[ViewVC] Diff of: cvs/OpenCL/OpenCL.pm

Comparing OpenCL/OpenCL.pm (file contents):
Revision 1.36 by root, Thu Apr 19 14:34:56 2012 UTC vs.
Revision 1.68 by root, Tue May 1 22:25:13 2012 UTC

…		…
43		43
44	OpenCL::Event objects are used to signal when something is complete.	44	OpenCL::Event objects are used to signal when something is complete.
45		45
46	=head2 HELPFUL RESOURCES	46	=head2 HELPFUL RESOURCES
47		47
48	The OpenCL spec used to develop this module (1.2 spec was available, but	48	The OpenCL specs used to develop this module:
49	no implementation was available to me :).
50		49
51	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
52		53
53	OpenCL manpages:	54	OpenCL manpages:
54		55
55	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/
56		58
57	If you are into UML class diagrams, the following diagram might help - if	59	If you are into UML class diagrams, the following diagram might help - if
58	not, it will be mildly cobfusing:	60	not, it will be mildly confusing (also, the class hierarchy of this module
		61	is much more fine-grained):
59		62
60	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/classDiagram.html	63	http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/classDiagram.html
61		64
62	Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it	65	Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it
63	is, but at least it's free of charge:	66	is, but at least it's free of charge:
64		67
65	http://developer.amd.com/zones/OpenCLZone/courses/Documents/Introduction_to_OpenCL_Programming%20Training_Guide%20%28201005%29.pdf	68	http://developer.amd.com/zones/OpenCLZone/courses/Documents/Introduction_to_OpenCL_Programming%20Training_Guide%20%28201005%29.pdf
…		…
138	=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,
139	then asynchronously.	142	then asynchronously.
140		143
141	my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut");	144	my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut");
142		145
143	$queue->enqueue_read_buffer ($buf, 1, 1, 3, my $data);	146	$queue->read_buffer ($buf, 1, 1, 3, my $data);
144	print "$data\n";	147	print "$data\n";
145		148
146	my $ev = $queue->enqueue_read_buffer ($buf, 0, 1, 3, my $data);	149	my $ev = $queue->read_buffer ($buf, 0, 1, 3, my $data);
147	$ev->wait;	150	$ev->wait;
148	print "$data\n"; # prints "elm"	151	print "$data\n"; # prints "elm"
149		152
150	=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
151	functions.	154	functions.
…		…
157	$id = get_global_id (0);	160	$id = get_global_id (0);
158	output [id] = input [id] * input [id];	161	output [id] = input [id] * input [id];
159	}	162	}
160	';	163	';
161		164
162	my $prog = $ctx->program_with_source ($src);	165	my $prog = $ctx->build_program ($src);
163
164	# build croaks on compile errors, so catch it and print the compile errors
165	eval { $prog->build ($dev); 1 }
166	or die $prog->build_log;
167
168	my $kernel = $prog->kernel ("squareit");	166	my $kernel = $prog->kernel ("squareit");
169		167
170	=head2 Create some input and output float buffers, then call the	168	=head2 Create some input and output float buffers, then call the
171	'squareit' kernel on them.	169	'squareit' kernel on them.
172		170
…		…
176	# set buffer	174	# set buffer
177	$kernel->set_buffer (0, $input);	175	$kernel->set_buffer (0, $input);
178	$kernel->set_buffer (1, $output);	176	$kernel->set_buffer (1, $output);
179		177
180	# execute it for all 4 numbers	178	# execute it for all 4 numbers
181	$queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef);	179	$queue->nd_range_kernel ($kernel, undef, [4], undef);
182		180
183	# enqueue a synchronous read	181	# enqueue a synchronous read
184	$queue->enqueue_read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);	182	$queue->read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
185		183
186	# print the results:	184	# print the results:
187	printf "%s\n", join ", ", unpack "f*", $data;	185	printf "%s\n", join ", ", unpack "f*", $data;
188		186
189	=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,
190	showing off barriers.	188	showing off barriers.
191		189
192	# execute it for all 4 numbers	190	# execute it for all 4 numbers
193	$queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef);	191	$queue->nd_range_kernel ($kernel, undef, [4], undef);
194		192
195	# enqueue a barrier to ensure in-order execution	193	# enqueue a barrier to ensure in-order execution
196	$queue->enqueue_barrier;	194	$queue->barrier;
197		195
198	# enqueue an async read	196	# enqueue an async read
199	$queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);	197	$queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
200		198
201	# wait for all requests to finish	199	# wait for all requests to finish
202	$queue->finish;	200	$queue->finish;
203		201
204	=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,
205	showing off event objects and wait lists.	203	showing off event objects and wait lists.
206		204
207	# execute it for all 4 numbers	205	# execute it for all 4 numbers
208	my $ev = $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef);	206	my $ev = $queue->nd_range_kernel ($kernel, undef, [4], undef);
209		207
210	# enqueue an async read	208	# enqueue an async read
211	$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);
212		210
213	# wait for the last event to complete	211	# wait for the last event to complete
214	$ev->wait;	212	$ev->wait;
		213
		214	=head2 Use the OpenGL module to share a texture between OpenCL and OpenGL and draw some julia
		215	set tunnel effect.
		216
		217	This is quite a long example to get you going - you can download it from
		218	L<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
		312	For those poor souls with only a sucky CPU OpenCL implementation, you
		313	currently have to read the image into some perl scalar, and then modify a
		314	texture or use glDrawPixels or so).
		315
		316	First, when you don't need gl sharing, you can create the context much simpler:
		317
		318	$ctx = $platform->context (undef, [$dev])
		319
		320	To use a texture, you would modify the above example by creating an
		321	OpenCL::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
		325	And in the darw loop, intead of acquire_gl_objects/release_gl_objects, you
		326	would 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
		330	And 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, $data;
		333
		334	The fully modified example can be found at
		335	L<http://cvs.schmorp.de/OpenCL/examples/juliaflight-nosharing>.
215		336
216	=head1 DOCUMENTATION	337	=head1 DOCUMENTATION
217		338
218	=head2 BASIC CONVENTIONS	339	=head2 BASIC CONVENTIONS
219		340
…		…
241	=item * Structures are often specified by flattening out their components	362	=item * Structures are often specified by flattening out their components
242	as with short vectors, and returned as arrayrefs.	363	as with short vectors, and returned as arrayrefs.
243		364
244	=item * When enqueuing commands, the wait list is specified by adding	365	=item * When enqueuing commands, the wait list is specified by adding
245	extra arguments to the function - anywhere a C<$wait_events...> argument	366	extra arguments to the function - anywhere a C<$wait_events...> argument
246	is documented this can be any number of event objects.	367	is documented this can be any number of event objects. As an extsnion
		368	implemented by this module, C<undef> values will be ignored in the event
		369	list.
247		370
248	=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
249	context, no event is created. In all other contexts an event is returned	372	context, no event is created. In all other contexts an event is returned
250	by the method.	373	by the method.
251		374
…		…
275		398
276	=head2 GLX SUPPORT	399	=head2 GLX SUPPORT
277		400
278	Due to the sad state that OpenGL support is in in Perl (mostly the OpenGL	401	Due to the sad state that OpenGL support is in in Perl (mostly the OpenGL
279	module, which has little to no documentation and has little to no support	402	module, which has little to no documentation and has little to no support
280	for glx), this module, as a special extension, treats context creation	403	for glX), this module, as a special extension, treats context creation
281	properties C<OpenCL::GLX_DISPLAY_KHR> and C<OpenCL::GL_CONTEXT_KHR>	404	properties C<OpenCL::GLX_DISPLAY_KHR> and C<OpenCL::GL_CONTEXT_KHR>
282	specially: If either or both of these are C<undef>, then the OpenCL	405	specially: If either or both of these are C<undef>, then the OpenCL
283	module tries to dynamically resolve C<glxGetCurrentDisplay> and	406	module tries to dynamically resolve C<glXGetCurrentDisplay> and
284	C<glxGetCurrentContext>, call these functions and use their return values	407	C<glXGetCurrentContext>, call these functions and use their return values
285	instead.	408	instead.
286		409
287	For this to work, the OpenGL library must be loaded, a GLX context must	410	For this to work, the OpenGL library must be loaded, a GLX context must
288	have been created and be made current, and C<dlsym> must be available and	411	have been created and be made current, and C<dlsym> must be available and
289	capable of finding the function via C<RTLD_DEFAULT>.	412	capable of finding the function via C<RTLD_DEFAULT>.
290		413
		414	=head2 EVENT SYSTEM
		415
		416	OpenCL can generate a number of (potentially) asynchronous events, for
		417	example, after compiling a program, to signal a context-related error or,
		418	perhaps most important, to signal completion of queued jobs (by setting
		419	callbacks on OpenCL::Event objects).
		420
		421	To facilitate this, this module maintains an event queue - each
		422	time an asynchronous event happens, it is queued, and perl will be
		423	interrupted. This is implemented via the L<Async::Interrupt> module. In
		424	addition, this module has L<AnyEvent> support, so it can seamlessly
		425	integrate itself into many event loops.
		426
		427	Since this module is a bit hard to understand, here are some case examples:
		428
		429	=head3 Don't use callbacks.
		430
		431	When your program never uses any callbacks, then there will never be any
		432	notifications you need to take care of, and therefore no need to worry
		433	about all this.
		434
		435	You can achieve a great deal by explicitly waiting for events, or using
		436	barriers and flush calls. In many programs, there is no need at all to
		437	tinker with asynchronous events.
		438
		439	=head3 Use AnyEvent
		440
		441	This module automatically registers a watcher that invokes all outstanding
		442	event callbacks when AnyEvent is initialised (and block asynchronous
		443	interruptions). Using this mode of operations is the safest and most
		444	recommended one.
		445
		446	To use this, simply use AnyEvent and this module normally, make sure you
		447	have 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	$e->cb (sub {
		457	warn "opencl is finished\n";
		458	})
		459
		460	main Gtk2;
		461
		462	Note that this module will not initialise AnyEvent for you. Before
		463	AnyEvent is initialised, the module will asynchronously interrupt perl
		464	instead. To avoid any surprises, it's best to explicitly initialise
		465	AnyEvent.
		466
		467	You can temporarily enable asynchronous interruptions (see next paragraph)
		468	by calling C<$OpenCL::INTERRUPT->unblock> and disable them again by
		469	calling C<$OpenCL::INTERRUPT->block>.
		470
		471	=head3 Let yourself be interrupted at any time
		472
		473	This mode is the default unless AnyEvent is loaded and initialised. In
		474	this mode, OpenCL asynchronously interrupts a running perl program. The
		475	emphasis is on both I<asynchronously> and I<running> here.
		476
		477	Asynchronously means that perl might execute your callbacks at any
		478	time. For example, in the following code (I<THAT YOU SHOULD NOT COPY>),
		479	the C<until> loop following the marker call will be interrupted by the
		480	callback:
		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
		488	The reason why you shouldn't blindly copy the above code is that
		489	busy waiting is a really really bad thing, and really really bad for
		490	performance.
		491
		492	While at first this asynchronous business might look exciting, it can be
		493	really hard, because you need to be prepared for the callback code to be
		494	executed at any time, which limits the amount of things the callback code
		495	can do safely.
		496
		497	This can be mitigated somewhat by using C<<
		498	$OpenCL::INTERRUPT->scope_block >> (see the L<Async::Interrupt>
		499	documentation for details).
		500
		501	The other problem is that your program must be actively I<running> to be
		502	interrupted. When you calculate stuff, your program is running. When you
		503	hang in some C functions or other block execution (by calling C<sleep>,
		504	C<select>, running an event loop and so on), your program is waiting, not
		505	running.
		506
		507	One way around that would be to attach a read watcher to your event loop,
		508	listening for events on C<< $OpenCL::INTERRUPT->pipe_fileno >>, using a
		509	dummy callback (C<sub { }>) to temporarily execute some perl code.
		510
		511	That is then awfully close to using the built-in AnyEvent support above,
		512	though, so consider that one instead.
		513
		514	=head3 Be creative
		515
		516	OpenCL exports the L<Async::Interrupt> object it uses in the global
		517	variable C<$OpenCL::INTERRUPT>. You can configure it in any way you like.
		518
		519	So if you want to feel like a real pro, err, wait, if you feel no risk
		520	menas no fun, you can experiment by implementing your own mode of
		521	operations.
		522
		523	=cut
		524
		525	package OpenCL;
		526
		527	use common::sense;
		528	use Carp ();
		529	use Async::Interrupt ();
		530
		531	our $POLL_FUNC; # set by XS
		532
		533	BEGIN {
		534	our $VERSION = '0.98';
		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
291	=head2 THE OpenCL PACKAGE	567	=head2 THE OpenCL PACKAGE
292		568
293	=over 4	569	=over 4
294		570
295	=item $int = OpenCL::errno	571	=item $int = OpenCL::errno
296		572
297	The last error returned by a function - it's only valid after an error occured	573	The last error returned by a function - it's only valid after an error occured
298	and before calling another OpenCL function.	574	and before calling another OpenCL function.
299		575
300	=item $str = OpenCL::err2str $errval	576	=item $str = OpenCL::err2str [$errval]
301		577
302	Comverts an error value into a human readable string.	578	Converts an error value into a human readable string. IF no error value is
		579	given, then the last error will be used (as returned by OpenCL::errno).
303		580
304	=item $str = OpenCL::enum2str $enum	581	=item $str = OpenCL::enum2str $enum
305		582
306	Converts most enum values (of parameter names, image format constants,	583	Converts most enum values (of parameter names, image format constants,
307	object types, addressing and filter modes, command types etc.) into a	584	object types, addressing and filter modes, command types etc.) into a
…		…
313		590
314	Returns all available OpenCL::Platform objects.	591	Returns all available OpenCL::Platform objects.
315		592
316	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html>	593	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html>
317		594
318	=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $notify = undef	595	=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr
319		596
320	Tries to create a context from a default device and platform - never worked for me.	597	Tries to create a context from a default device and platform type - never worked for me.
321		598
322	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>	599	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
323		600
		601	=item $ctx = OpenCL::context $properties, \@devices, $callback->($err, $pvt) = $print_stderr)
		602
		603	Create a new OpenCL::Context object using the given device object(s). This
		604	function isn't implemented yet, use C<< $platform->context >> instead.
		605
		606	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
		607
324	=item OpenCL::wait_for_events $wait_events...	608	=item OpenCL::wait_for_events $wait_events...
325		609
326	Waits for all events to complete.	610	Waits for all events to complete.
327		611
328	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>	612	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
329		613
		614	=item OpenCL::poll
		615
		616	Checks if there are any outstanding events (see L<EVENT SYSTEM>) and
		617	invokes their callbacks.
		618
		619	=item $OpenCL::INTERRUPT
		620
		621	The L<Async::Interrupt> object used to signal asynchronous events (see
		622	L<EVENT SYSTEM>).
		623
		624	=cut
		625
		626	our $INTERRUPT = new Async::Interrupt c_cb => [$POLL_FUNC, 0];
		627
		628	&_eq_initialise ($INTERRUPT->signal_func);
		629
		630	=item $OpenCL::WATCHER
		631
		632	The L<AnyEvent> watcher object used to watch for asynchronous events (see
		633	L<EVENT SYSTEM>). This variable is C<undef> until L<AnyEvent> has been
		634	loaded I<and> initialised (e.g. by calling C<AnyEvent::detect>).
		635
		636	=cut
		637
		638	our $WATCHER;
		639
		640	sub _init_anyevent {
		641	$INTERRUPT->block;
		642	$WATCHER = AE::io ($INTERRUPT->pipe_fileno, 0, sub { $INTERRUPT->handle });
		643	}
		644
		645	if (defined $AnyEvent::MODEL) {
		646	_init_anyevent;
		647	} else {
		648	push @AnyEvent::post_detect, \&_init_anyevent;
		649	}
		650
330	=back	651	=back
331		652
		653	=head2 THE OpenCL::Object CLASS
		654
		655	This is the base class for all objects in the OpenCL module. The only
		656	method it implements is the C<id> method, which is only useful if you want
		657	to interface to OpenCL on the C level.
		658
		659	=over 4
		660
		661	=item $iv = $obj->id
		662
		663	OpenCL objects are represented by pointers or integers on the C level. If
		664	you want to interface to an OpenCL object directly on the C level, then
		665	you need this value, which is returned by this method. You should use an
		666	C<IV> type in your code and cast that to the correct type.
		667
		668	=cut
		669
		670	sub OpenCL::Object::id {
		671	ref $_[0] eq "SCALAR"
		672	? ${ $_[0] }
		673	: $_[0][0]
		674	}
		675
		676	=back
		677
332	=head2 THE OpenCL::Platform CLASS	678	=head2 THE OpenCL::Platform CLASS
333		679
334	=over 4	680	=over 4
335		681
336	=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL)	682	=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL)
337		683
338	Returns a list of matching OpenCL::Device objects.	684	Returns a list of matching OpenCL::Device objects.
339		685
340	=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $notify = undef)	686	=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr)
341		687
342	Tries to create a context. Never worked for me, and you need devices explicitly anyway.	688	Tries to create a context. Never worked for me, and you need devices explicitly anyway.
343		689
344	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>	690	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
345		691
346	=item $ctx = $platform->context ($properties = undef, @$devices, $notify = undef)	692	=item $ctx = $platform->context ($properties, \@devices, $callback->($err, $pvt) = $print_stderr)
347		693
348	Create a new OpenCL::Context object using the given device object(s)- a	694	Create a new OpenCL::Context object using the given device object(s)- a
349	CL_CONTEXT_PLATFORM property is supplied automatically.	695	CL_CONTEXT_PLATFORM property is supplied automatically.
350		696
351	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>	697	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
…		…
359	It's best to avoid this method and use one of the following convenience	705	It's best to avoid this method and use one of the following convenience
360	wrappers.	706	wrappers.
361		707
362	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html>	708	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html>
363		709
		710	=item $platform->unload_compiler
		711
		712	Attempts to unload the compiler for this platform, for endless
		713	profit. Does nothing on OpenCL 1.1.
		714
		715	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clUnloadPlatformCompiler.html>
		716
364	=for gengetinfo begin platform	717	=for gengetinfo begin platform
365		718
366	=item $string = $platform->profile	719	=item $string = $platform->profile
367		720
368	Calls C<clGetPlatformInfo> with C<CL_PLATFORM_PROFILE> and returns the result.	721	Calls C<clGetPlatformInfo> with C<CL_PLATFORM_PROFILE> and returns the result.
…		…
653		1006
654	=item @device_partition_property_exts = $device->affinity_domains_ext	1007	=item @device_partition_property_exts = $device->affinity_domains_ext
655		1008
656	Calls C<clGetDeviceInfo> with C<CL_DEVICE_AFFINITY_DOMAINS_EXT> and returns the result.	1009	Calls C<clGetDeviceInfo> with C<CL_DEVICE_AFFINITY_DOMAINS_EXT> and returns the result.
657		1010
658	=item $uint = $device->reference_count_ext	1011	=item $uint = $device->reference_count_ext
659		1012
660	Calls C<clGetDeviceInfo> with C<CL_DEVICE_REFERENCE_COUNT_EXT > and returns the result.	1013	Calls C<clGetDeviceInfo> with C<CL_DEVICE_REFERENCE_COUNT_EXT> and returns the result.
661		1014
662	=item @device_partition_property_exts = $device->partition_style_ext	1015	=item @device_partition_property_exts = $device->partition_style_ext
663		1016
664	Calls C<clGetDeviceInfo> with C<CL_DEVICE_PARTITION_STYLE_EXT> and returns the result.	1017	Calls C<clGetDeviceInfo> with C<CL_DEVICE_PARTITION_STYLE_EXT> and returns the result.
665		1018
…		…
669		1022
670	=head2 THE OpenCL::Context CLASS	1023	=head2 THE OpenCL::Context CLASS
671		1024
672	=over 4	1025	=over 4
673		1026
		1027	=item $prog = $ctx->build_program ($program, $options = "")
		1028
		1029	This convenience function tries to build the program on all devices in
		1030	the context. If the build fails, then the function will C<croak> with the
		1031	build log. Otherwise ti returns the program object.
		1032
		1033	The C<$program> can either be a C<OpenCL::Program> object or a string
		1034	containing the program. In the latter case, a program objetc will be
		1035	created automatically.
		1036
		1037	=cut
		1038
		1039	sub 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
674	=item $queue = $ctx->queue ($device, $properties)	1060	=item $queue = $ctx->queue ($device, $properties)
675		1061
676	Create a new OpenCL::Queue object from the context and the given device.	1062	Create a new OpenCL::Queue object from the context and the given device.
677		1063
678	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html>	1064	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html>
		1065
		1066	Example: create an out-of-order queue.
		1067
		1068	$queue = $ctx->queue ($device, OpenCL::QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE);
679		1069
680	=item $ev = $ctx->user_event	1070	=item $ev = $ctx->user_event
681		1071
682	Creates a new OpenCL::UserEvent object.	1072	Creates a new OpenCL::UserEvent object.
683		1073
…		…
693	=item $buf = $ctx->buffer_sv ($flags, $data)	1083	=item $buf = $ctx->buffer_sv ($flags, $data)
694		1084
695	Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object and	1085	Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object and
696	initialise it with the given data values.	1086	initialise it with the given data values.
697		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
		1090	Creates a new OpenCL::Image object and optionally initialises it with
		1091	the given data values.
		1092
		1093	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage.html>
		1094
698	=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $row_pitch = 0, $data = undef)	1095	=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $row_pitch = 0, $data = undef)
699		1096
700	Creates a new OpenCL::Image2D object and optionally initialises it with	1097	Creates a new OpenCL::Image2D object and optionally initialises it with
701	the given data values.	1098	the given data values.
702		1099
…		…
714	Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object that refers to the given	1111	Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object that refers to the given
715	OpenGL buffer object.	1112	OpenGL buffer object.
716		1113
717	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLBuffer.html	1114	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLBuffer.html
718		1115
		1116	=item $img = $ctx->gl_texture ($flags, $target, $miplevel, $texture)
		1117
		1118	Creates a new OpenCL::Image object that refers to the given OpenGL
		1119	texture object or buffer.
		1120
		1121	http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture.html
		1122
719	=item $ctx->gl_texture2d ($flags, $target, $miplevel, $texture)	1123	=item $img = $ctx->gl_texture2d ($flags, $target, $miplevel, $texture)
720		1124
721	Creates a new OpenCL::Image2D object that refers to the given OpenGL	1125	Creates a new OpenCL::Image2D object that refers to the given OpenGL
722	2D texture object.	1126	2D texture object.
723		1127
724	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture2D.html	1128	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture2D.html
725		1129
726	=item $ctx->gl_texture3d ($flags, $target, $miplevel, $texture)	1130	=item $img = $ctx->gl_texture3d ($flags, $target, $miplevel, $texture)
727		1131
728	Creates a new OpenCL::Image3D object that refers to the given OpenGL	1132	Creates a new OpenCL::Image3D object that refers to the given OpenGL
729	3D texture object.	1133	3D texture object.
730		1134
731	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture3D.html	1135	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture3D.html
…		…
785	=back	1189	=back
786		1190
787	=head2 THE OpenCL::Queue CLASS	1191	=head2 THE OpenCL::Queue CLASS
788		1192
789	An OpenCL::Queue represents an execution queue for OpenCL. You execute	1193	An OpenCL::Queue represents an execution queue for OpenCL. You execute
790	requests by calling their respective C<enqueue_xxx> method and waitinf for	1194	requests by calling their respective method and waiting for it to complete
791	it to complete in some way.	1195	in some way.
792		1196
793	All the enqueue methods return an event object that can be used to wait	1197	Most methods that enqueue some request return an event object that can
794	for completion, unless the method is called in void context, in which case	1198	be used to wait for completion (optionally using a callback), unless
795	no event object is created.	1199	the method is called in void context, in which case no event object is
		1200	created.
796		1201
797	They also allow you to specify any number of other event objects that this	1202	They also allow you to specify any number of other event objects that this
798	request has to wait for before it starts executing, by simply passing the	1203	request has to wait for before it starts executing, by simply passing the
799	event objects as extra parameters to the enqueue methods.	1204	event objects as extra parameters to the enqueue methods. To simplify
		1205	program design, this module ignores any C<undef> values in the list of
		1206	events. This makes it possible to code operations such as this, without
		1207	having to put a valid event object into C<$event> first:
		1208
		1209	$event = $queue->xxx (..., $event);
800		1210
801	Queues execute in-order by default, without any parallelism, so in most	1211	Queues execute in-order by default, without any parallelism, so in most
802	cases (i.e. you use only one queue) it's not necessary to wait for or	1212	cases (i.e. you use only one queue) it's not necessary to wait for or
803	create event objects.	1213	create event objects, althoguh an our of order queue is often a bit
		1214	faster.
804		1215
805	=over 4	1216	=over 4
806		1217
807	=item $ev = $queue->enqueue_read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...)	1218	=item $ev = $queue->read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...)
808		1219
809	Reads data from buffer into the given string.	1220	Reads data from buffer into the given string.
810		1221
811	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html>	1222	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html>
812		1223
813	=item $ev = $queue->enqueue_write_buffer ($buffer, $blocking, $offset, $data, $wait_events...)	1224	=item $ev = $queue->write_buffer ($buffer, $blocking, $offset, $data, $wait_events...)
814		1225
815	Writes data to buffer from the given string.	1226	Writes data to buffer from the given string.
816		1227
817	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html>	1228	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html>
818		1229
819	=item $ev = $queue->enqueue_copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...)	1230	=item $ev = $queue->copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...)
820		1231
821	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html>	1232	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html>
822		1233
823	=item $ev = $queue->enqueue_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...)	1234	=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...)
824		1235
825	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBufferRect.html	1236	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBufferRect.html
826		1237
827	=item $ev = $queue->enqueue_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...)	1238	=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...)
828		1239
829	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBufferRect.html	1240	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBufferRect.html
830		1241
831	=item $ev = $queue->enqueue_read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
832
833	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferRect.html>
834
835	=item $ev = $queue->enqueue_copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)	1242	=item $ev = $queue->copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)
		1243
		1244	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>
		1245
		1246	=item $ev = $queue->read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
		1247
		1248	C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
		1249	module uses the image width (and height) to supply default values.
836		1250
837	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html>	1251	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html>
838		1252
839	=item $ev = $queue->enqueue_write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)	1253	=item $ev = $queue->write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
840		1254
		1255	C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
		1256	module uses the image width (and height) to supply default values.
841	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html>	1257	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html>
842		1258
843	=item $ev = $queue->enqueue_copy_image ($src_image, $dst_image, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)	1259	=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...)
844		1260
845	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html>	1261	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html>
846		1262
847	=item $ev = $queue->enqueue_copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...)	1263	=item $ev = $queue->copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...)
848		1264
849	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html>	1265	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html>
850		1266
851	=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, $dst_row_pitch, $dst_slice_pitch, $wait_event...)	1267	=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...)
852		1268
853	Yeah.	1269	Yeah.
854		1270
855	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>.	1271	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>.
856		1272
		1273	=item $ev = $queue->fill_buffer ($mem, $pattern, $offset, $size, ...)
		1274
		1275	Fills the given buffer object with repeated applications of C<$pattern>,
		1276	starting at C<$offset> for C<$size> octets.
		1277
		1278	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillBuffer.html>
		1279
		1280	=item $ev = $queue->fill_image ($img, $r, $g, $b, $a, $x, $y, $z, $width, $height, $depth, ...)
		1281
		1282	Fills the given image area with the given rgba colour components. The
		1283	components are normally floating point values between C<0> and C<1>,
		1284	except when the image channel data type is a signe dor unsigned
		1285	unnormalised format, in which case the range is determined by the format.
		1286
		1287	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillImage.html>
		1288
857	=item $ev = $queue->enqueue_task ($kernel, $wait_events...)	1289	=item $ev = $queue->task ($kernel, $wait_events...)
858		1290
859	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html>	1291	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html>
860		1292
861	=item $ev = $queue->enqueue_nd_range_kernel ($kernel, @$global_work_offset, @$global_work_size, @$local_work_size, $wait_events...)	1293	=item $ev = $queue->nd_range_kernel ($kernel, \@global_work_offset, \@global_work_size, \@local_work_size, $wait_events...)
862		1294
863	Enqueues a kernel execution.	1295	Enqueues a kernel execution.
864		1296
865	@$global_work_size must be specified as a reference to an array of	1297	\@global_work_size must be specified as a reference to an array of
866	integers specifying the work sizes (element counts).	1298	integers specifying the work sizes (element counts).
867		1299
868	@$global_work_offset must be either C<undef> (in which case all offsets	1300	\@global_work_offset must be either C<undef> (in which case all offsets
869	are C<0>), or a reference to an array of work offsets, with the same number	1301	are C<0>), or a reference to an array of work offsets, with the same number
870	of elements as @$global_work_size.	1302	of elements as \@global_work_size.
871		1303
872	@$local_work_size must be either C<undef> (in which case the	1304	\@local_work_size must be either C<undef> (in which case the
873	implementation is supposed to choose good local work sizes), or a	1305	implementation is supposed to choose good local work sizes), or a
874	reference to an array of local work sizes, with the same number of	1306	reference to an array of local work sizes, with the same number of
875	elements as @$global_work_size.	1307	elements as \@global_work_size.
876		1308
877	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html>	1309	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html>
878		1310
879	=item $ev = $queue->enqueue_marker ($wait_events...)
880
881	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMarker.html>
882
883	=item $ev = $queue->enqueue_acquire_gl_objects ([object, ...], $wait_events...)	1311	=item $ev = $queue->acquire_gl_objects ([object, ...], $wait_events...)
884		1312
885	Enqueues a list (an array-ref of OpenCL::Memory objects) to be acquired	1313	Enqueues a list (an array-ref of OpenCL::Memory objects) to be acquired
886	for subsequent OpenCL usage.	1314	for subsequent OpenCL usage.
887		1315
888	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueAcquireGLObjects.html>	1316	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueAcquireGLObjects.html>
889		1317
890	=item $ev = $queue->enqueue_release_gl_objects ([object, ...], $wait_events...)	1318	=item $ev = $queue->release_gl_objects ([object, ...], $wait_events...)
891		1319
892	Enqueues a list (an array-ref of OpenCL::Memory objects) to be released	1320	Enqueues a list (an array-ref of OpenCL::Memory objects) to be released
893	for subsequent OpenGL usage.	1321	for subsequent OpenGL usage.
894		1322
895	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReleaseGLObjects.html>	1323	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReleaseGLObjects.html>
896		1324
897	=item $ev = $queue->enqueue_wait_for_events ($wait_events...)	1325	=item $ev = $queue->wait_for_events ($wait_events...)
898		1326
899	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html>	1327	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html>
900		1328
901	=item $queue->enqueue_barrier	1329	=item $ev = $queue->marker ($wait_events...)
902		1330
		1331	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMarkerWithWaitList.html>
		1332
		1333	=item $ev = $queue->barrier ($wait_events...)
		1334
903	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueBarrier.html>	1335	L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueBarrierWithWaitList.html>
904		1336
905	=item $queue->flush	1337	=item $queue->flush
906		1338
907	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html>	1339	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html>
908		1340
…		…
933	=item $command_queue_properties = $command_queue->properties	1365	=item $command_queue_properties = $command_queue->properties
934		1366
935	Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_PROPERTIES> and returns the result.	1367	Calls C<clGetCommandQueueInfo> with C<CL_QUEUE_PROPERTIES> and returns the result.
936		1368
937	=for gengetinfo end command_queue	1369	=for gengetinfo end command_queue
		1370
		1371	=back
		1372
		1373	=head3 MEMORY MAPPED BUFFERS
		1374
		1375	OpenCL allows you to map buffers and images to host memory (read: perl
		1376	scalars). This is done much like reading or copying a buffer, by enqueuing
		1377	a map or unmap operation on the command queue.
		1378
		1379	The map operations return a C<OpenCL::Mapped> object - see L<THE
		1380	OpenCL::Mapped CLASS> section for details on what to do with these
		1381	objects.
		1382
		1383	The object will be unmapped automatically when the mapped object is
		1384	destroyed (you can use a barrier to make sure the unmap has finished,
		1385	before using the buffer in a kernel), but you can also enqueue an unmap
		1386	operation manually.
		1387
		1388	=over 4
		1389
		1390	=item $mapped_buffer = $queue->map_buffer ($buf, $data, $blocking=1, $map_flags=OpenCL::MAP_READ\|OpenCL::MAP_WRITE, $offset=0, $size=0, $wait_events...)
		1391
		1392	Maps the given buffer into host memory and returns a C<OpenCL::MappedBuffer> object.
		1393
		1394	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapBuffer.html>
		1395
		1396	=item $mapped_image = $queue->map_image ($img, $data, $blocking=1, $map_flags=OpenCL::MAP_READ\|OpenCL::MAP_WRITE, $x=0, $y=0, $z=0, $width=0, $height=0, $depth=0, $wait_events...)
		1397
		1398	Maps the given image area into host memory and return a
		1399	C<OpenCL::MappedImage> object. Although there are default values for most
		1400	arguments, you currently have to specify all arguments, otherwise the call
		1401	will fail.
		1402
		1403	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapImage.html>
		1404
		1405	=item $ev = $queue->unmap ($mapped, $wait_events...)
		1406
		1407	Unmaps the data from host memory. You must not call any methods that
		1408	modify the data, or modify the data scalar directly, after calling this
		1409	method.
		1410
		1411	The mapped event object will always be passed as part of the
		1412	$wait_events. The mapped event object will be replaced by the new event
		1413	object that this request creates.
938		1414
939	=back	1415	=back
940		1416
941	=head2 THE OpenCL::Memory CLASS	1417	=head2 THE OpenCL::Memory CLASS
942		1418
…		…
1023		1499
1024	=back	1500	=back
1025		1501
1026	=head2 THE OpenCL::Image CLASS	1502	=head2 THE OpenCL::Image CLASS
1027		1503
1028	This is the superclass of all image objects - OpenCL::Image2D and OpenCL::Image3D.	1504	This is the superclass of all image objects - OpenCL::Image1D,
		1505	OpenCL::Image1DArray, OpenCL::Image1DBuffer, OpenCL::Image2D,
		1506	OpenCL::Image2DArray and OpenCL::Image3D.
1029		1507
1030	=over 4	1508	=over 4
1031		1509
1032	=item $packed_value = $ev->image_info ($name)	1510	=item $packed_value = $image->image_info ($name)
1033		1511
1034	See C<< $platform->info >> for details.	1512	See C<< $platform->info >> for details.
1035		1513
1036	The reason this method is not called C<info> is that there already is an	1514	The reason this method is not called C<info> is that there already is an
1037	C<< ->info >> method inherited from C<OpenCL::Memory>.	1515	C<< ->info >> method inherited from C<OpenCL::Memory>.
1038		1516
1039	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetImageInfo.html>	1517	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetImageInfo.html>
1040		1518
		1519	=item ($channel_order, $channel_data_type) = $image->format
		1520
		1521	Returns the channel order and type used to create the image by calling
		1522	C<clGetImageInfo> with C<CL_IMAGE_FORMAT>.
		1523
1041	=for gengetinfo begin image	1524	=for gengetinfo begin image
1042		1525
1043	=item $int = $image->element_size	1526	=item $int = $image->element_size
1044		1527
1045	Calls C<clGetImageInfo> with C<CL_IMAGE_ELEMENT_SIZE> and returns the result.	1528	Calls C<clGetImageInfo> with C<CL_IMAGE_ELEMENT_SIZE> and returns the result.
…		…
1068		1551
1069	=for gengetinfo begin gl_texture	1552	=for gengetinfo begin gl_texture
1070		1553
1071	=item $GLenum = $gl_texture->target	1554	=item $GLenum = $gl_texture->target
1072		1555
1073	Calls C<clGetGlTextureInfo> with C<CL_GL_TEXTURE_TARGET> and returns the result.	1556	Calls C<clGetGLTextureInfo> with C<CL_GL_TEXTURE_TARGET> and returns the result.
1074		1557
1075	=item $GLint = $gl_texture->gl_mipmap_level	1558	=item $GLint = $gl_texture->gl_mipmap_level
1076		1559
1077	Calls C<clGetGlTextureInfo> with C<CL_GL_MIPMAP_LEVEL> and returns the result.	1560	Calls C<clGetGLTextureInfo> with C<CL_GL_MIPMAP_LEVEL> and returns the result.
1078		1561
1079	=for gengetinfo end gl_texture	1562	=for gengetinfo end gl_texture
1080		1563
1081	=back	1564	=back
1082		1565
…		…
1118		1601
1119	=head2 THE OpenCL::Program CLASS	1602	=head2 THE OpenCL::Program CLASS
1120		1603
1121	=over 4	1604	=over 4
1122		1605
1123	=item $program->build ($device, $options = "")	1606	=item $program->build (\@devices = undef, $options = "", $cb->($program) = undef)
1124		1607
1125	Tries to build the program with the givne options.	1608	Tries to build the program with the given options. See also the
		1609	C<$ctx->build> convenience function.
		1610
		1611	If a callback is specified, then it will be called when compilation is
		1612	finished. Note that many OpenCL implementations block your program while
		1613	compiling whether you use a callback or not. See C<build_async> if you
		1614	want to make sure the build is done in the background.
		1615
		1616	Note that some OpenCL implementations act up badly, and don't call the
		1617	callback in some error cases (but call it in others). This implementation
		1618	assumes the callback will always be called, and leaks memory if this is
		1619	not so. So best make sure you don't pass in invalid values.
		1620
		1621	Some implementations fail with C<OpenCL::INVALID_BINARY> when the
		1622	compilation state is successful but some later stage fails.
1126		1623
1127	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html>	1624	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html>
		1625
		1626	=item $program->build_async (\@devices = undef, $options = "", $cb->($program) = undef)
		1627
		1628	Similar to C<< ->build >>, except it starts a thread, and never fails (you
		1629	need to check the compilation status form the callback, or by polling).
1128		1630
1129	=item $packed_value = $program->build_info ($device, $name)	1631	=item $packed_value = $program->build_info ($device, $name)
1130		1632
1131	Similar to C<< $platform->info >>, but returns build info for a previous	1633	Similar to C<< $platform->info >>, but returns build info for a previous
1132	build attempt for the given device.	1634	build attempt for the given device.
…		…
1137		1639
1138	Creates an OpenCL::Kernel object out of the named C<__kernel> function in	1640	Creates an OpenCL::Kernel object out of the named C<__kernel> function in
1139	the program.	1641	the program.
1140		1642
1141	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html>	1643	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html>
		1644
		1645	=item @kernels = $program->kernels_in_program
		1646
		1647	Returns all kernels successfully compiled for all devices in program.
		1648
		1649	http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernelsInProgram.html
1142		1650
1143	=for gengetinfo begin program_build	1651	=for gengetinfo begin program_build
1144		1652
1145	=item $build_status = $program->build_status ($device)	1653	=item $build_status = $program->build_status ($device)
1146		1654
…		…
1268		1776
1269	Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_PRIVATE_MEM_SIZE> and returns the result.	1777	Calls C<clGetKernelWorkGroupInfo> with C<CL_KERNEL_PRIVATE_MEM_SIZE> and returns the result.
1270		1778
1271	=for gengetinfo end kernel_work_group	1779	=for gengetinfo end kernel_work_group
1272		1780
		1781	=item $kernel->setf ($format, ...)
		1782
		1783	Sets the arguments of a kernel. Since OpenCL 1.1 doesn't have a generic
		1784	way to set arguments (and with OpenCL 1.2 it might be rather slow), you
		1785	need to specify a format argument, much as with C<printf>, to tell OpenCL
		1786	what type of argument it is.
		1787
		1788	The format arguments are single letters:
		1789
		1790	c char
		1791	C unsigned char
		1792	s short
		1793	S unsigned short
		1794	i int
		1795	I unsigned int
		1796	l long
		1797	L unsigned long
		1798
		1799	h half float (0..65535)
		1800	f float
		1801	d double
		1802
		1803	z local (octet size)
		1804
		1805	m memory object (buffer or image)
		1806	a sampler
		1807	e event
		1808
		1809	Space characters in the format string are ignored.
		1810
		1811	Example: set the arguments for a kernel that expects an int, two floats, a buffer and an image.
		1812
		1813	$kernel->setf ("i ff mm", 5, 0.5, 3, $buffer, $image);
		1814
1273	=item $kernel->set_TYPE ($index, $value)	1815	=item $kernel->set_TYPE ($index, $value)
1274		1816
		1817	=item $kernel->set_char ($index, $value)
		1818
		1819	=item $kernel->set_uchar ($index, $value)
		1820
		1821	=item $kernel->set_short ($index, $value)
		1822
		1823	=item $kernel->set_ushort ($index, $value)
		1824
		1825	=item $kernel->set_int ($index, $value)
		1826
		1827	=item $kernel->set_uint ($index, $value)
		1828
		1829	=item $kernel->set_long ($index, $value)
		1830
		1831	=item $kernel->set_ulong ($index, $value)
		1832
		1833	=item $kernel->set_half ($index, $value)
		1834
		1835	=item $kernel->set_float ($index, $value)
		1836
		1837	=item $kernel->set_double ($index, $value)
		1838
		1839	=item $kernel->set_memory ($index, $value)
		1840
		1841	=item $kernel->set_buffer ($index, $value)
		1842
		1843	=item $kernel->set_image ($index, $value)
		1844
		1845	=item $kernel->set_sampler ($index, $value)
		1846
		1847	=item $kernel->set_local ($index, $value)
		1848
		1849	=item $kernel->set_event ($index, $value)
		1850
1275	This is a family of methods to set the kernel argument with the number C<$index> to the give C<$value>.	1851	This is a family of methods to set the kernel argument with the number
1276		1852	C<$index> to the give C<$value>.
1277	TYPE is one of C<char>, C<uchar>, C<short>, C<ushort>, C<int>, C<uint>,
1278	C<long>, C<ulong>, C<half>, C<float>, C<double>, C<memory>, C<buffer>,
1279	C<image2d>, C<image3d>, C<sampler> or C<event>.
1280		1853
1281	Chars and integers (including the half type) are specified as integers,	1854	Chars and integers (including the half type) are specified as integers,
1282	float and double as floating point values, memory/buffer/image2d/image3d	1855	float and double as floating point values, memory/buffer/image must be
1283	must be an object of that type or C<undef>, and sampler and event must be	1856	an object of that type or C<undef>, local-memory arguments are set by
1284	objects of that type.	1857	specifying the size, and sampler and event must be objects of that type.
		1858
		1859	Note that C<set_memory> works for all memory objects (all types of buffers
		1860	and images) - the main purpose of the more specific C<set_TYPE> functions
		1861	is type checking.
		1862
		1863	Setting an argument for a kernel does NOT keep a reference to the object -
		1864	for example, if you set an argument to some image object, free the image,
		1865	and call the kernel, you will run into undefined behaviour.
1285		1866
1286	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html>	1867	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html>
1287		1868
1288	=back	1869	=back
1289		1870
…		…
1298		1879
1299	Waits for the event to complete.	1880	Waits for the event to complete.
1300		1881
1301	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>	1882	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
1302		1883
		1884	=item $ev->cb ($exec_callback_type, $callback->($event, $event_command_exec_status))
		1885
		1886	Adds a callback to the callback stack for the given event type. There is
		1887	no way to remove a callback again.
		1888
		1889	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetEventCallback.html>
		1890
1303	=item $packed_value = $ev->info ($name)	1891	=item $packed_value = $ev->info ($name)
1304		1892
1305	See C<< $platform->info >> for details.	1893	See C<< $platform->info >> for details.
1306		1894
1307	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html>	1895	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html>
…		…
1367		1955
1368	=over 4	1956	=over 4
1369		1957
1370	=item $ev->set_status ($execution_status)	1958	=item $ev->set_status ($execution_status)
1371		1959
		1960	Sets the execution status of the user event. Can only be called once,
		1961	either with OpenCL::COMPLETE or a negative number as status.
		1962
1372	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html>	1963	L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html>
1373		1964
1374	=back	1965	=back
1375		1966
		1967	=head2 THE OpenCL::Mapped CLASS
		1968
		1969	This class represents objects mapped into host memory. They are
		1970	represented by a blessed string scalar. The string data is the mapped
		1971	memory area, that is, if you read or write it, then the mapped object is
		1972	accessed directly.
		1973
		1974	You must only ever use operations that modify the string in-place - for
		1975	example, a C<substr> that doesn't change the length, or maybe a regex that
		1976	doesn't change the length. Any other operation might cause the data to be
		1977	copied.
		1978
		1979	When the object is destroyed it will enqueue an implicit unmap operation
		1980	on the queue that was used to create it.
		1981
		1982	Keep in mind that you I<need> to unmap (or destroy) mapped objects before
		1983	OpenCL sees the changes, even if some implementations don't need this
		1984	sometimes.
		1985
		1986	Example, replace the first two floats in the mapped buffer by 1 and 2.
		1987
		1988	my $mapped = $queue->map_buffer ($buf, ...
		1989	$mapped->event->wait; # make sure it's there
		1990
		1991	# now replace first 8 bytes by new data, which is exactly 8 bytes long
		1992	# we blindly assume device endianness to equal host endianness
		1993	# (and of course, we assume iee 754 single precision floats :)
		1994	substr $$mapped, 0, 8, pack "f*", 1, 2;
		1995
		1996	=over 4
		1997
		1998	=item $ev = $mapped->unmap ($wait_events...)
		1999
		2000	Unmaps the mapped memory object, using the queue originally used to create
		2001	it, quite similarly to C<< $queue->unmap ($mapped, ...) >>.
		2002
		2003	=item $bool = $mapped->mapped
		2004
		2005	Returns whether the object is still mapped - true before an C<unmap> is
		2006	enqueued, false afterwards.
		2007
		2008	=item $ev = $mapped->event
		2009
		2010	Return the event object associated with the mapped object. Initially, this
		2011	will be the event object created when mapping the object, and after an
		2012	unmap, this will be the event object that the unmap operation created.
		2013
		2014	=item $mapped->wait
		2015
		2016	Same as C<< $mapped->event->wait >> - makes sure no operations on this
		2017	mapped object are outstanding.
		2018
		2019	=item $bytes = $mapped->size
		2020
		2021	Returns the size of the mapped area, in bytes. Same as C<length $$mapped>.
		2022
		2023	=item $ptr = $mapped->ptr
		2024
		2025	Returns the raw memory address of the mapped area.
		2026
		2027	=item $mapped->set ($offset, $data)
		2028
		2029	Replaces the data at the given C<$offset> in the memory area by the new
		2030	C<$data>. This method is safer than direct manipulation of C<$mapped>
		2031	because it does bounds-checking, but also slower.
		2032
		2033	=item $data = $mapped->get ($offset, $length)
		2034
		2035	Returns (without copying) a scalar representing the data at the given
		2036	C<$offset> and C<$length> in the mapped memory area. This is the same as
		2037	the following substr, except much slower;
		2038
		2039	$data = substr $$mapped, $offset, $length
		2040
1376	=cut	2041	=cut
1377		2042
1378	package OpenCL;	2043	sub OpenCL::Mapped::get {
1379		2044	substr ${$_[0]}, $_[1], $_[2]
1380	use common::sense;
1381
1382	BEGIN {
1383	our $VERSION = '0.92';
1384
1385	require XSLoader;
1386	XSLoader::load (__PACKAGE__, $VERSION);
1387
1388	@OpenCL::Buffer::ISA =
1389	@OpenCL::Image::ISA = OpenCL::Memory::;
1390
1391	@OpenCL::BufferObj::ISA = OpenCL::Buffer::;
1392
1393	@OpenCL::Image2D::ISA =
1394	@OpenCL::Image3D::ISA = OpenCL::Image::;
1395
1396	@OpenCL::UserEvent::ISA = OpenCL::Event::;
1397	}	2045	}
		2046
		2047	=back
		2048
		2049	=head2 THE OpenCL::MappedBuffer CLASS
		2050
		2051	This is a subclass of OpenCL::Mapped, representing mapped buffers.
		2052
		2053	=head2 THE OpenCL::MappedImage CLASS
		2054
		2055	This is a subclass of OpenCL::Mapped, representing mapped images.
		2056
		2057	=over 4
		2058
		2059	=item $bytes = $mapped->row_pitch
		2060
		2061	=item $bytes = $mapped->slice_pitch
		2062
		2063	Return the row or slice pitch of the image that has been mapped.
		2064
		2065	=back
		2066
		2067
		2068	=cut
1398		2069
1399	1;	2070	1;
1400		2071
1401	=head1 AUTHOR	2072	=head1 AUTHOR
1402		2073

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing OpenCL/OpenCL.pm (file contents): Revision 1.36 by root, Thu Apr 19 14:34:56 2012 UTC vs. Revision 1.68 by root, Tue May 1 22:25:13 2012 UTC

Diff Legend

Comparing OpenCL/OpenCL.pm (file contents):
Revision 1.36 by root, Thu Apr 19 14:34:56 2012 UTC vs.
Revision 1.68 by root, Tue May 1 22:25:13 2012 UTC