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 1.87 by root, Mon May 7 01:01:33 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 - download these and keept
16no implementation was available to me :). 49hema round, they are required reference material:
17 50
18 http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf 51 http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf
52 http://www.khronos.org/registry/cl/specs/opencl-1.2.pdf
53 http://www.khronos.org/registry/cl/specs/opencl-1.2-extensions.pdf
19 54
20OpenCL manpages: 55OpenCL manpages:
21 56
22 http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/ 57 http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/
58 http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/
59
60If you are into UML class diagrams, the following diagram might help - if
61not, it will be mildly confusing (also, the class hierarchy of this module
62is much more fine-grained):
63
64 http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/classDiagram.html
65
66Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it
67is, but at least it's free of charge:
68
69 http://developer.amd.com/zones/OpenCLZone/courses/Documents/Introduction_to_OpenCL_Programming%20Training_Guide%20%28201005%29.pdf
70
71And here's NVIDIA's OpenCL Best Practises Guide:
72
73 http://developer.download.nvidia.com/compute/cuda/3_2/toolkit/docs/OpenCL_Best_Practices_Guide.pdf
74
75=head1 BASIC WORKFLOW
76
77To get something done, you basically have to do this once (refer to the
78examples below for actual code, this is just a high-level description):
79
80Find some platform (e.g. the first one) and some device(s) (e.g. the first
81device of the platform), and create a context from those.
82
83Create program objects from your OpenCL source code, then build (compile)
84the programs for each device you want to run them on.
85
86Create kernel objects for all kernels you want to use (surprisingly, these
87are not device-specific).
88
89Then, to execute stuff, you repeat these steps, possibly resuing or
90sharing some buffers:
91
92Create some input and output buffers from your context. Set these as
93arguments to your kernel.
94
95Enqueue buffer writes to initialise your input buffers (when not
96initialised at creation time).
97
98Enqueue the kernel execution.
99
100Enqueue buffer reads for your output buffer to read results.
23 101
24=head1 EXAMPLES 102=head1 EXAMPLES
25 103
26=head2 Enumerate all devices and get contexts for them. 104=head2 Enumerate all devices and get contexts for them.
27 105
106Best run this once to get a feel for the platforms and devices in your
107system.
108
28 for my $platform (OpenCL::platforms) { 109 for my $platform (OpenCL::platforms) {
29 warn $platform->info (OpenCL::PLATFORM_NAME); 110 printf "platform: %s\n", $platform->name;
30 warn $platform->info (OpenCL::PLATFORM_EXTENSIONS); 111 printf "extensions: %s\n", $platform->extensions;
31 for my $device ($platform->devices) { 112 for my $device ($platform->devices) {
32 warn $device->info (OpenCL::DEVICE_NAME); 113 printf "+ device: %s\n", $device->name;
33 my $ctx = $device->context_simple; 114 my $ctx = $platform->context (undef, [$device]);
34 # do stuff 115 # do stuff
35 } 116 }
36 } 117 }
37 118
38=head2 Get a useful context and a command queue. 119=head2 Get a useful context and a command queue.
39 120
40 my $dev = ((OpenCL::platforms)[0]->devices)[0]; 121This is a useful boilerplate for any OpenCL program that only wants to use
41 my $ctx = $dev->context_simple; 122one device,
42 my $queue = $ctx->command_queue_simple ($dev); 123
124 my ($platform) = OpenCL::platforms; # find first platform
125 my ($dev) = $platform->devices; # find first device of platform
126 my $ctx = $platform->context (undef, [$dev]); # create context out of those
127 my $queue = $ctx->queue ($dev); # create a command queue for the device
43 128
44=head2 Print all supported image formats of a context. 129=head2 Print all supported image formats of a context.
45 130
131Best run this once for your context, to see whats available and how to
132gather information.
133
46 for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) { 134 for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) {
47 say "supported image formats for ", OpenCL::enum2str $type; 135 print "supported image formats for ", OpenCL::enum2str $type, "\n";
48 136
49 for my $f ($ctx->supported_image_formats (0, $type)) { 137 for my $f ($ctx->supported_image_formats (0, $type)) {
50 printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1]; 138 printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1];
51 } 139 }
52 } 140 }
54=head2 Create a buffer with some predefined data, read it back synchronously, 142=head2 Create a buffer with some predefined data, read it back synchronously,
55then asynchronously. 143then asynchronously.
56 144
57 my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut"); 145 my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut");
58 146
59 $queue->enqueue_read_buffer ($buf, 1, 1, 3, my $data); 147 $queue->read_buffer ($buf, 1, 1, 3, my $data);
60 warn $data; 148 print "$data\n";
61 149
62 my $ev = $queue->enqueue_read_buffer ($buf, 0, 1, 3, my $data); 150 my $ev = $queue->read_buffer ($buf, 0, 1, 3, my $data);
63 $ev->wait; 151 $ev->wait;
64 warn $data; 152 print "$data\n"; # prints "elm"
65 153
66=head2 Create and build a program, then create a kernel out of one of its 154=head2 Create and build a program, then create a kernel out of one of its
67functions. 155functions.
68 156
69 my $src = ' 157 my $src = '
70 __kernel void 158 kernel void
71 squareit (__global float *input, __global float *output) 159 squareit (global float *input, global float *output)
72 { 160 {
73 size_t id = get_global_id (0); 161 $id = get_global_id (0);
74 output [id] = input [id] * input [id]; 162 output [id] = input [id] * input [id];
75 } 163 }
76 '; 164 ';
77 165
78 my $prog = $ctx->program_with_source ($src); 166 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"); 167 my $kernel = $prog->kernel ("squareit");
84 168
85=head2 Create some input and output float buffers, then call squareit on them. 169=head2 Create some input and output float buffers, then call the
170'squareit' kernel on them.
86 171
87 my $input = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, pack "f*", 1, 2, 3, 4.5); 172 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); 173 my $output = $ctx->buffer (0, OpenCL::SIZEOF_FLOAT * 5);
89 174
90 # set buffer 175 # set buffer
91 $kernel->set_buffer (0, $input); 176 $kernel->set_buffer (0, $input);
92 $kernel->set_buffer (1, $output); 177 $kernel->set_buffer (1, $output);
93 178
94 # execute it for all 4 numbers 179 # execute it for all 4 numbers
95 $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 180 $queue->nd_range_kernel ($kernel, undef, [4], undef);
96 181
97 # enqueue a synchronous read 182 # enqueue a synchronous read
98 $queue->enqueue_read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); 183 $queue->read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
99 184
100 # print the results: 185 # print the results:
101 say join ", ", unpack "f*", $data; 186 printf "%s\n", join ", ", unpack "f*", $data;
102 187
103=head2 The same enqueue operations as before, but assuming an out-of-order queue, 188=head2 The same enqueue operations as before, but assuming an out-of-order queue,
104showing off barriers. 189showing off barriers.
105 190
106 # execute it for all 4 numbers 191 # execute it for all 4 numbers
107 $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 192 $queue->nd_range_kernel ($kernel, undef, [4], undef);
108 193
109 # enqueue a barrier to ensure in-order execution 194 # enqueue a barrier to ensure in-order execution
110 $queue->enqueue_barrier; 195 $queue->barrier;
111 196
112 # enqueue an async read 197 # enqueue an async read
113 $queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); 198 $queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data);
114 199
115 # wait for all requests to finish 200 # wait for all requests to finish
116 $queue->finish; 201 $queue->finish;
117 202
118=head2 The same enqueue operations as before, but assuming an out-of-order queue, 203=head2 The same enqueue operations as before, but assuming an out-of-order queue,
119showing off event objects and wait lists. 204showing off event objects and wait lists.
120 205
121 # execute it for all 4 numbers 206 # execute it for all 4 numbers
122 my $ev = $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); 207 my $ev = $queue->nd_range_kernel ($kernel, undef, [4], undef);
123 208
124 # enqueue an async read 209 # enqueue an async read
125 $ev = $queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev); 210 $ev = $queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev);
126 211
127 # wait for the last event to complete 212 # wait for the last event to complete
128 $ev->wait; 213 $ev->wait;
129 214
215=head2 Use the OpenGL module to share a texture between OpenCL and OpenGL and draw some julia
216set flight effect.
217
218This is quite a long example to get you going - you can also download it
219from L<http://cvs.schmorp.de/OpenCL/examples/juliaflight>.
220
221 use OpenGL ":all";
222 use OpenCL;
223
224 my $S = $ARGV[0] || 256; # window/texture size, smaller is faster
225
226 # open a window and create a gl texture
227 OpenGL::glpOpenWindow width => $S, height => $S;
228 my $texid = glGenTextures_p 1;
229 glBindTexture GL_TEXTURE_2D, $texid;
230 glTexImage2D_c GL_TEXTURE_2D, 0, GL_RGBA8, $S, $S, 0, GL_RGBA, GL_UNSIGNED_BYTE, 0;
231
232 # find and use the first opencl device that let's us get a shared opengl context
233 my $platform;
234 my $dev;
235 my $ctx;
236
237 for (OpenCL::platforms) {
238 $platform = $_;
239 for ($platform->devices) {
240 $dev = $_;
241 $ctx = $platform->context ([OpenCL::GLX_DISPLAY_KHR, undef, OpenCL::GL_CONTEXT_KHR, undef], [$dev])
242 and last;
243 }
244 }
245
246 $ctx
247 or die "cannot find suitable OpenCL device\n";
248
249 my $queue = $ctx->queue ($dev);
250
251 # now attach an opencl image2d object to the opengl texture
252 my $tex = $ctx->gl_texture2d (OpenCL::MEM_WRITE_ONLY, GL_TEXTURE_2D, 0, $texid);
253
254 # now the boring opencl code
255 my $src = <<EOF;
256 kernel void
257 juliatunnel (write_only image2d_t img, float time)
258 {
259 int2 xy = (int2)(get_global_id (0), get_global_id (1));
260 float2 p = convert_float2 (xy) / $S.f * 2.f - 1.f;
261
262 float2 m = (float2)(1.f, p.y) / fabs (p.x); // tunnel
263 m.x = fabs (fmod (m.x + time * 0.05f, 4.f) - 2.f);
264
265 float2 z = m;
266 float2 c = (float2)(sin (time * 0.01133f), cos (time * 0.02521f));
267
268 for (int i = 0; i < 25 && dot (z, z) < 4.f; ++i) // standard julia
269 z = (float2)(z.x * z.x - z.y * z.y, 2.f * z.x * z.y) + c;
270
271 float3 colour = (float3)(z.x, z.y, atan2 (z.y, z.x));
272 write_imagef (img, xy, (float4)(colour * p.x * p.x, 1.));
273 }
274 EOF
275
276 my $prog = $ctx->build_program ($src);
277 my $kernel = $prog->kernel ("juliatunnel");
278
279 # program compiled, kernel ready, now draw and loop
280
281 for (my $time; ; ++$time) {
282 # acquire objects from opengl
283 $queue->acquire_gl_objects ([$tex]);
284
285 # configure and run our kernel
286 $kernel->setf ("mf", $tex, $time*2); # mf = memory object, float
287 $queue->nd_range_kernel ($kernel, undef, [$S, $S], undef);
288
289 # release objects to opengl again
290 $queue->release_gl_objects ([$tex]);
291
292 # wait
293 $queue->finish;
294
295 # now draw the texture, the defaults should be all right
296 glTexParameterf GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST;
297
298 glEnable GL_TEXTURE_2D;
299 glBegin GL_QUADS;
300 glTexCoord2f 0, 1; glVertex3i -1, -1, -1;
301 glTexCoord2f 0, 0; glVertex3i 1, -1, -1;
302 glTexCoord2f 1, 0; glVertex3i 1, 1, -1;
303 glTexCoord2f 1, 1; glVertex3i -1, 1, -1;
304 glEnd;
305
306 glXSwapBuffers;
307
308 select undef, undef, undef, 1/60;
309 }
310
311=head2 How to modify the previous example to not rely on GL sharing.
312
313For those poor souls with only a sucky CPU OpenCL implementation, you
314currently have to read the image into some perl scalar, and then modify a
315texture or use glDrawPixels or so).
316
317First, when you don't need gl sharing, you can create the context much simpler:
318
319 $ctx = $platform->context (undef, [$dev])
320
321To use a texture, you would modify the above example by creating an
322OpenCL::Image manually instead of deriving it from a texture:
323
324 my $tex = $ctx->image2d (OpenCL::MEM_WRITE_ONLY, OpenCL::RGBA, OpenCL::UNORM_INT8, $S, $S);
325
326And in the draw loop, intead of acquire_gl_objects/release_gl_objects, you
327would read the image2d after the kernel has written it:
328
329 $queue->read_image ($tex, 0, 0, 0, 0, $S, $S, 1, 0, 0, my $data);
330
331And then you would upload the pixel data to the texture (or use glDrawPixels):
332
333 glTexSubImage2D_s GL_TEXTURE_2D, 0, 0, 0, $S, $S, GL_RGBA, GL_UNSIGNED_BYTE, $data;
334
335The fully modified example can be found at
336L<http://cvs.schmorp.de/OpenCL/examples/juliaflight-nosharing>.
337
338=head2 Julia sets look soooo 80ies.
339
340Then colour them differently, e.g. using orbit traps! Replace the loop and
341colour calculation from the previous examples by this:
342
343 float2 dm = (float2)(1.f, 1.f);
344
345 for (int i = 0; i < 25; ++i)
346 {
347 z = (float2)(z.x * z.x - z.y * z.y, 2.f * z.x * z.y) + c;
348 dm = fmin (dm, (float2)(fabs (dot (z, z) - 1.f), fabs (z.x - 1.f)));
349 }
350
351 float3 colour = (float3)(dm.x * dm.y, dm.x * dm.y, dm.x);
352
353Also try C<-10.f> instead of C<-1.f>.
354
130=head1 DOCUMENTATION 355=head1 DOCUMENTATION
131 356
132=head2 BASIC CONVENTIONS 357=head2 BASIC CONVENTIONS
133 358
134This is not a 1:1 C-style translation of OpenCL to Perl - instead I 359This 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 360I attempted to make the interface as type-safe as possible by introducing
136object syntax where it makes sense. There are a number of important 361object syntax where it makes sense. There are a number of important
137differences between the OpenCL C API and this module: 362differences between the OpenCL C API and this module:
138 363
139=over 4 364=over 4
140 365
141=item * Object lifetime managament is automatic - there is no need 366=item * Object lifetime managament is automatic - there is no need
142to free objects explicitly (C<clReleaseXXX>), the release function 367to free objects explicitly (C<clReleaseXXX>), the release function
143is called automatically once all Perl references to it go away. 368is called automatically once all Perl references to it go away.
144 369
145=item * OpenCL uses CamelCase for function names (C<clGetPlatformInfo>), 370=item * OpenCL uses CamelCase for function names
371(e.g. C<clGetPlatformIDs>, C<clGetPlatformInfo>), while this module
146while this module uses underscores as word separator and often leaves out 372uses underscores as word separator and often leaves out prefixes
147prefixes (C<< $platform->info >>). 373(C<OpenCL::platforms>, C<< $platform->info >>).
148 374
149=item * OpenCL often specifies fixed vector function arguments as short 375=item * OpenCL often specifies fixed vector function arguments as short
150arrays (C<size_t origin[3]>), while this module explicitly expects the 376arrays (C<size_t origin[3]>), while this module explicitly expects the
151components as separate arguments- 377components as separate arguments (C<$orig_x, $orig_y, $orig_z>) in
378function calls.
152 379
153=item * Where possible, the row_pitch value is calculated from the perl 380=item * Structures are often specified by flattening out their components
154scalar length and need not be specified. 381as with short vectors, and returned as arrayrefs.
155 382
156=item * When enqueuing commands, the wait list is specified by adding 383=item * When enqueuing commands, the wait list is specified by adding
157extra arguments to the function - everywhere a C<$wait_events...> argument 384extra arguments to the function - anywhere a C<$wait_events...> argument
158is documented this can be any number of event objects. 385is documented this can be any number of event objects. As an extsnion
386implemented by this module, C<undef> values will be ignored in the event
387list.
159 388
160=item * When enqueuing commands, if the enqueue method is called in void 389=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 390context, no event is created. In all other contexts an event is returned
162by the method. 391by the method.
163 392
164=item * This module expects all functions to return C<CL_SUCCESS>. If any 393=item * This module expects all functions to return C<OpenCL::SUCCESS>. If any
165other status is returned the function will throw an exception, so you 394other status is returned the function will throw an exception, so you
166don't normally have to to any error checking. 395don't normally have to to any error checking.
167 396
168=back 397=back
169 398
399=head2 CONSTANTS
400
401All C<CL_xxx> constants that this module supports are always available
402in the C<OpenCL> namespace as C<OpenCL::xxx> (i.e. without the C<CL_>
403prefix). Constants which are not defined in the header files used during
404compilation, or otherwise are not available, will have the value C<0> (in
405some cases, this will make them indistinguishable from real constants,
406sorry).
407
408The latest version of this module knows and exports the constants
409listed in L<http://cvs.schmorp.de/OpenCL/constiv.h>.
410
411=head2 OPENCL 1.1 VS. OPENCL 1.2
412
413This module supports both OpenCL version 1.1 and 1.2, although the OpenCL
4141.2 interface hasn't been tested much for lack of availability of an
415actual implementation.
416
417Every function or method in this manual page that interfaces to a
418particular OpenCL function has a link to the its C manual page.
419
420If the link contains a F<1.1>, then this function is an OpenCL 1.1
421function. Most but not all also exist in OpenCL 1.2, and this module
422tries to emulate the missing ones for you, when told to do so at
423compiletime. You can check whether a function was removed in OpenCL 1.2 by
424replacing the F<1.1> component in the URL by F<1.2>.
425
426If the link contains a F<1.2>, then this is a OpenCL 1.2-only
427function. Even if the module was compiled with OpenCL 1.2 header files
428and has an 1.2 OpenCL library, calling such a function on a platform that
429doesn't implement 1.2 causes undefined behaviour, usually a crash (But
430this is not guaranteed).
431
432You can find out whether this module was compiled to prefer 1.1
433functionality by ooking at C<OpenCL::PREFER_1_1> - if it is true, then
4341.1 functions generally are implemented using 1.1 OpenCL functions. If it
435is false, then 1.1 functions missing from 1.2 are emulated by calling 1.2
436fucntions.
437
438This is a somewhat sorry state of affairs, but the Khronos group choose to
439make every release of OpenCL source and binary incompatible with previous
440releases.
441
442=head2 PERL AND OPENCL TYPES
443
444This handy(?) table lists OpenCL types and their perl, PDL and pack/unpack
445format equivalents:
446
447 OpenCL perl PDL pack/unpack
448 char IV - c
449 uchar IV byte C
450 short IV short s
451 ushort IV ushort S
452 int IV long? l
453 uint IV - L
454 long IV longlong q
455 ulong IV - Q
456 float NV float f
457 half IV ushort S
458 double NV double d
459
460=head2 GLX SUPPORT
461
462Due to the sad state that OpenGL support is in in Perl (mostly the OpenGL
463module, which has little to no documentation and has little to no support
464for glX), this module, as a special extension, treats context creation
465properties C<OpenCL::GLX_DISPLAY_KHR> and C<OpenCL::GL_CONTEXT_KHR>
466specially: If either or both of these are C<undef>, then the OpenCL
467module tries to dynamically resolve C<glXGetCurrentDisplay> and
468C<glXGetCurrentContext>, call these functions and use their return values
469instead.
470
471For this to work, the OpenGL library must be loaded, a GLX context must
472have been created and be made current, and C<dlsym> must be available and
473capable of finding the function via C<RTLD_DEFAULT>.
474
475=head2 EVENT SYSTEM
476
477OpenCL can generate a number of (potentially) asynchronous events, for
478example, after compiling a program, to signal a context-related error or,
479perhaps most important, to signal completion of queued jobs (by setting
480callbacks on OpenCL::Event objects).
481
482The OpenCL module converts all these callbacks into events - you can
483still register callbacks, but they are not executed when your OpenCL
484implementation calls the actual callback, but only later. Therefore, none
485of the limitations of OpenCL callbacks apply to the perl implementation:
486it is perfectly safe to make blocking operations from event callbacks, and
487enqueued operations don't need to be flushed.
488
489To facilitate this, this module maintains an event queue - each
490time an asynchronous event happens, it is queued, and perl will be
491interrupted. This is implemented via the L<Async::Interrupt> module. In
492addition, this module has L<AnyEvent> support, so it can seamlessly
493integrate itself into many event loops.
494
495Since L<Async::Interrupt> is a bit hard to understand, here are some case examples:
496
497=head3 Don't use callbacks.
498
499When your program never uses any callbacks, then there will never be any
500notifications you need to take care of, and therefore no need to worry
501about all this.
502
503You can achieve a great deal by explicitly waiting for events, or using
504barriers and flush calls. In many programs, there is no need at all to
505tinker with asynchronous events.
506
507=head3 Use AnyEvent
508
509This module automatically registers a watcher that invokes all outstanding
510event callbacks when AnyEvent is initialised (and block asynchronous
511interruptions). Using this mode of operations is the safest and most
512recommended one.
513
514To use this, simply use AnyEvent and this module normally, make sure you
515have an event loop running:
516
517 use Gtk2 -init;
518 use AnyEvent;
519
520 # initialise AnyEvent, by creating a watcher, or:
521 AnyEvent::detect;
522
523 my $e = $queue->marker;
524 $e->cb (sub {
525 warn "opencl is finished\n";
526 })
527
528 main Gtk2;
529
530Note that this module will not initialise AnyEvent for you. Before
531AnyEvent is initialised, the module will asynchronously interrupt perl
532instead. To avoid any surprises, it's best to explicitly initialise
533AnyEvent.
534
535You can temporarily enable asynchronous interruptions (see next paragraph)
536by calling C<$OpenCL::INTERRUPT->unblock> and disable them again by
537calling C<$OpenCL::INTERRUPT->block>.
538
539=head3 Let yourself be interrupted at any time
540
541This mode is the default unless AnyEvent is loaded and initialised. In
542this mode, OpenCL asynchronously interrupts a running perl program. The
543emphasis is on both I<asynchronously> and I<running> here.
544
545Asynchronously means that perl might execute your callbacks at any
546time. For example, in the following code (I<THAT YOU SHOULD NOT COPY>),
547the C<until> loop following the marker call will be interrupted by the
548callback:
549
550 my $e = $queue->marker;
551 my $flag;
552 $e->cb (sub { $flag = 1 });
553 1 until $flag;
554 # $flag is now 1
555
556The reason why you shouldn't blindly copy the above code is that
557busy waiting is a really really bad thing, and really really bad for
558performance.
559
560While at first this asynchronous business might look exciting, it can be
561really hard, because you need to be prepared for the callback code to be
562executed at any time, which limits the amount of things the callback code
563can do safely.
564
565This can be mitigated somewhat by using C<<
566$OpenCL::INTERRUPT->scope_block >> (see the L<Async::Interrupt>
567documentation for details).
568
569The other problem is that your program must be actively I<running> to be
570interrupted. When you calculate stuff, your program is running. When you
571hang in some C functions or other block execution (by calling C<sleep>,
572C<select>, running an event loop and so on), your program is waiting, not
573running.
574
575One way around that would be to attach a read watcher to your event loop,
576listening for events on C<< $OpenCL::INTERRUPT->pipe_fileno >>, using a
577dummy callback (C<sub { }>) to temporarily execute some perl code.
578
579That is then awfully close to using the built-in AnyEvent support above,
580though, so consider that one instead.
581
582=head3 Be creative
583
584OpenCL exports the L<Async::Interrupt> object it uses in the global
585variable C<$OpenCL::INTERRUPT>. You can configure it in any way you like.
586
587So if you want to feel like a real pro, err, wait, if you feel no risk
588menas no fun, you can experiment by implementing your own mode of
589operations.
590
591=cut
592
593package OpenCL;
594
595use common::sense;
596use Carp ();
597use Async::Interrupt ();
598
599our $POLL_FUNC; # set by XS
600
601BEGIN {
602 our $VERSION = '1.01';
603
604 require XSLoader;
605 XSLoader::load (__PACKAGE__, $VERSION);
606
607 @OpenCL::Platform::ISA =
608 @OpenCL::Device::ISA =
609 @OpenCL::Context::ISA =
610 @OpenCL::Queue::ISA =
611 @OpenCL::Memory::ISA =
612 @OpenCL::Sampler::ISA =
613 @OpenCL::Program::ISA =
614 @OpenCL::Kernel::ISA =
615 @OpenCL::Event::ISA = OpenCL::Object::;
616
617 @OpenCL::SubDevice::ISA = OpenCL::Device::;
618
619 @OpenCL::Buffer::ISA =
620 @OpenCL::Image::ISA = OpenCL::Memory::;
621
622 @OpenCL::BufferObj::ISA = OpenCL::Buffer::;
623
624 @OpenCL::Image2D::ISA =
625 @OpenCL::Image3D::ISA =
626 @OpenCL::Image2DArray::ISA =
627 @OpenCL::Image1D::ISA =
628 @OpenCL::Image1DArray::ISA =
629 @OpenCL::Image1DBuffer::ISA = OpenCL::Image::;
630
631 @OpenCL::UserEvent::ISA = OpenCL::Event::;
632
633 @OpenCL::MappedBuffer::ISA =
634 @OpenCL::MappedImage::ISA = OpenCL::Mapped::;
635}
636
170=head2 THE OpenCL PACKAGE 637=head2 THE OpenCL PACKAGE
171 638
172=over 4 639=over 4
173 640
174=item $int = OpenCL::errno 641=item $int = OpenCL::errno
175 642
176The last error returned by a function - it's only changed on errors. 643The last error returned by a function - it's only valid after an error occured
644and before calling another OpenCL function.
177 645
178=item $str = OpenCL::err2str $errval 646=item $str = OpenCL::err2str [$errval]
179 647
180Comverts an error value into a human readable string. 648Converts an error value into a human readable string. If no error value is
649given, then the last error will be used (as returned by OpenCL::errno).
181 650
651The latest version of this module knows the error constants
652listed in L<http://cvs.schmorp.de/OpenCL/errstr.h>.
653
182=item $str = OpenCL::err2str $enum 654=item $str = OpenCL::enum2str $enum
183 655
184Converts most enum values (inof parameter names, image format constants, 656Converts most enum values (of parameter names, image format constants,
185object types, addressing and filter modes, command types etc.) into a 657object types, addressing and filter modes, command types etc.) into a
186human readbale string. When confronted with some random integer it can be 658human readable string. When confronted with some random integer it can be
187very helpful to pass it through this function to maybe get some readable 659very helpful to pass it through this function to maybe get some readable
188string out of it. 660string out of it.
189 661
662The latest version of this module knows the enumaration constants
663listed in L<http://cvs.schmorp.de/OpenCL/enumstr.h>.
664
190=item @platforms = OpenCL::platforms 665=item @platforms = OpenCL::platforms
191 666
192Returns all available OpenCL::Platform objects. 667Returns all available OpenCL::Platform objects.
193 668
194L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html> 669L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html>
195 670
196=item $ctx = OpenCL::context_from_type_simple $type = OpenCL::DEVICE_TYPE_DEFAULT 671=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr
197 672
198Tries to create a context from a default device and platform - never worked for me. 673Tries to create a context from a default device and platform type - never worked for me.
674Consider using C<< $platform->context_from_type >> instead.
675
676type: OpenCL::DEVICE_TYPE_DEFAULT, OpenCL::DEVICE_TYPE_CPU, OpenCL::DEVICE_TYPE_GPU,
677OpenCL::DEVICE_TYPE_ACCELERATOR, OpenCL::DEVICE_TYPE_CUSTOM, OpenCL::DEVICE_TYPE_ALL.
199 678
200L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> 679L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
201 680
681=item $ctx = OpenCL::context $properties, \@devices, $callback->($err, $pvt) = $print_stderr)
682
683Create a new OpenCL::Context object using the given device object(s).
684Consider using C<< $platform->context >> instead.
685
686L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
687
202=item OpenCL::wait_for_events $wait_events... 688=item OpenCL::wait_for_events $wait_events...
203 689
204Waits for all events to complete. 690Waits for all events to complete.
205 691
206L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> 692L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
207 693
694=item OpenCL::poll
695
696Checks if there are any outstanding events (see L<EVENT SYSTEM>) and
697invokes their callbacks.
698
699=item $OpenCL::INTERRUPT
700
701The L<Async::Interrupt> object used to signal asynchronous events (see
702L<EVENT SYSTEM>).
703
704=cut
705
706our $INTERRUPT = new Async::Interrupt c_cb => [$POLL_FUNC, 0];
707
708&_eq_initialise ($INTERRUPT->signal_func);
709
710=item $OpenCL::WATCHER
711
712The L<AnyEvent> watcher object used to watch for asynchronous events (see
713L<EVENT SYSTEM>). This variable is C<undef> until L<AnyEvent> has been
714loaded I<and> initialised (e.g. by calling C<AnyEvent::detect>).
715
716=cut
717
718our $WATCHER;
719
720sub _init_anyevent {
721 $INTERRUPT->block;
722 $WATCHER = AE::io ($INTERRUPT->pipe_fileno, 0, sub { $INTERRUPT->handle });
723}
724
725if (defined $AnyEvent::MODEL) {
726 _init_anyevent;
727} else {
728 push @AnyEvent::post_detect, \&_init_anyevent;
729}
730
208=back 731=back
209 732
733=head2 THE OpenCL::Object CLASS
734
735This is the base class for all objects in the OpenCL module. The only
736method it implements is the C<id> method, which is only useful if you want
737to interface to OpenCL on the C level.
738
739=over 4
740
741=item $iv = $obj->id
742
743OpenCL objects are represented by pointers or integers on the C level. If
744you want to interface to an OpenCL object directly on the C level, then
745you need this value, which is returned by this method. You should use an
746C<IV> type in your code and cast that to the correct type.
747
748=cut
749
750sub OpenCL::Object::id {
751 ref $_[0] eq "SCALAR"
752 ? ${ $_[0] }
753 : $_[0][0]
754}
755
756=back
757
210=head2 THE OpenCL::Platform CLASS 758=head2 THE OpenCL::Platform CLASS
211 759
212=over 4 760=over 4
213 761
762=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL)
763
764Returns a list of matching OpenCL::Device objects.
765
766=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr)
767
768Tries to create a context. Never worked for me, and you need devices explicitly anyway.
769
770L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html>
771
772=item $ctx = $platform->context ($properties, \@devices, $callback->($err, $pvt) = $print_stderr)
773
774Create a new OpenCL::Context object using the given device object(s)- a
775OpenCL::CONTEXT_PLATFORM property is supplied automatically.
776
777L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html>
778
214=item $packed_value = $platform->info ($name) 779=item $packed_value = $platform->info ($name)
215 780
216Calls C<clGetPlatformInfo> and returns the packed, raw value - for 781Calls C<clGetPlatformInfo> and returns the packed, raw value - for
217strings, this will be the string, for other values you probably need to 782strings, this will be the string (possibly including terminating \0), for
218use the correct C<unpack>. This might get improved in the future. Hopefully. 783other values you probably need to use the correct C<unpack>.
784
785It's best to avoid this method and use one of the following convenience
786wrappers.
219 787
220L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html> 788L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html>
221 789
222=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL) 790=item $platform->unload_compiler
223 791
224Returns a list of matching OpenCL::Device objects. 792Attempts to unload the compiler for this platform, for endless
793profit. Does nothing on OpenCL 1.1.
225 794
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> 795L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clUnloadPlatformCompiler.html>
796
797=for gengetinfo begin platform
798
799=item $string = $platform->profile
800
801Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_PROFILE> and returns the result.
802
803=item $string = $platform->version
804
805Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_VERSION> and returns the result.
806
807=item $string = $platform->name
808
809Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_NAME> and returns the result.
810
811=item $string = $platform->vendor
812
813Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_VENDOR> and returns the result.
814
815=item $string = $platform->extensions
816
817Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_EXTENSIONS> and returns the result.
818
819=for gengetinfo end platform
231 820
232=back 821=back
233 822
234=head2 THE OpenCL::Device CLASS 823=head2 THE OpenCL::Device CLASS
235 824
237 826
238=item $packed_value = $device->info ($name) 827=item $packed_value = $device->info ($name)
239 828
240See C<< $platform->info >> for details. 829See C<< $platform->info >> for details.
241 830
831type: OpenCL::DEVICE_TYPE_DEFAULT, OpenCL::DEVICE_TYPE_CPU,
832OpenCL::DEVICE_TYPE_GPU, OpenCL::DEVICE_TYPE_ACCELERATOR,
833OpenCL::DEVICE_TYPE_CUSTOM, OpenCL::DEVICE_TYPE_ALL.
834
835fp_config: OpenCL::FP_DENORM, OpenCL::FP_INF_NAN, OpenCL::FP_ROUND_TO_NEAREST,
836OpenCL::FP_ROUND_TO_ZERO, OpenCL::FP_ROUND_TO_INF, OpenCL::FP_FMA,
837OpenCL::FP_SOFT_FLOAT, OpenCL::FP_CORRECTLY_ROUNDED_DIVIDE_SQRT.
838
839mem_cache_type: OpenCL::NONE, OpenCL::READ_ONLY_CACHE, OpenCL::READ_WRITE_CACHE.
840
841local_mem_type: OpenCL::LOCAL, OpenCL::GLOBAL.
842
843exec_capabilities: OpenCL::EXEC_KERNEL, OpenCL::EXEC_NATIVE_KERNEL.
844
845command_queue_properties: OpenCL::QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE,
846OpenCL::QUEUE_PROFILING_ENABLE.
847
848partition_properties: OpenCL::DEVICE_PARTITION_EQUALLY,
849OpenCL::DEVICE_PARTITION_BY_COUNTS, OpenCL::DEVICE_PARTITION_BY_COUNTS_LIST_END,
850OpenCL::DEVICE_PARTITION_BY_AFFINITY_DOMAIN.
851
852affinity_domain: OpenCL::DEVICE_AFFINITY_DOMAIN_NUMA,
853OpenCL::DEVICE_AFFINITY_DOMAIN_L4_CACHE, OpenCL::DEVICE_AFFINITY_DOMAIN_L3_CACHE,
854OpenCL::DEVICE_AFFINITY_DOMAIN_L2_CACHE, OpenCL::DEVICE_AFFINITY_DOMAIN_L1_CACHE,
855OpenCL::DEVICE_AFFINITY_DOMAIN_NEXT_PARTITIONABLE.
856
242L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html> 857L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html>
243 858
244=item $ctx = $device->context_simple 859=item @devices = $device->sub_devices (\@properties)
245 860
246Convenience function to create a new OpenCL::Context object. 861Creates OpencL::SubDevice objects by partitioning an existing device.
247 862
248L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html> 863L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateSubDevices.html>
864
865=for gengetinfo begin device
866
867=item $device_type = $device->type
868
869Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_TYPE> and returns the result.
870
871=item $uint = $device->vendor_id
872
873Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VENDOR_ID> and returns the result.
874
875=item $uint = $device->max_compute_units
876
877Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_COMPUTE_UNITS> and returns the result.
878
879=item $uint = $device->max_work_item_dimensions
880
881Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_ITEM_DIMENSIONS> and returns the result.
882
883=item $int = $device->max_work_group_size
884
885Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_GROUP_SIZE> and returns the result.
886
887=item @ints = $device->max_work_item_sizes
888
889Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_ITEM_SIZES> and returns the result.
890
891=item $uint = $device->preferred_vector_width_char
892
893Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_CHAR> and returns the result.
894
895=item $uint = $device->preferred_vector_width_short
896
897Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_SHORT> and returns the result.
898
899=item $uint = $device->preferred_vector_width_int
900
901Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_INT> and returns the result.
902
903=item $uint = $device->preferred_vector_width_long
904
905Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_LONG> and returns the result.
906
907=item $uint = $device->preferred_vector_width_float
908
909Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_FLOAT> and returns the result.
910
911=item $uint = $device->preferred_vector_width_double
912
913Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE> and returns the result.
914
915=item $uint = $device->max_clock_frequency
916
917Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CLOCK_FREQUENCY> and returns the result.
918
919=item $bitfield = $device->address_bits
920
921Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ADDRESS_BITS> and returns the result.
922
923=item $uint = $device->max_read_image_args
924
925Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_READ_IMAGE_ARGS> and returns the result.
926
927=item $uint = $device->max_write_image_args
928
929Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WRITE_IMAGE_ARGS> and returns the result.
930
931=item $ulong = $device->max_mem_alloc_size
932
933Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_MEM_ALLOC_SIZE> and returns the result.
934
935=item $int = $device->image2d_max_width
936
937Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE2D_MAX_WIDTH> and returns the result.
938
939=item $int = $device->image2d_max_height
940
941Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE2D_MAX_HEIGHT> and returns the result.
942
943=item $int = $device->image3d_max_width
944
945Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_WIDTH> and returns the result.
946
947=item $int = $device->image3d_max_height
948
949Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_HEIGHT> and returns the result.
950
951=item $int = $device->image3d_max_depth
952
953Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_DEPTH> and returns the result.
954
955=item $uint = $device->image_support
956
957Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE_SUPPORT> and returns the result.
958
959=item $int = $device->max_parameter_size
960
961Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_PARAMETER_SIZE> and returns the result.
962
963=item $uint = $device->max_samplers
964
965Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_SAMPLERS> and returns the result.
966
967=item $uint = $device->mem_base_addr_align
968
969Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MEM_BASE_ADDR_ALIGN> and returns the result.
970
971=item $uint = $device->min_data_type_align_size
972
973Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MIN_DATA_TYPE_ALIGN_SIZE> and returns the result.
974
975=item $device_fp_config = $device->single_fp_config
976
977Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_SINGLE_FP_CONFIG> and returns the result.
978
979=item $device_mem_cache_type = $device->global_mem_cache_type
980
981Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHE_TYPE> and returns the result.
982
983=item $uint = $device->global_mem_cacheline_size
984
985Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHELINE_SIZE> and returns the result.
986
987=item $ulong = $device->global_mem_cache_size
988
989Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHE_SIZE> and returns the result.
990
991=item $ulong = $device->global_mem_size
992
993Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_SIZE> and returns the result.
994
995=item $ulong = $device->max_constant_buffer_size
996
997Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CONSTANT_BUFFER_SIZE> and returns the result.
998
999=item $uint = $device->max_constant_args
1000
1001Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CONSTANT_ARGS> and returns the result.
1002
1003=item $device_local_mem_type = $device->local_mem_type
1004
1005Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_LOCAL_MEM_TYPE> and returns the result.
1006
1007=item $ulong = $device->local_mem_size
1008
1009Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_LOCAL_MEM_SIZE> and returns the result.
1010
1011=item $boolean = $device->error_correction_support
1012
1013Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ERROR_CORRECTION_SUPPORT> and returns the result.
1014
1015=item $int = $device->profiling_timer_resolution
1016
1017Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PROFILING_TIMER_RESOLUTION> and returns the result.
1018
1019=item $boolean = $device->endian_little
1020
1021Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ENDIAN_LITTLE> and returns the result.
1022
1023=item $boolean = $device->available
1024
1025Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_AVAILABLE> and returns the result.
1026
1027=item $boolean = $device->compiler_available
1028
1029Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_COMPILER_AVAILABLE> and returns the result.
1030
1031=item $device_exec_capabilities = $device->execution_capabilities
1032
1033Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_EXECUTION_CAPABILITIES> and returns the result.
1034
1035=item $command_queue_properties = $device->properties
1036
1037Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_QUEUE_PROPERTIES> and returns the result.
1038
1039=item $ = $device->platform
1040
1041Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PLATFORM> and returns the result.
1042
1043=item $string = $device->name
1044
1045Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NAME> and returns the result.
1046
1047=item $string = $device->vendor
1048
1049Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VENDOR> and returns the result.
1050
1051=item $string = $device->driver_version
1052
1053Calls C<clGetDeviceInfo> with C<OpenCL::DRIVER_VERSION> and returns the result.
1054
1055=item $string = $device->profile
1056
1057Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PROFILE> and returns the result.
1058
1059=item $string = $device->version
1060
1061Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VERSION> and returns the result.
1062
1063=item $string = $device->extensions
1064
1065Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_EXTENSIONS> and returns the result.
1066
1067=item $uint = $device->preferred_vector_width_half
1068
1069Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_HALF> and returns the result.
1070
1071=item $uint = $device->native_vector_width_char
1072
1073Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_CHAR> and returns the result.
1074
1075=item $uint = $device->native_vector_width_short
1076
1077Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_SHORT> and returns the result.
1078
1079=item $uint = $device->native_vector_width_int
1080
1081Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_INT> and returns the result.
1082
1083=item $uint = $device->native_vector_width_long
1084
1085Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_LONG> and returns the result.
1086
1087=item $uint = $device->native_vector_width_float
1088
1089Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_FLOAT> and returns the result.
1090
1091=item $uint = $device->native_vector_width_double
1092
1093Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE> and returns the result.
1094
1095=item $uint = $device->native_vector_width_half
1096
1097Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_HALF> and returns the result.
1098
1099=item $device_fp_config = $device->double_fp_config
1100
1101Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_DOUBLE_FP_CONFIG> and returns the result.
1102
1103=item $device_fp_config = $device->half_fp_config
1104
1105Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_HALF_FP_CONFIG> and returns the result.
1106
1107=item $boolean = $device->host_unified_memory
1108
1109Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_HOST_UNIFIED_MEMORY> and returns the result.
1110
1111=item $device = $device->parent_device_ext
1112
1113Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARENT_DEVICE_EXT> and returns the result.
1114
1115=item @device_partition_property_exts = $device->partition_types_ext
1116
1117Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARTITION_TYPES_EXT> and returns the result.
1118
1119=item @device_partition_property_exts = $device->affinity_domains_ext
1120
1121Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_AFFINITY_DOMAINS_EXT> and returns the result.
1122
1123=item $uint = $device->reference_count_ext
1124
1125Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_REFERENCE_COUNT_EXT> and returns the result.
1126
1127=item @device_partition_property_exts = $device->partition_style_ext
1128
1129Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARTITION_STYLE_EXT> and returns the result.
1130
1131=for gengetinfo end device
249 1132
250=back 1133=back
251 1134
252=head2 THE OpenCL::Context CLASS 1135=head2 THE OpenCL::Context CLASS
253 1136
1137An OpenCL::Context is basically a container, or manager, for a number of
1138devices of a platform. It is used to create all sorts of secondary objects
1139such as buffers, queues, programs and so on.
1140
1141All context creation functions and methods take a list of properties
1142(type-value pairs). All property values can be specified as integers -
1143some additionally support other types:
1144
254=over 4 1145=over 4
255 1146
256=item $packed_value = $ctx->info ($name) 1147=item OpenCL::CONTEXT_PLATFORM
257 1148
258See C<< $platform->info >> for details. 1149Also accepts OpenCL::Platform objects.
259 1150
260L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html> 1151=item OpenCL::GLX_DISPLAY_KHR
261 1152
262=item $queue = $ctx->command_queue_simple ($device) 1153Also accepts C<undef>, in which case a deep and troubling hack is engaged
1154to find the current glx display (see L<GLX SUPPORT>).
263 1155
1156=item OpenCL::GL_CONTEXT_KHR
1157
1158Also accepts C<undef>, in which case a deep and troubling hack is engaged
1159to find the current glx context (see L<GLX SUPPORT>).
1160
1161=back
1162
1163=over 4
1164
1165=item $prog = $ctx->build_program ($program, $options = "")
1166
1167This convenience function tries to build the program on all devices in
1168the context. If the build fails, then the function will C<croak> with the
1169build log. Otherwise ti returns the program object.
1170
1171The C<$program> can either be a C<OpenCL::Program> object or a string
1172containing the program. In the latter case, a program objetc will be
1173created automatically.
1174
1175=cut
1176
1177sub OpenCL::Context::build_program {
1178 my ($self, $prog, $options) = @_;
1179
1180 $prog = $self->program_with_source ($prog)
1181 unless ref $prog;
1182
1183 eval { $prog->build (undef, $options); 1 }
1184 or errno == BUILD_PROGRAM_FAILURE
1185 or errno == INVALID_BINARY # workaround nvidia bug
1186 or Carp::croak "OpenCL::Context->build_program: " . err2str;
1187
1188 # we check status for all devices
1189 for my $dev ($self->devices) {
1190 $prog->build_status ($dev) == BUILD_SUCCESS
1191 or Carp::croak "Building OpenCL program for device '" . $dev->name . "' failed:\n"
1192 . $prog->build_log ($dev);
1193 }
1194
1195 $prog
1196}
1197
1198=item $queue = $ctx->queue ($device, $properties)
1199
264Convenience function to create a new OpenCL::Queue object from the context and the given device. 1200Create a new OpenCL::Queue object from the context and the given device.
265 1201
266L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html> 1202L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html>
267 1203
1204Example: create an out-of-order queue.
1205
1206 $queue = $ctx->queue ($device, OpenCL::QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE);
1207
268=item $ev = $ctx->user_event 1208=item $ev = $ctx->user_event
269 1209
270Creates a new OpenCL::UserEvent object. 1210Creates a new OpenCL::UserEvent object.
271 1211
272L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html> 1212L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html>
273 1213
274=item $buf = $ctx->buffer ($flags, $len) 1214=item $buf = $ctx->buffer ($flags, $len)
275 1215
276Creates a new OpenCL::Buffer object with the given flags and octet-size. 1216Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object with the
1217given flags and octet-size.
1218
1219flags: OpenCL::MEM_READ_WRITE, OpenCL::MEM_WRITE_ONLY, OpenCL::MEM_READ_ONLY,
1220OpenCL::MEM_USE_HOST_PTR, OpenCL::MEM_ALLOC_HOST_PTR, OpenCL::MEM_COPY_HOST_PTR,
1221OpenCL::MEM_HOST_WRITE_ONLY, OpenCL::MEM_HOST_READ_ONLY, OpenCL::MEM_HOST_NO_ACCESS.
277 1222
278L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html> 1223L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html>
279 1224
280=item $buf = $ctx->buffer_sv ($flags, $data) 1225=item $buf = $ctx->buffer_sv ($flags, $data)
281 1226
282Creates a new OpenCL::Buffer object and initialise it with the given data values. 1227Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object and
1228initialise it with the given data values.
283 1229
1230=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)
1231
1232Creates a new OpenCL::Image object and optionally initialises it with
1233the given data values.
1234
1235channel_order: OpenCL::R, OpenCL::A, OpenCL::RG, OpenCL::RA, OpenCL::RGB,
1236OpenCL::RGBA, OpenCL::BGRA, OpenCL::ARGB, OpenCL::INTENSITY, OpenCL::LUMINANCE,
1237OpenCL::Rx, OpenCL::RGx, OpenCL::RGBx.
1238
1239channel_type: OpenCL::SNORM_INT8, OpenCL::SNORM_INT16, OpenCL::UNORM_INT8,
1240OpenCL::UNORM_INT16, OpenCL::UNORM_SHORT_565, OpenCL::UNORM_SHORT_555,
1241OpenCL::UNORM_INT_101010, OpenCL::SIGNED_INT8, OpenCL::SIGNED_INT16,
1242OpenCL::SIGNED_INT32, OpenCL::UNSIGNED_INT8, OpenCL::UNSIGNED_INT16,
1243OpenCL::UNSIGNED_INT32, OpenCL::HALF_FLOAT, OpenCL::FLOAT.
1244
1245type: OpenCL::MEM_OBJECT_BUFFER, OpenCL::MEM_OBJECT_IMAGE2D,
1246OpenCL::MEM_OBJECT_IMAGE3D, OpenCL::MEM_OBJECT_IMAGE2D_ARRAY,
1247OpenCL::MEM_OBJECT_IMAGE1D, OpenCL::MEM_OBJECT_IMAGE1D_ARRAY,
1248OpenCL::MEM_OBJECT_IMAGE1D_BUFFER.
1249
1250L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage.html>
1251
284=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $data) 1252=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $row_pitch = 0, $data = undef)
285 1253
286Creates a new OpenCL::Image2D object and optionally initialises it with the given data values. 1254Creates a new OpenCL::Image2D object and optionally initialises it with
1255the given data values.
287 1256
288L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html> 1257L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html>
289 1258
290=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $slice_pitch, $data) 1259=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $row_pitch = 0, $slice_pitch = 0, $data = undef)
291 1260
292Creates a new OpenCL::Image3D object and optionally initialises it with the given data values. 1261Creates a new OpenCL::Image3D object and optionally initialises it with
1262the given data values.
293 1263
294L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html> 1264L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html>
1265
1266=item $buffer = $ctx->gl_buffer ($flags, $bufobj)
1267
1268Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object that refers to the given
1269OpenGL buffer object.
1270
1271flags: OpenCL::MEM_READ_WRITE, OpenCL::MEM_READ_ONLY, OpenCL::MEM_WRITE_ONLY.
1272
1273http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLBuffer.html
1274
1275=item $img = $ctx->gl_texture ($flags, $target, $miplevel, $texture)
1276
1277Creates a new OpenCL::Image object that refers to the given OpenGL
1278texture object or buffer.
1279
1280target: GL_TEXTURE_1D, GL_TEXTURE_1D_ARRAY, GL_TEXTURE_BUFFER,
1281GL_TEXTURE_2D, GL_TEXTURE_2D_ARRAY, GL_TEXTURE_3D,
1282GL_TEXTURE_CUBE_MAP_POSITIVE_X, GL_TEXTURE_CUBE_MAP_POSITIVE_Y,
1283GL_TEXTURE_CUBE_MAP_POSITIVE_Z, GL_TEXTURE_CUBE_MAP_NEGATIVE_X,
1284GL_TEXTURE_CUBE_MAP_NEGATIVE_Y, GL_TEXTURE_CUBE_MAP_NEGATIVE_Z,
1285GL_TEXTURE_RECTANGLE/GL_TEXTURE_RECTANGLE_ARB.
1286
1287http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture.html
1288
1289=item $img = $ctx->gl_texture2d ($flags, $target, $miplevel, $texture)
1290
1291Creates a new OpenCL::Image2D object that refers to the given OpenGL
12922D texture object.
1293
1294http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture2D.html
1295
1296=item $img = $ctx->gl_texture3d ($flags, $target, $miplevel, $texture)
1297
1298Creates a new OpenCL::Image3D object that refers to the given OpenGL
12993D texture object.
1300
1301http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture3D.html
1302
1303=item $ctx->gl_renderbuffer ($flags, $renderbuffer)
1304
1305Creates a new OpenCL::Image2D object that refers to the given OpenGL
1306render buffer.
1307
1308http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLRenderbuffer.html
295 1309
296=item @formats = $ctx->supported_image_formats ($flags, $image_type) 1310=item @formats = $ctx->supported_image_formats ($flags, $image_type)
297 1311
298Returns a list of matching image formats - each format is an arrayref with 1312Returns a list of matching image formats - each format is an arrayref with
299two values, $channel_order and $channel_type, in it. 1313two values, $channel_order and $channel_type, in it.
302 1316
303=item $sampler = $ctx->sampler ($normalized_coords, $addressing_mode, $filter_mode) 1317=item $sampler = $ctx->sampler ($normalized_coords, $addressing_mode, $filter_mode)
304 1318
305Creates a new OpenCL::Sampler object. 1319Creates a new OpenCL::Sampler object.
306 1320
1321addressing_mode: OpenCL::ADDRESS_NONE, OpenCL::ADDRESS_CLAMP_TO_EDGE,
1322OpenCL::ADDRESS_CLAMP, OpenCL::ADDRESS_REPEAT, OpenCL::ADDRESS_MIRRORED_REPEAT.
1323
1324filter_mode: OpenCL::FILTER_NEAREST, OpenCL::FILTER_LINEAR.
1325
307L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSampler.html> 1326L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSampler.html>
308 1327
309=item $program = $ctx->program_with_source ($string) 1328=item $program = $ctx->program_with_source ($string)
310 1329
311Creates a new OpenCL::Program object from the given source code. 1330Creates a new OpenCL::Program object from the given source code.
312 1331
313L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html> 1332L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html>
314 1333
1334=item ($program, \@status) = $ctx->program_with_binary (\@devices, \@binaries)
1335
1336Creates a new OpenCL::Program object from the given binaries.
1337
1338L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithBinary.html>
1339
1340Example: clone an existing program object that contains a successfully
1341compiled program, no matter how useless this is.
1342
1343 my $clone = $ctx->program_with_binary ([$prog->devices], [$prog->binaries]);
1344
1345=item $program = $ctx->program_with_built_in_kernels (\@devices, $kernel_names)
1346
1347Creates a new OpenCL::Program object from the given built-in kernel names.
1348
1349L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateProgramWithBuiltInKernels.html>
1350
1351=item $program = $ctx->link_program (\@devices, $options, \@programs, $cb->($program) = undef)
1352
1353Links all (already compiled) program objects specified in C<@programs>
1354together and returns a new OpenCL::Program object with the result.
1355
1356L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clLinkProgram.html>
1357
1358=item $packed_value = $ctx->info ($name)
1359
1360See C<< $platform->info >> for details.
1361
1362L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html>
1363
1364=for gengetinfo begin context
1365
1366=item $uint = $context->reference_count
1367
1368Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_REFERENCE_COUNT> and returns the result.
1369
1370=item @devices = $context->devices
1371
1372Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_DEVICES> and returns the result.
1373
1374=item @property_ints = $context->properties
1375
1376Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_PROPERTIES> and returns the result.
1377
1378=item $uint = $context->num_devices
1379
1380Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_NUM_DEVICES> and returns the result.
1381
1382=for gengetinfo end context
1383
315=back 1384=back
316 1385
317=head2 THE OpenCL::Queue CLASS 1386=head2 THE OpenCL::Queue CLASS
318 1387
319An OpenCL::Queue represents an execution queue for OpenCL. You execute 1388An OpenCL::Queue represents an execution queue for OpenCL. You execute
320requests by calling their respective C<enqueue_xxx> method and waitinf for 1389requests by calling their respective method and waiting for it to complete
321it to complete in some way. 1390in some way.
322 1391
323All the enqueue methods return an event object that can be used to wait 1392Most methods that enqueue some request return an event object that can
324for completion, unless the method is called in void context, in which case 1393be used to wait for completion (optionally using a callback), unless
325no event object is created. 1394the method is called in void context, in which case no event object is
1395created.
326 1396
327They also allow you to specify any number of other event objects that this 1397They also allow you to specify any number of other event objects that this
328request has to wait for before it starts executing, by simply passing the 1398request has to wait for before it starts executing, by simply passing the
329event objects as extra parameters to the enqueue methods. 1399event objects as extra parameters to the enqueue methods. To simplify
1400program design, this module ignores any C<undef> values in the list of
1401events. This makes it possible to code operations such as this, without
1402having to put a valid event object into C<$event> first:
1403
1404 $event = $queue->xxx (..., $event);
330 1405
331Queues execute in-order by default, without any parallelism, so in most 1406Queues execute in-order by default, without any parallelism, so in most
332cases it's not necessary to wait for or create event objects. 1407cases (i.e. you use only one queue) it's not necessary to wait for or
1408create event objects, althoguh an our of order queue is often a bit
1409faster.
333 1410
334=over 4 1411=over 4
335 1412
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...) 1413=item $ev = $queue->read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...)
343 1414
344Reads data from buffer into the given string. 1415Reads data from buffer into the given string.
345 1416
346L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html> 1417L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html>
347 1418
348=item $ev = $queue->enqueue_write_buffer ($buffer, $blocking, $offset, $data, $wait_events...) 1419=item $ev = $queue->write_buffer ($buffer, $blocking, $offset, $data, $wait_events...)
349 1420
350Writes data to buffer from the given string. 1421Writes data to buffer from the given string.
351 1422
352L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html> 1423L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html>
353 1424
354=item $ev = $queue->enqueue_copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...) 1425=item $ev = $queue->copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...)
355 1426
356L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html> 1427L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html>
357 1428
1429$eue->read_buffer_rect ($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...)
1430
1431http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBufferRect.html
1432
1433=item $ev = $queue->write_buffer_rect ($buf, $blocking, $buf_y, $host_x, $host_z, $height, $buf_row_pitch, $host_row_pitch, $data, $wait_events...)
1434
1435http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBufferRect.html
1436
1437=item $ev = $queue->copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...)
1438
1439L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>
1440
358=item $ev = $queue->enqueue_read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) 1441=item $ev = $queue->read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
1442
1443C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
1444module uses the image width (and height) to supply default values.
359 1445
360L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html> 1446L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html>
361 1447
362=item $ev = $queue->enqueue_write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $data, $wait_events...) 1448=item $ev = $queue->write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...)
363 1449
1450C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL
1451module 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> 1452L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html>
365 1453
1454=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...)
1455
1456L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html>
1457
1458=item $ev = $queue->copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...)
1459
1460L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html>
1461
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...) 1462=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 1463
368Yeah. 1464Yeah.
369 1465
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>. 1466L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>.
375 1467
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, ...) 1468=item $ev = $queue->fill_buffer ($mem, $pattern, $offset, $size, ...)
377 1469
1470Fills the given buffer object with repeated applications of C<$pattern>,
1471starting at C<$offset> for C<$size> octets.
1472
1473L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillBuffer.html>
1474
1475=item $ev = $queue->fill_image ($img, $r, $g, $b, $a, $x, $y, $z, $width, $height, $depth, ...)
1476
1477Fills the given image area with the given rgba colour components. The
1478components are normally floating point values between C<0> and C<1>,
1479except when the image channel data type is a signe dor unsigned
1480unnormalised format, in which case the range is determined by the format.
1481
378L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html> 1482L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillImage.html>
379 1483
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...) 1484=item $ev = $queue->task ($kernel, $wait_events...)
385 1485
386L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html> 1486L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html>
387 1487
388=item $ev = $queue->enqueue_nd_range_kernel ($kernel, @$global_work_offset, @$global_work_size, @$local_work_size, $wait_events...) 1488=item $ev = $queue->nd_range_kernel ($kernel, \@global_work_offset, \@global_work_size, \@local_work_size, $wait_events...)
389 1489
390Enqueues a kernel execution. 1490Enqueues a kernel execution.
391 1491
392@$global_work_size must be specified as a reference to an array of 1492\@global_work_size must be specified as a reference to an array of
393integers specifying the work sizes (element counts). 1493integers specifying the work sizes (element counts).
394 1494
395@$global_work_offset must be either C<undef> (in which case all offsets 1495\@global_work_offset must be either C<undef> (in which case all offsets
396are C<0>), or a reference to an array of work offsets, with the same number 1496are C<0>), or a reference to an array of work offsets, with the same number
397of elements as @$global_work_size. 1497of elements as \@global_work_size.
398 1498
399@$local_work_size must be either C<undef> (in which case the 1499\@local_work_size must be either C<undef> (in which case the
400implementation is supposed to choose good local work sizes), or a 1500implementation is supposed to choose good local work sizes), or a
401reference to an array of local work sizes, with the same number of 1501reference to an array of local work sizes, with the same number of
402elements as @$global_work_size. 1502elements as \@global_work_size.
403 1503
404L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html> 1504L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html>
405 1505
406=item $ev = $queue->enqueue_marker 1506=item $ev = $queue->migrate_mem_objects (\@mem_objects, $flags, $wait_events...)
407 1507
1508Migrates a number of OpenCL::Memory objects to or from the device.
1509
1510flags: OpenCL::MIGRATE_MEM_OBJECT_HOST, OpenCL::MIGRATE_MEM_OBJECT_CONTENT_UNDEFINED
1511
1512L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMigrateMemObjects.html>
1513
1514=item $ev = $queue->acquire_gl_objects ([object, ...], $wait_events...)
1515
1516Enqueues a list (an array-ref of OpenCL::Memory objects) to be acquired
1517for subsequent OpenCL usage.
1518
408L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMarker.html> 1519L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueAcquireGLObjects.html>
409 1520
1521=item $ev = $queue->release_gl_objects ([object, ...], $wait_events...)
1522
1523Enqueues a list (an array-ref of OpenCL::Memory objects) to be released
1524for subsequent OpenGL usage.
1525
1526L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReleaseGLObjects.html>
1527
410=item $ev = $queue->enqueue_wait_for_events ($wait_events...) 1528=item $ev = $queue->wait_for_events ($wait_events...)
411 1529
412L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html> 1530L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html>
413 1531
414=item $queue->enqueue_barrier 1532=item $ev = $queue->marker ($wait_events...)
415 1533
1534L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMarkerWithWaitList.html>
1535
1536=item $ev = $queue->barrier ($wait_events...)
1537
416L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueBarrier.html> 1538L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueBarrierWithWaitList.html>
417 1539
418=item $queue->flush 1540=item $queue->flush
419 1541
420L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html> 1542L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html>
421 1543
422=item $queue->finish 1544=item $queue->finish
423 1545
424L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html> 1546L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html>
425 1547
1548=item $packed_value = $queue->info ($name)
1549
1550See C<< $platform->info >> for details.
1551
1552L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetCommandQueueInfo.html>
1553
1554=for gengetinfo begin command_queue
1555
1556=item $ctx = $command_queue->context
1557
1558Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_CONTEXT> and returns the result.
1559
1560=item $device = $command_queue->device
1561
1562Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_DEVICE> and returns the result.
1563
1564=item $uint = $command_queue->reference_count
1565
1566Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_REFERENCE_COUNT> and returns the result.
1567
1568=item $command_queue_properties = $command_queue->properties
1569
1570Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_PROPERTIES> and returns the result.
1571
1572=for gengetinfo end command_queue
1573
426=back 1574=back
427 1575
1576=head3 MEMORY MAPPED BUFFERS
1577
1578OpenCL allows you to map buffers and images to host memory (read: perl
1579scalars). This is done much like reading or copying a buffer, by enqueuing
1580a map or unmap operation on the command queue.
1581
1582The map operations return an C<OpenCL::Mapped> object - see L<THE
1583OpenCL::Mapped CLASS> section for details on what to do with these
1584objects.
1585
1586The object will be unmapped automatically when the mapped object is
1587destroyed (you can use a barrier to make sure the unmap has finished,
1588before using the buffer in a kernel), but you can also enqueue an unmap
1589operation manually.
1590
1591=over 4
1592
1593=item $mapped_buffer = $queue->map_buffer ($buf, $blocking=1, $map_flags=OpenCL::MAP_READ|OpenCL::MAP_WRITE, $offset=0, $size=undef, $wait_events...)
1594
1595Maps the given buffer into host memory and returns an
1596C<OpenCL::MappedBuffer> object. If C<$size> is specified as undef, then
1597the map will extend to the end of the buffer.
1598
1599map_flags: OpenCL::MAP_READ, OpenCL::MAP_WRITE, OpenCL::MAP_WRITE_INVALIDATE_REGION.
1600
1601L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapBuffer.html>
1602
1603Example: map the buffer $buf fully and replace the first 4 bytes by "abcd", then unmap.
1604
1605 {
1606 my $mapped = $queue->map_buffer ($buf, 1, OpenCL::MAP_WRITE);
1607 substr $$mapped, 0, 4, "abcd";
1608 } # asynchronously unmap because $mapped is destroyed
1609
1610=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...)
1611
1612Maps the given image area into host memory and return an
1613C<OpenCL::MappedImage> object.
1614
1615If any of C<$width>, C<$height> and/or C<$depth> are C<undef> then they
1616will be replaced by the maximum possible value.
1617
1618L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapImage.html>
1619
1620Example: map an image (with OpenCL::UNSIGNED_INT8 channel type) and set
1621the first channel of the leftmost column to 5, then explicitly unmap
1622it. You are not necessarily meant to do it this way, this example just
1623shows you the accessors to use :)
1624
1625 my $mapped = $queue->map_image ($image, 1, OpenCL::MAP_WRITE);
1626
1627 $mapped->write ($_ * $mapped->row_pitch, pack "C", 5)
1628 for 0 .. $mapped->height - 1;
1629
1630 $mapped->unmap;.
1631 $mapped->wait; # only needed for out of order queues normally
1632
1633=item $ev = $queue->unmap ($mapped, $wait_events...)
1634
1635Unmaps the data from host memory. You must not call any methods that
1636modify the data, or modify the data scalar directly, after calling this
1637method.
1638
1639The mapped event object will always be passed as part of the
1640$wait_events. The mapped event object will be replaced by the new event
1641object that this request creates.
1642
1643=back
1644
428=head2 THE OpenCL::Memory CLASS 1645=head2 THE OpenCL::Memory CLASS
429 1646
430This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image, 1647This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image,
431OpenCL::Image2D and OpenCL::Image3D. The subclasses of this class 1648OpenCL::Image2D and OpenCL::Image3D.
432currently only exist to allow type-checking.
433 1649
434=over 4 1650=over 4
435 1651
436=item $packed_value = $memory->info ($name) 1652=item $packed_value = $memory->info ($name)
437 1653
438See C<< $platform->info >> for details. 1654See C<< $platform->info >> for details.
439 1655
440L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html> 1656L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html>
441 1657
1658=item $memory->destructor_callback ($cb->())
1659
1660Sets a callback that will be invoked after the memory object is destructed.
1661
1662L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetMemObjectDestructorCallback.html>
1663
1664=for gengetinfo begin mem
1665
1666=item $mem_object_type = $mem->type
1667
1668Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_TYPE> and returns the result.
1669
1670=item $mem_flags = $mem->flags
1671
1672Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_FLAGS> and returns the result.
1673
1674=item $int = $mem->size
1675
1676Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_SIZE> and returns the result.
1677
1678=item $ptr_value = $mem->host_ptr
1679
1680Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_HOST_PTR> and returns the result.
1681
1682=item $uint = $mem->map_count
1683
1684Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_MAP_COUNT> and returns the result.
1685
1686=item $uint = $mem->reference_count
1687
1688Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_REFERENCE_COUNT> and returns the result.
1689
1690=item $ctx = $mem->context
1691
1692Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_CONTEXT> and returns the result.
1693
1694=item $mem = $mem->associated_memobject
1695
1696Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_ASSOCIATED_MEMOBJECT> and returns the result.
1697
1698=item $int = $mem->offset
1699
1700Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_OFFSET> and returns the result.
1701
1702=for gengetinfo end mem
1703
1704=item ($type, $name) = $mem->gl_object_info
1705
1706Returns the OpenGL object type (e.g. OpenCL::GL_OBJECT_TEXTURE2D) and the
1707object "name" (e.g. the texture name) used to create this memory object.
1708
1709L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetGLObjectInfo.html>
1710
442=back 1711=back
443 1712
1713=head2 THE OpenCL::Buffer CLASS
1714
1715This is a subclass of OpenCL::Memory, and the superclass of
1716OpenCL::BufferObj. Its purpose is simply to distinguish between buffers
1717and sub-buffers.
1718
1719=head2 THE OpenCL::BufferObj CLASS
1720
1721This is a subclass of OpenCL::Buffer and thus OpenCL::Memory. It exists
1722because one cna create sub buffers of OpenLC::BufferObj objects, but not
1723sub buffers from these sub buffers.
1724
1725=over 4
1726
1727=item $subbuf = $buf_obj->sub_buffer_region ($flags, $origin, $size)
1728
1729Creates an OpenCL::Buffer objects from this buffer and returns it. The
1730C<buffer_create_type> is assumed to be C<OpenCL::BUFFER_CREATE_TYPE_REGION>.
1731
1732L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSubBuffer.html>
1733
1734=back
1735
1736=head2 THE OpenCL::Image CLASS
1737
1738This is the superclass of all image objects - OpenCL::Image1D,
1739OpenCL::Image1DArray, OpenCL::Image1DBuffer, OpenCL::Image2D,
1740OpenCL::Image2DArray and OpenCL::Image3D.
1741
1742=over 4
1743
1744=item $packed_value = $image->image_info ($name)
1745
1746See C<< $platform->info >> for details.
1747
1748The reason this method is not called C<info> is that there already is an
1749C<< ->info >> method inherited from C<OpenCL::Memory>.
1750
1751L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetImageInfo.html>
1752
1753=item ($channel_order, $channel_data_type) = $image->format
1754
1755Returns the channel order and type used to create the image by calling
1756C<clGetImageInfo> with C<OpenCL::IMAGE_FORMAT>.
1757
1758=for gengetinfo begin image
1759
1760=item $int = $image->element_size
1761
1762Calls C<clGetImageInfo> with C<OpenCL::IMAGE_ELEMENT_SIZE> and returns the result.
1763
1764=item $int = $image->row_pitch
1765
1766Calls C<clGetImageInfo> with C<OpenCL::IMAGE_ROW_PITCH> and returns the result.
1767
1768=item $int = $image->slice_pitch
1769
1770Calls C<clGetImageInfo> with C<OpenCL::IMAGE_SLICE_PITCH> and returns the result.
1771
1772=item $int = $image->width
1773
1774Calls C<clGetImageInfo> with C<OpenCL::IMAGE_WIDTH> and returns the result.
1775
1776=item $int = $image->height
1777
1778Calls C<clGetImageInfo> with C<OpenCL::IMAGE_HEIGHT> and returns the result.
1779
1780=item $int = $image->depth
1781
1782Calls C<clGetImageInfo> with C<OpenCL::IMAGE_DEPTH> and returns the result.
1783
1784=for gengetinfo end image
1785
1786=for gengetinfo begin gl_texture
1787
1788=item $GLenum = $gl_texture->target
1789
1790Calls C<clGetGLTextureInfo> with C<OpenCL::GL_TEXTURE_TARGET> and returns the result.
1791
1792=item $GLint = $gl_texture->gl_mipmap_level
1793
1794Calls C<clGetGLTextureInfo> with C<OpenCL::GL_MIPMAP_LEVEL> and returns the result.
1795
1796=for gengetinfo end gl_texture
1797
1798=back
1799
444=head2 THE OpenCL::Sampler CLASS 1800=head2 THE OpenCL::Sampler CLASS
445 1801
446=over 4 1802=over 4
447 1803
448=item $packed_value = $sampler->info ($name) 1804=item $packed_value = $sampler->info ($name)
449 1805
450See C<< $platform->info >> for details. 1806See C<< $platform->info >> for details.
451 1807
452L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html> 1808L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html>
453 1809
1810=for gengetinfo begin sampler
1811
1812=item $uint = $sampler->reference_count
1813
1814Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_REFERENCE_COUNT> and returns the result.
1815
1816=item $ctx = $sampler->context
1817
1818Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_CONTEXT> and returns the result.
1819
1820=item $addressing_mode = $sampler->normalized_coords
1821
1822Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_NORMALIZED_COORDS> and returns the result.
1823
1824=item $filter_mode = $sampler->addressing_mode
1825
1826Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_ADDRESSING_MODE> and returns the result.
1827
1828=item $boolean = $sampler->filter_mode
1829
1830Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_FILTER_MODE> and returns the result.
1831
1832=for gengetinfo end sampler
1833
454=back 1834=back
455 1835
456=head2 THE OpenCL::Program CLASS 1836=head2 THE OpenCL::Program CLASS
457 1837
458=over 4 1838=over 4
459 1839
460=item $packed_value = $program->info ($name) 1840=item $program->build (\@devices = undef, $options = "", $cb->($program) = undef)
461 1841
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. 1842Tries to build the program with the given options. See also the
1843C<$ctx->build> convenience function.
1844
1845If a callback is specified, then it will be called when compilation is
1846finished. Note that many OpenCL implementations block your program while
1847compiling whether you use a callback or not. See C<build_async> if you
1848want to make sure the build is done in the background.
1849
1850Note that some OpenCL implementations act up badly, and don't call the
1851callback in some error cases (but call it in others). This implementation
1852assumes the callback will always be called, and leaks memory if this is
1853not so. So best make sure you don't pass in invalid values.
1854
1855Some implementations fail with C<OpenCL::INVALID_BINARY> when the
1856compilation state is successful but some later stage fails.
1857
1858options: C<-D name>, C<-D name=definition>, C<-I dir>,
1859C<-cl-single-precision-constant>, C<-cl-denorms-are-zero>,
1860C<-cl-fp32-correctly-rounded-divide-sqrt>, C<-cl-opt-disable>,
1861C<-cl-mad-enable>, C<-cl-no-signed-zeros>, C<-cl-unsafe-math-optimizations>,
1862C<-cl-finite-math-only>, C<-cl-fast-relaxed-math>,
1863C<-w>, C<-Werror>, C<-cl-std=CL1.1/CL1.2>, C<-cl-kernel-arg-info>,
1864C<-create-library>, C<-enable-link-options>.
1865
1866build_status: OpenCL::BUILD_SUCCESS, OpenCL::BUILD_NONE,
1867OpenCL::BUILD_ERROR, OpenCL::BUILD_IN_PROGRESS.
469 1868
470L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html> 1869L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html>
1870
1871=item $program->build_async (\@devices = undef, $options = "", $cb->($program) = undef)
1872
1873Similar to C<< ->build >>, except it starts a thread, and never fails (you
1874need to check the compilation status form the callback, or by polling).
1875
1876=item $program->compile (\@devices = undef, $options = "", \%headers = undef, $cb->($program) = undef)
1877
1878Compiles the given program for the given devices (or all devices if
1879undef). If C<$headers> is given, it must be a hashref with include name =>
1880OpenCL::Program pairs.
1881
1882L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCompileProgram.html>
471 1883
472=item $packed_value = $program->build_info ($device, $name) 1884=item $packed_value = $program->build_info ($device, $name)
473 1885
474Similar to C<< $platform->info >>, but returns build info for a previous 1886Similar to C<< $platform->info >>, but returns build info for a previous
475build attempt for the given device. 1887build attempt for the given device.
476 1888
1889binary_type: OpenCL::PROGRAM_BINARY_TYPE_NONE,
1890OpenCL::PROGRAM_BINARY_TYPE_COMPILED_OBJECT,
1891OpenCL::PROGRAM_BINARY_TYPE_LIBRARY,
1892OpenCL::PROGRAM_BINARY_TYPE_EXECUTABLE.
1893
477L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetBuildInfo.html> 1894L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetBuildInfo.html>
478 1895
479=item $kernel = $program->kernel ($function_name) 1896=item $kernel = $program->kernel ($function_name)
480 1897
481Creates an OpenCL::Kernel object out of the named C<__kernel> function in 1898Creates an OpenCL::Kernel object out of the named C<__kernel> function in
482the program. 1899the program.
483 1900
484L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html> 1901L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html>
485 1902
1903=item @kernels = $program->kernels_in_program
1904
1905Returns all kernels successfully compiled for all devices in program.
1906
1907http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernelsInProgram.html
1908
1909=for gengetinfo begin program_build
1910
1911=item $build_status = $program->build_status ($device)
1912
1913Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_STATUS> and returns the result.
1914
1915=item $string = $program->build_options ($device)
1916
1917Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_OPTIONS> and returns the result.
1918
1919=item $string = $program->build_log ($device)
1920
1921Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_LOG> and returns the result.
1922
1923=item $binary_type = $program->binary_type ($device)
1924
1925Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BINARY_TYPE> and returns the result.
1926
1927=for gengetinfo end program_build
1928
1929=item $packed_value = $program->info ($name)
1930
1931See C<< $platform->info >> for details.
1932
1933L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html>
1934
1935=for gengetinfo begin program
1936
1937=item $uint = $program->reference_count
1938
1939Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_REFERENCE_COUNT> and returns the result.
1940
1941=item $ctx = $program->context
1942
1943Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_CONTEXT> and returns the result.
1944
1945=item $uint = $program->num_devices
1946
1947Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_NUM_DEVICES> and returns the result.
1948
1949=item @devices = $program->devices
1950
1951Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_DEVICES> and returns the result.
1952
1953=item $string = $program->source
1954
1955Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_SOURCE> and returns the result.
1956
1957=item @ints = $program->binary_sizes
1958
1959Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_BINARY_SIZES> and returns the result.
1960
1961=for gengetinfo end program
1962
1963=item @blobs = $program->binaries
1964
1965Returns a string for the compiled binary for every device associated with
1966the program, empty strings indicate missing programs, and an empty result
1967means no program binaries are available.
1968
1969These "binaries" are often, in fact, informative low-level assembly
1970sources.
1971
1972L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html>
1973
486=back 1974=back
487 1975
488=head2 THE OpenCL::Kernel CLASS 1976=head2 THE OpenCL::Kernel CLASS
489 1977
490=over 4 1978=over 4
493 1981
494See C<< $platform->info >> for details. 1982See C<< $platform->info >> for details.
495 1983
496L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html> 1984L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html>
497 1985
1986=for gengetinfo begin kernel
1987
1988=item $string = $kernel->function_name
1989
1990Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_FUNCTION_NAME> and returns the result.
1991
1992=item $uint = $kernel->num_args
1993
1994Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_NUM_ARGS> and returns the result.
1995
1996=item $uint = $kernel->reference_count
1997
1998Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_REFERENCE_COUNT> and returns the result.
1999
2000=item $ctx = $kernel->context
2001
2002Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_CONTEXT> and returns the result.
2003
2004=item $program = $kernel->program
2005
2006Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_PROGRAM> and returns the result.
2007
2008=for gengetinfo end kernel
2009
2010=item $packed_value = $kernel->work_group_info ($device, $name)
2011
2012See C<< $platform->info >> for details.
2013
2014L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelWorkGroupInfo.html>
2015
2016=for gengetinfo begin kernel_work_group
2017
2018=item $int = $kernel->work_group_size ($device)
2019
2020Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_WORK_GROUP_SIZE> and returns the result.
2021
2022=item @ints = $kernel->compile_work_group_size ($device)
2023
2024Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_COMPILE_WORK_GROUP_SIZE> and returns the result.
2025
2026=item $ulong = $kernel->local_mem_size ($device)
2027
2028Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_LOCAL_MEM_SIZE> and returns the result.
2029
2030=item $int = $kernel->preferred_work_group_size_multiple ($device)
2031
2032Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE> and returns the result.
2033
2034=item $ulong = $kernel->private_mem_size ($device)
2035
2036Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_PRIVATE_MEM_SIZE> and returns the result.
2037
2038=for gengetinfo end kernel_work_group
2039
2040=item $packed_value = $kernel->arg_info ($idx, $name)
2041
2042See C<< $platform->info >> for details.
2043
2044L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clGetKernelArgInfo.html>
2045
2046=for gengetinfo begin kernel_arg
2047
2048=item $kernel_arg_address_qualifier = $kernel->arg_address_qualifier ($idx)
2049
2050Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_ADDRESS_QUALIFIER> and returns the result.
2051
2052=item $kernel_arg_access_qualifier = $kernel->arg_access_qualifier ($idx)
2053
2054Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_ACCESS_QUALIFIER> and returns the result.
2055
2056=item $string = $kernel->arg_type_name ($idx)
2057
2058Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_TYPE_NAME> and returns the result.
2059
2060=item $kernel_arg_type_qualifier = $kernel->arg_type_qualifier ($idx)
2061
2062Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_TYPE_QUALIFIER> and returns the result.
2063
2064=item $string = $kernel->arg_name ($idx)
2065
2066Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_NAME> and returns the result.
2067
2068=for gengetinfo end kernel_arg
2069
2070=item $kernel->setf ($format, ...)
2071
2072Sets the arguments of a kernel. Since OpenCL 1.1 doesn't have a generic
2073way to set arguments (and with OpenCL 1.2 it might be rather slow), you
2074need to specify a format argument, much as with C<printf>, to tell OpenCL
2075what type of argument it is.
2076
2077The format arguments are single letters:
2078
2079 c char
2080 C unsigned char
2081 s short
2082 S unsigned short
2083 i int
2084 I unsigned int
2085 l long
2086 L unsigned long
2087
2088 h half float (0..65535)
2089 f float
2090 d double
2091
2092 z local (octet size)
2093
2094 m memory object (buffer or image)
2095 a sampler
2096 e event
2097
2098Space characters in the format string are ignored.
2099
2100Example: set the arguments for a kernel that expects an int, two floats, a buffer and an image.
2101
2102 $kernel->setf ("i ff mm", 5, 0.5, 3, $buffer, $image);
2103
498=item $kernel->set_TYPE ($index, $value) 2104=item $kernel->set_TYPE ($index, $value)
499 2105
2106=item $kernel->set_char ($index, $value)
2107
2108=item $kernel->set_uchar ($index, $value)
2109
2110=item $kernel->set_short ($index, $value)
2111
2112=item $kernel->set_ushort ($index, $value)
2113
2114=item $kernel->set_int ($index, $value)
2115
2116=item $kernel->set_uint ($index, $value)
2117
2118=item $kernel->set_long ($index, $value)
2119
2120=item $kernel->set_ulong ($index, $value)
2121
2122=item $kernel->set_half ($index, $value)
2123
2124=item $kernel->set_float ($index, $value)
2125
2126=item $kernel->set_double ($index, $value)
2127
2128=item $kernel->set_memory ($index, $value)
2129
2130=item $kernel->set_buffer ($index, $value)
2131
2132=item $kernel->set_image ($index, $value)
2133
2134=item $kernel->set_sampler ($index, $value)
2135
2136=item $kernel->set_local ($index, $value)
2137
2138=item $kernel->set_event ($index, $value)
2139
500This is a family of methods to set the kernel argument with the number C<$index> to the give C<$value>. 2140This is a family of methods to set the kernel argument with the number
501 2141C<$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 2142
506Chars and integers (including the half type) are specified as integers, 2143Chars and integers (including the half type) are specified as integers,
507float and double as floating point values, memory/buffer/image2d/image3d 2144float 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 2145an object of that type or C<undef>, local-memory arguments are set by
509objects of that type. 2146specifying the size, and sampler and event must be objects of that type.
2147
2148Note that C<set_memory> works for all memory objects (all types of buffers
2149and images) - the main purpose of the more specific C<set_TYPE> functions
2150is type checking.
2151
2152Setting an argument for a kernel does NOT keep a reference to the object -
2153for example, if you set an argument to some image object, free the image,
2154and call the kernel, you will run into undefined behaviour.
510 2155
511L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html> 2156L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html>
512 2157
513=back 2158=back
514 2159
517This is the superclass for all event objects (including OpenCL::UserEvent 2162This is the superclass for all event objects (including OpenCL::UserEvent
518objects). 2163objects).
519 2164
520=over 4 2165=over 4
521 2166
2167=item $ev->wait
2168
2169Waits for the event to complete.
2170
2171L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html>
2172
2173=item $ev->cb ($exec_callback_type, $callback->($event, $event_command_exec_status))
2174
2175Adds a callback to the callback stack for the given event type. There is
2176no way to remove a callback again.
2177
2178L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetEventCallback.html>
2179
522=item $packed_value = $ev->info ($name) 2180=item $packed_value = $ev->info ($name)
523 2181
524See C<< $platform->info >> for details. 2182See C<< $platform->info >> for details.
525 2183
526L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html> 2184L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html>
527 2185
528=item $ev->wait 2186=for gengetinfo begin event
529 2187
530Waits for the event to complete. 2188=item $queue = $event->command_queue
531 2189
2190Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_QUEUE> and returns the result.
2191
2192=item $command_type = $event->command_type
2193
2194Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_TYPE> and returns the result.
2195
2196=item $uint = $event->reference_count
2197
2198Calls C<clGetEventInfo> with C<OpenCL::EVENT_REFERENCE_COUNT> and returns the result.
2199
2200=item $uint = $event->command_execution_status
2201
2202Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_EXECUTION_STATUS> and returns the result.
2203
2204=item $ctx = $event->context
2205
2206Calls C<clGetEventInfo> with C<OpenCL::EVENT_CONTEXT> and returns the result.
2207
2208=for gengetinfo end event
2209
2210=item $packed_value = $ev->profiling_info ($name)
2211
2212See C<< $platform->info >> for details.
2213
2214The reason this method is not called C<info> is that there already is an
2215C<< ->info >> method.
2216
532L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> 2217L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProfilingInfo.html>
2218
2219=for gengetinfo begin profiling
2220
2221=item $ulong = $event->profiling_command_queued
2222
2223Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_QUEUED> and returns the result.
2224
2225=item $ulong = $event->profiling_command_submit
2226
2227Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_SUBMIT> and returns the result.
2228
2229=item $ulong = $event->profiling_command_start
2230
2231Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_START> and returns the result.
2232
2233=item $ulong = $event->profiling_command_end
2234
2235Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_END> and returns the result.
2236
2237=for gengetinfo end profiling
533 2238
534=back 2239=back
535 2240
536=head2 THE OpenCL::UserEvent CLASS 2241=head2 THE OpenCL::UserEvent CLASS
537 2242
539 2244
540=over 4 2245=over 4
541 2246
542=item $ev->set_status ($execution_status) 2247=item $ev->set_status ($execution_status)
543 2248
2249Sets the execution status of the user event. Can only be called once,
2250either with OpenCL::COMPLETE or a negative number as status.
2251
2252execution_status: OpenCL::COMPLETE or a negative integer.
2253
544L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html> 2254L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html>
545 2255
546=back 2256=back
547 2257
2258=head2 THE OpenCL::Mapped CLASS
2259
2260This class represents objects mapped into host memory. They are
2261represented by a blessed string scalar. The string data is the mapped
2262memory area, that is, if you read or write it, then the mapped object is
2263accessed directly.
2264
2265You must only ever use operations that modify the string in-place - for
2266example, a C<substr> that doesn't change the length, or maybe a regex that
2267doesn't change the length. Any other operation might cause the data to be
2268copied.
2269
2270When the object is destroyed it will enqueue an implicit unmap operation
2271on the queue that was used to create it.
2272
2273Keep in mind that you I<need> to unmap (or destroy) mapped objects before
2274OpenCL sees the changes, even if some implementations don't need this
2275sometimes.
2276
2277Example, replace the first two floats in the mapped buffer by 1 and 2.
2278
2279 my $mapped = $queue->map_buffer ($buf, ...
2280 $mapped->event->wait; # make sure it's there
2281
2282 # now replace first 8 bytes by new data, which is exactly 8 bytes long
2283 # we blindly assume device endianness to equal host endianness
2284 # (and of course, we assume iee 754 single precision floats :)
2285 substr $$mapped, 0, 8, pack "f*", 1, 2;
2286
2287=over 4
2288
2289=item $ev = $mapped->unmap ($wait_events...)
2290
2291Unmaps the mapped memory object, using the queue originally used to create
2292it, quite similarly to C<< $queue->unmap ($mapped, ...) >>.
2293
2294=item $bool = $mapped->mapped
2295
2296Returns whether the object is still mapped - true before an C<unmap> is
2297enqueued, false afterwards.
2298
2299=item $ev = $mapped->event
2300
2301Return the event object associated with the mapped object. Initially, this
2302will be the event object created when mapping the object, and after an
2303unmap, this will be the event object that the unmap operation created.
2304
2305=item $mapped->wait
2306
2307Same as C<< $mapped->event->wait >> - makes sure no operations on this
2308mapped object are outstanding.
2309
2310=item $bytes = $mapped->size
2311
2312Returns the size of the mapped area, in bytes. Same as C<length $$mapped>.
2313
2314=item $ptr = $mapped->ptr
2315
2316Returns the raw memory address of the mapped area.
2317
2318=item $mapped->set ($offset, $data)
2319
2320Replaces the data at the given C<$offset> in the memory area by the new
2321C<$data>. This method is safer than direct manipulation of C<$mapped>
2322because it does bounds-checking, but also slower.
2323
2324=item $data = $mapped->get ($offset, $length)
2325
2326Returns (without copying) a scalar representing the data at the given
2327C<$offset> and C<$length> in the mapped memory area. This is the same as
2328the following substr, except much slower;
2329
2330 $data = substr $$mapped, $offset, $length
2331
548=cut 2332=cut
549 2333
550package OpenCL; 2334sub OpenCL::Mapped::get {
551 2335 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} 2336}
2337
2338=back
2339
2340=head2 THE OpenCL::MappedBuffer CLASS
2341
2342This is a subclass of OpenCL::Mapped, representing mapped buffers.
2343
2344=head2 THE OpenCL::MappedImage CLASS
2345
2346This is a subclass of OpenCL::Mapped, representing mapped images.
2347
2348=over 4
2349
2350=item $pixels = $mapped->width
2351
2352=item $pixels = $mapped->height
2353
2354=item $pixels = $mapped->depth
2355
2356Return the width/height/depth of the mapped image region, in pixels.
2357
2358=item $bytes = $mapped->row_pitch
2359
2360=item $bytes = $mapped->slice_pitch
2361
2362Return the row or slice pitch of the image that has been mapped.
2363
2364=item $bytes = $mapped->element_size
2365
2366Return the size of a single pixel.
2367
2368=item $data = $mapped->get_row ($count, $x=0, $y=0, $z=0)
2369
2370Return C<$count> pixels from the given coordinates. The pixel data must
2371be completely contained within a single row.
2372
2373If C<$count> is C<undef>, then all the remaining pixels in that row are
2374returned.
2375
2376=item $mapped->set_row ($data, $x=0, $y=0, $z=0)
2377
2378Write the given pixel data at the given coordinate. The pixel data must
2379be completely contained within a single row.
2380
2381=back
2382
2383=cut
568 2384
5691; 23851;
570 2386
571=head1 AUTHOR 2387=head1 AUTHOR
572 2388

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines