify the data, or modify the data scalar directly, after calling this
1443method.
1444
1445The mapped event object will always be passed as part of the
1446$wait_events. The mapped event object will be replaced by the new event
1447object that this request creates.
1448
1449=back
1450
428=head2 THE OpenCL::Memory CLASS 1451=head2 THE OpenCL::Memory CLASS
429 1452
430This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image, 1453This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image,
431OpenCL::Image2D and OpenCL::Image3D. The subclasses of this class 1454OpenCL::Image2D and OpenCL::Image3D.
432currently only exist to allow type-checking.
433 1455
434=over 4 1456=over 4
435 1457
436=item $packed_value = $memory->info ($name) 1458=item $packed_value = $memory->info ($name)
437 1459
438See C<< $platform->info >> for details. 1460See C<< $platform->info >> for details.
439 1461
440L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html> 1462L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html>
441 1463
1464=for gengetinfo begin mem
1465
1466=item $mem_object_type = $mem->type
1467
1468Calls C<clGetMemObjectInfo> with C<CL_MEM_TYPE> and returns the result.
1469
1470=item $mem_flags = $mem->flags
1471
1472Calls C<clGetMemObjectInfo> with C<CL_MEM_FLAGS> and returns the result.
1473
1474=item $int = $mem->size
1475
1476Calls C<clGetMemObjectInfo> with C<CL_MEM_SIZE> and returns the result.
1477
1478=item $ptr_value = $mem->host_ptr
1479
1480Calls C<clGetMemObjectInfo> with C<CL_MEM_HOST_PTR> and returns the result.
1481
1482=item $uint = $mem->map_count
1483
1484Calls C<clGetMemObjectInfo> with C<CL_MEM_MAP_COUNT> and returns the result.
1485
1486=item $uint = $mem->reference_count
1487
1488Calls C<clGetMemObjectInfo> with C<CL_MEM_REFERENCE_COUNT> and returns the result.
1489
1490=item $ctx = $mem->context
1491
1492Calls C<clGetMemObjectInfo> with C<CL_MEM_CONTEXT> and returns the result.
1493
1494=item $mem = $mem->associated_memobject
1495
1496Calls C<clGetMemObjectInfo> with C<CL_MEM_ASSOCIATED_MEMOBJECT> and returns the result.
1497
1498=item $int = $mem->offset
1499
1500Calls C<clGetMemObjectInfo> with C<CL_MEM_OFFSET> and returns the result.
1501
1502=for gengetinfo end mem
1503
1504=item ($type, $name) = $mem->gl_object_info
1505
1506Returns the OpenGL object type (e.g. OpenCL::GL_OBJECT_TEXTURE2D) and the
1507object "name" (e.g. the texture name) used to create this memory object.
1508
1509L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetGLObjectInfo.html>
1510
442=back 1511=back
443 1512
1513=head2 THE OpenCL::Buffer CLASS
1514
1515This is a subclass of OpenCL::Memory, and the superclass of
1516OpenCL::BufferObj. Its purpose is simply to distinguish between buffers
1517and sub-buffers.
1518
1519=head2 THE OpenCL::BufferObj CLASS
1520
1521This is a subclass of OpenCL::Buffer and thus OpenCL::Memory. It exists
1522because one cna create sub buffers of OpenLC::BufferObj objects, but not
1523sub buffers from these sub buffers.
1524
1525=over 4
1526
1527=item $subbuf = $buf_obj->sub_buffer_region ($flags, $origin, $size)
1528
1529Creates an OpenCL::Buffer objects from this buffer and returns it. The
1530C<buffer_create_type> is assumed to be C<CL_BUFFER_CREATE_TYPE_REGION>.
1531
1532L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSubBuffer.html>
1533
1534=back
1535
1536=head2 THE OpenCL::Image CLASS
1537
1538This is the superclass of all image objects - OpenCL::Image1D,
1539OpenCL::Image1DArray, OpenCL::Image1DBuffer, OpenCL::Image2D,
1540OpenCL::Image2DArray and OpenCL::Image3D.
1541
1542=over 4
1543
1544=item $packed_value = $image->image_info ($name)
1545
1546See C<< $platform->info >> for details.
1547
1548The reason this method is not called C<info> is that there already is an
1549C<< ->info >> method inherited from C<OpenCL::Memory>.
1550
1551L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetImageInfo.html>
1552
1553=item ($channel_order, $channel_data_type) = $image->format
1554
1555Returns the channel order and type used to create the image by calling
1556C<clGetImageInfo> with C<CL_IMAGE_FORMAT>.
1557
1558=for gengetinfo begin image
1559
1560=item $int = $image->element_size
1561
1562Calls C<clGetImageInfo> with C<CL_IMAGE_ELEMENT_SIZE> and returns the result.
1563
1564=item $int = $image->row_pitch
1565
1566Calls C<clGetImageInfo> with C<CL_IMAGE_ROW_PITCH> and returns the result.
1567
1568=item $int = $image->slice_pitch
1569
1570Calls C<clGetImageInfo> with C<CL_IMAGE_SLICE_PITCH> and returns the result.
1571
1572=item $int = $image->width
1573
1574Calls C<clGetImageInfo> with C<CL_IMAGE_WIDTH> and returns the result.
1575
1576=item $int = $image->height
1577
1578Calls C<clGetImageInfo> with C<CL_IMAGE_HEIGHT> and returns the result.
1579
1580=item $int = $image->depth
1581
1582Calls C<clGetImageInfo> with C<CL_IMAGE_DEPTH> and returns the result.
1583
1584=for gengetinfo end image
1585
1586=for gengetinfo begin gl_texture
1587
1588=item $GLenum = $gl_texture->target
1589
1590Calls C<clGetGLTextureInfo> with C<CL_GL_TEXTURE_TARGET> and returns the result.
1591
1592=item $GLint = $gl_texture->gl_mipmap_level
1593
1594Calls C<clGetGLTextureInfo> with C<CL_GL_MIPMAP_LEVEL> and returns the result.
1595
1596=for gengetinfo end gl_texture
1597
1598=back
1599
444=head2 THE OpenCL::Sampler CLASS 1600=head2 THE OpenCL::Sampler CLASS
445 1601
446=over 4 1602=over 4
447 1603
448=item $packed_value = $sampler->info ($name) 1604=item $packed_value = $sampler->info ($name)
449 1605
450See C<< $platform->info >> for details. 1606See C<< $platform->info >> for details.
451 1607
452L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html> 1608L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html>
453 1609
1610=for gengetinfo begin sampler
1611
1612=item $uint = $sampler->reference_count
1613
1614Calls C<clGetSamplerInfo> with C<CL_SAMPLER_REFERENCE_COUNT> and returns the result.
1615
1616=item $ctx = $sampler->context
1617
1618Calls C<clGetSamplerInfo> with C<CL_SAMPLER_CONTEXT> and returns the result.
1619
1620=item $addressing_mode = $sampler->normalized_coords
1621
1622Calls C<clGetSamplerInfo> with C<CL_SAMPLER_NORMALIZED_COORDS> and returns the result.
1623
1624=item $filter_mode = $sampler->addressing_mode
1625
1626Calls C<clGetSamplerInfo> with C<CL_SAMPLER_ADDRESSING_MODE> and returns the result.
1627
1628=item $boolean = $sampler->filter_mode
1629
1630Calls C<clGetSamplerInfo> with C<CL_SAMPLER_FILTER_MODE> and returns the result.
1631
1632=for gengetinfo end sampler
1633
454=back 1634=back
455 1635
456=head2 THE OpenCL::Program CLASS 1636=head2 THE OpenCL::Program CLASS
457 1637
458=over 4 1638=over 4
459 1639
460=item $packed_value = $program->info ($name) 1640=item $program->build (\@devices = undef, $options = "", $cb->($program) = undef)
461 1641
462See C<< $platform->info >> for details.
463
464L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html>
465
466=item $program->build ($device, $options = "")
467
468Tries to build the program with the givne options. 1642Tries to build the program with the given options. See also the
1643C<$ctx->build> convenience function.
1644
1645If a callback is specified, then it will be called when compilation is
1646finished. Note that many OpenCL implementations block your program while
1647compiling whether you use a callback or not. See C<build_async> if you
1648want to make sure the build is done in the background.
1649
1650Note that some OpenCL implementations act up badly, and don't call the
1651callback in some error cases (but call it in others). This implementation
1652assumes the callback will always be called, and leaks memory if this is
1653not so. So best make sure you don't pass in invalid values.
1654
1655Some implementations fail with C<OpenCL::INVALID_BINARY> when the
1656compilation state is successful but some later stage fails.
469 1657
470L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html> 1658L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html>
1659
1660=item $program->build_async (\@devices = undef, $options = "", $cb->($program) = undef)
1661
1662Similar to C<< ->build >>, except it starts a thread, and never fails (you
1663need to check the compilation status form the callback, or by polling).
471 1664
472=item $packed_value = $program->build_info ($device, $name) 1665=item $packed_value = $program->build_info ($device, $name)
473 1666
474Similar to C<< $platform->info >>, but returns build info for a previous 1667Similar to C<< $platform->info >>, but returns build info for a previous
475build attempt for the given device. 1668build attempt for the given device.
481Creates an OpenCL::Kernel object out of the named C<__kernel> function in 1674Creates an OpenCL::Kernel object out of the named C<__kernel> function in
482the program. 1675the program.
483 1676
484L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html> 1677L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html>
485 1678
1679=item @kernels = $program->kernels_in_program
1680
1681Returns all kernels successfully compiled for all devices in program.
1682
1683http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernelsInProgram.html
1684
1685=for gengetinfo begin program_build
1686
1687=item $build_status = $program->build_status ($device)
1688
1689Calls C<clGetProgramBuildInfo> with C<CL_PROGRAM_BUILD_STATUS> and returns the result.
1690
1691=item $string = $program->build_options ($device)
1692
1693Calls C<clGetProgramBuildInfo> with C<CL_PROGRAM_BUILD_OPTIONS> and returns the result.
1694
1695=item $string = $program->build_log ($device)
1696
1697Calls C<clGetProgramBuildInfo> with C<CL_PROGRAM_BUILD_LOG> and returns the result.
1698
1699=for gengetinfo end program_build
1700
1701=item $packed_value = $program->info ($name)
1702
1703See C<< $platform->info >> for details.
1704
1705L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html>
1706
1707=for gengetinfo begin program
1708
1709=item $uint = $program->reference_count
1710
1711Calls C<clGetProgramInfo> with C<CL_PROGRAM_REFERENCE_COUNT> and returns the result.
1712
1713=item $ctx = $program->context
1714
1715Calls C<clGetProgramInfo> with C<CL_PROGRAM_CONTEXT> and returns the result.
1716
1717=item $uint = $program->num_devices
1718
1719Calls C<clGetProgramInfo> with C<CL_PROGRAM_NUM_DEVICES> and returns the result.
1720
1721=item @devices = $program->devices
1722
1723Calls C<clGetProgramInfo> with C<CL_PROGRAM_DEVICES> and returns the result.
1724
1725=item $string = $program->source
1726
1727Calls C<clGetProgramInfo> with C<CL_PROGRAM_SOURCE> and returns the result.
1728
1729=item @ints = $program->binary_sizes
1730
1731Calls C<clGetProgramInfo> with C<CL_PROGRAM_BINARY_SIZES> and returns the result.
1732
1733=for gengetinfo end program
1734
1735=item @blobs = $program->binaries
1736
1737Returns a string for the compiled binary for every device associated with
1738the program, empty strings indicate missing programs, and an empty result
1739means no program binaries are available.
1740
1741These "binaries" are often, in fact, informative low-level assembly
1742sources.
1743
1744L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html>
1745
486=back 1746=back
487 1747
488=head2 THE OpenCL::Kernel CLASS 1748=head2 THE OpenCL::Kernel CLASS
489 1749
490=over 4 1750=over 4
493 1753
494See C<< $platform->info >> for details. 1754See C<< $platform->info >> for details.
495 1755
496L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html> 1756L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html>
497 1757
1758=for gengetinfo begin kernel
1759
1760=item $string = $kernel->function_name
1761
1762Calls C<clGetKernelInfo> with C<CL_KERNEL_FUNCTION_NAME> and returns the result.
1763
1764=item $uint = $kernel->num_args
1765
1766Calls C<clGetKernelInfo> with C<CL_KERNEL_NUM_ARGS> and returns the result.
1767
1768=item $uint = $kernel->reference_count
1769
1770Calls C<clGetKernelInfo> with C<CL_KERNEL_REFERENCE_COUNT> and returns the result.
1771
1772=item $ctx = $kernel->context
1773
1774Calls C<clGetKernelInfo> with C<CL_KERNEL_CONTEXT> and returns the result.
1775
1776=item $program = $kernel->program
1777
1778Calls C<clGetKernelInfo> with C<CL_KERNEL_PROGRAM> and returns the result.
1779
1780=for gengetinfo end kernel
1781
1782=item $packed_value = $kernel->work_group_info ($device, $name)
1783
1784See C<< $platform->info >> for details.
1785
1786The reason this method is not called C<info> is that there already is an
1787C<< ->info >> method.
1788
1789L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelWorkGroupInfo.html>
1790
1791=for gengetinfo begin kernel_work_group
1792
1793=item $int = $kernel->work_group_size ($device)
1794
1795Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_WORK_GROUP_SIZE> and returns the result.
1796
1797=item @ints = $kernel->compile_work_group_size ($device)
1798
1799Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_COMPILE_WORK_GROUP_SIZE> and returns the result.
1800
1801=item $ulong = $kernel->local_mem_size ($device)
1802
1803Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_LOCAL_MEM_SIZE> and returns the result.
1804
1805=item $int = $kernel->preferred_work_group_size_multiple ($device)
1806
1807Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE> and returns the result.
1808
1809=item $ulong = $kernel->private_mem_size ($device)
1810
1811Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_PRIVATE_MEM_SIZE> and returns the result.
1812
1813=for gengetinfo end kernel_work_group
1814
1815=item $kernel->setf ($format, ...)
1816
1817Sets the arguments of a kernel. Since OpenCL 1.1 doesn't have a generic
1818way to set arguments (and with OpenCL 1.2 it might be rather slow), you
1819need to specify a format argument, much as with C<printf>, to tell OpenCL
1820what type of argument it is.
1821
1822The format arguments are single letters:
1823
1824 c char
1825 C unsigned char
1826 s short
1827 S unsigned short
1828 i int
1829 I unsigned int
1830 l long
1831 L unsigned long
1832
1833 h half float (0..65535)
1834 f float
1835 d double
1836
1837 z local (octet size)
1838
1839 m memory object (buffer or image)
1840 a sampler
1841 e event
1842
1843Space characters in the format string are ignored.
1844
1845Example: set the arguments for a kernel that expects an int, two floats, a buffer and an image.
1846
1847 $kernel->setf ("i ff mm", 5, 0.5, 3, $buffer, $image);
1848
498=item $kernel->set_TYPE ($index, $value) 1849=item $kernel->set_TYPE ($index, $value)
499 1850
1851=item $kernel->set_char ($index, $value)
1852
1853=item $kernel->set_uchar ($index, $value)
1854
1855=item $kernel->set_short ($index, $value)
1856
1857=item $kernel->set_ushort ($index, $value)
1858
1859=item $kernel->set_int ($index, $value)
1860
1861=item $kernel->set_uint ($index, $value)
1862
1863=item $kernel->set_long ($index, $value)
1864
1865=item $kernel->set_ulong ($index, $value)
1866
1867=item $kernel->set_half ($index, $value)
1868
1869=item $kernel->set_float ($index, $value)
1870
1871=item $kernel->set_double ($index, $value)
1872
1873=item $kernel->set_memory ($index, $value)
1874
1875=item $kernel->set_buffer ($index, $value)
1876
1877=item $kernel->set_image ($index, $value)
1878
1879=item $kernel->set_sampler ($index, $value)
1880
1881=item $kernel->set_local ($index, $value)
1882
1883=item $kernel->set_event ($index, $value)
1884
500This is a family of methods to set the kernel argument with the number C<$index> to the give C<$value>. 1885This is a family of methods to set the kernel argument with the number
501 1886C<$index> to the give C<$value>.
502TYPE is one of C<char>, C<uchar>, C<short>, C<ushort>, C<int>, C<uint>,
503C<long>, C<ulong>, C<half>, C<float>, C<double>, C<memory>, C<buffer>,
504C<image2d>, C<image3d>, C<sampler> or C<event>.
505 1887
506Chars and integers (including the half type) are specified as integers, 1888Chars and integers (including the half type) are specified as integers,
507float and double as floating point values, memory/buffer/image2d/image3d 1889float and double as floating point values, memory/buffer/image must be
508must be an object of that type or C<undef>, and sampler and event must be 1890an object of that type or C<undef>, local-memory arguments are set by
509objects of that type. 1891specifying the size, and sampler and event must be objects of that type.
1892
1893Note that C<set_memory> works for all memory objects (all types of buffers
1894and images) - the main purpose of the more specific C<set_TYPE> functions
1895is type checking.
1896
1897Setting an argument for a kernel does NOT keep a reference to the object -
1898for example, if you set an argument to some image object, free the image,
1899and call the kernel, you will run into undefined behaviour.
510 1900
511L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html> 1901L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html>
512 1902
513=back 1903=back
514 1904
517This is the superclass for all event objects (including OpenCL::UserEvent 1907This is the superclass for all event objects (including OpenCL::UserEvent
518objects). 1908objects).
519 1909
520=over 4 1910=over 4
521 1911
1912=item $ev->wait
1913
1914Waits for the event to complete.
1915
1916L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
1917
1918=item $ev->cb ($exec_callback_type, $callback->($event, $event_command_exec_status))
1919
1920Adds a callback to the callback stack for the given event type. There is
1921no way to remove a callback again.
1922
1923L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetEventCallback.html>
1924
522=item $packed_value = $ev->info ($name) 1925=item $packed_value = $ev->info ($name)
523 1926
524See C<< $platform->info >> for details. 1927See C<< $platform->info >> for details.
525 1928
526L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html> 1929L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html>
527 1930
528=item $ev->wait 1931=for gengetinfo begin event
529 1932
530Waits for the event to complete. 1933=item $queue = $event->command_queue
531 1934
1935Calls C<clGetEventInfo> with C<CL_EVENT_COMMAND_QUEUE> and returns the result.
1936
1937=item $command_type = $event->command_type
1938
1939Calls C<clGetEventInfo> with C<CL_EVENT_COMMAND_TYPE> and returns the result.
1940
1941=item $uint = $event->reference_count
1942
1943Calls C<clGetEventInfo> with C<CL_EVENT_REFERENCE_COUNT> and returns the result.
1944
1945=item $uint = $event->command_execution_status
1946
1947Calls C<clGetEventInfo> with C<CL_EVENT_COMMAND_EXECUTION_STATUS> and returns the result.
1948
1949=item $ctx = $event->context
1950
1951Calls C<clGetEventInfo> with C<CL_EVENT_CONTEXT> and returns the result.
1952
1953=for gengetinfo end event
1954
1955=item $packed_value = $ev->profiling_info ($name)
1956
1957See C<< $platform->info >> for details.
1958
1959The reason this method is not called C<info> is that there already is an
1960C<< ->info >> method.
1961
532L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> 1962L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProfilingInfo.html>
1963
1964=for gengetinfo begin profiling
1965
1966=item $ulong = $event->profiling_command_queued
1967
1968Calls C<clGetEventProfilingInfo> with C<CL_PROFILING_COMMAND_QUEUED> and returns the result.
1969
1970=item $ulong = $event->profiling_command_submit
1971
1972Calls C<clGetEventProfilingInfo> with C<CL_PROFILING_COMMAND_SUBMIT> and returns the result.
1973
1974=item $ulong = $event->profiling_command_start
1975
1976Calls C<clGetEventProfilingInfo> with C<CL_PROFILING_COMMAND_START> and returns the result.
1977
1978=item $ulong = $event->profiling_command_end
1979
1980Calls C<clGetEventProfilingInfo> with C<CL_PROFILING_COMMAND_END> and returns the result.
1981
1982=for gengetinfo end profiling
533 1983
534=back 1984=back
535 1985
536=head2 THE OpenCL::UserEvent CLASS 1986=head2 THE OpenCL::UserEvent CLASS
537 1987
539 1989
540=over 4 1990=over 4
541 1991
542=item $ev->set_status ($execution_status) 1992=item $ev->set_status ($execution_status)
543 1993
1994Sets the execution status of the user event. Can only be called once,
1995either with OpenCL::COMPLETE or a negative number as status.
1996
544L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html> 1997L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html>
545 1998
546=back 1999=back
547 2000
2001=head2 THE OpenCL::Mapped CLASS
2002
2003This class represents objects mapped into host memory. They are
2004represented by a blessed string scalar. The string data is the mapped
2005memory area, that is, if you read or write it, then the mapped object is
2006accessed directly.
2007
2008You must only ever use operations that modify the string in-place - for
2009example, a C<substr> that doesn't change the length, or maybe a regex that
2010doesn't change the length. Any other operation might cause the data to be
2011copied.
2012
2013When the object is destroyed it will enqueue an implicit unmap operation
2014on the queue that was used to create it.
2015
2016Keep in mind that you I<need> to unmap (or destroy) mapped objects before
2017OpenCL sees the changes, even if some implementations don't need this
2018sometimes.
2019
2020Example, replace the first two floats in the mapped buffer by 1 and 2.
2021
2022 my $mapped = $queue->map_buffer ($buf, ...
2023 $mapped->event->wait; # make sure it's there
2024
2025 # now replace first 8 bytes by new data, which is exactly 8 bytes long
2026 # we blindly assume device endianness to equal host endianness
2027 # (and of course, we assume iee 754 single precision floats :)
2028 substr $$mapped, 0, 8, pack "f*", 1, 2;
2029
2030=over 4
2031
2032=item $ev = $mapped->unmap ($wait_events...)
2033
2034Unmaps the mapped memory object, using the queue originally used to create
2035it, quite similarly to C<< $queue->unmap ($mapped, ...) >>.
2036
2037=item $bool = $mapped->mapped
2038
2039Returns whether the object is still mapped - true before an C<unmap> is
2040enqueued, false afterwards.
2041
2042=item $ev = $mapped->event
2043
2044Return the event object associated with the mapped object. Initially, this
2045will be the event object created when mapping the object, and after an
2046unmap, this will be the event object that the unmap operation created.
2047
2048=item $mapped->wait
2049
2050Same as C<< $mapped->event->wait >> - makes sure no operations on this
2051mapped object are outstanding.
2052
2053=item $bytes = $mapped->size
2054
2055Returns the size of the mapped area, in bytes. Same as C<length $$mapped>.
2056
2057=item $ptr = $mapped->ptr
2058
2059Returns the raw memory address of the mapped area.
2060
2061=item $mapped->set ($offset, $data)
2062
2063Replaces the data at the given C<$offset> in the memory area by the new
2064C<$data>. This method is safer than direct manipulation of C<$mapped>
2065because it does bounds-checking, but also slower.
2066
2067=item $data = $mapped->get ($offset, $length)
2068
2069Returns (without copying) a scalar representing the data at the given
2070C<$offset> and C<$length> in the mapped memory area. This is the same as
2071the following substr, except much slower;
2072
2073 $data = substr $$mapped, $offset, $length
2074
548=cut 2075=cut
549 2076
550package OpenCL; 2077sub OpenCL::Mapped::get {
551 2078 substr ${$_[0]}, $_[1], $_[2]
552use common::sense;
553
554BEGIN {
555 our $VERSION = '0.03';
556
557 require XSLoader;
558 XSLoader::load (__PACKAGE__, $VERSION);
559
560 @OpenCL::Buffer::ISA =
561 @OpenCL::Image::ISA = OpenCL::Memory::;
562
563 @OpenCL::Image2D::ISA =
564 @OpenCL::Image3D::ISA = OpenCL::Image::;
565
566 @OpenCL::UserEvent::ISA = OpenCL::Event::;
567} 2079}
2080
2081=back
2082
2083=head2 THE OpenCL::MappedBuffer CLASS
2084
2085This is a subclass of OpenCL::Mapped, representing mapped buffers.
2086
2087=head2 THE OpenCL::MappedImage CLASS
2088
2089This is a subclass of OpenCL::Mapped, representing mapped images.
2090
2091=over 4
2092
2093=item $bytes = $mapped->row_pitch
2094
2095=item $bytes = $mapped->slice_pitch
2096
2097Return the row or slice pitch of the image that has been mapped.
2098
2099=back
2100
2101
2102=cut
568 2103
5691; 21041;
570 2105
571=head1 AUTHOR 2106=head1 AUTHOR
572 2107

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines