1 |
=head1 NAME |
2 |
|
3 |
OpenCL - Open Computing Language Bindings |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use OpenCL; |
8 |
|
9 |
=head1 DESCRIPTION |
10 |
|
11 |
This is an early release which might be useful, but hasn't seen much testing. |
12 |
|
13 |
=head2 OpenCL FROM 10000 FEET HEIGHT |
14 |
|
15 |
Here is a high level overview of OpenCL: |
16 |
|
17 |
First you need to find one or more OpenCL::Platforms (kind of like |
18 |
vendors) - usually there is only one. |
19 |
|
20 |
Each platform gives you access to a number of OpenCL::Device objects, e.g. |
21 |
your graphics card. |
22 |
|
23 |
From a platform and some device(s), you create an OpenCL::Context, which is |
24 |
a very central object in OpenCL: Once you have a context you can create |
25 |
most other objects: |
26 |
|
27 |
OpenCL::Program objects, which store source code and, after building for a |
28 |
specific device ("compiling and linking"), also binary programs. For each |
29 |
kernel function in a program you can then create an OpenCL::Kernel object |
30 |
which represents basically a function call with argument values. |
31 |
|
32 |
OpenCL::Memory objects of various flavours: OpenCL::Buffers objects (flat |
33 |
memory areas, think arrays or structs) and OpenCL::Image objects (think 2d |
34 |
or 3d array) for bulk data and input and output for kernels. |
35 |
|
36 |
OpenCL::Sampler objects, which are kind of like texture filter modes in |
37 |
OpenGL. |
38 |
|
39 |
OpenCL::Queue objects - command queues, which allow you to submit memory |
40 |
reads, writes and copies, as well as kernel calls to your devices. They |
41 |
also offer a variety of methods to synchronise request execution, for |
42 |
example with barriers or OpenCL::Event objects. |
43 |
|
44 |
OpenCL::Event objects are used to signal when something is complete. |
45 |
|
46 |
=head2 HELPFUL RESOURCES |
47 |
|
48 |
The OpenCL spec used to develop this module (1.2 spec was available, but |
49 |
no implementation was available to me :). |
50 |
|
51 |
http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf |
52 |
|
53 |
OpenCL manpages: |
54 |
|
55 |
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/ |
56 |
|
57 |
Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it |
58 |
is, but at least it's free of charge: |
59 |
|
60 |
http://developer.amd.com/zones/OpenCLZone/courses/Documents/Introduction_to_OpenCL_Programming%20Training_Guide%20%28201005%29.pdf |
61 |
|
62 |
If you are into UML class diagrams, the following diagram might help - if |
63 |
not, it will be mildly cofusing: |
64 |
|
65 |
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/classDiagram.html |
66 |
|
67 |
=head1 BASIC WORKFLOW |
68 |
|
69 |
To get something done, you basically have to do this once (refer to the |
70 |
examples below for actual code, this is just a high-level description): |
71 |
|
72 |
Find some platform (e.g. the first one) and some device(s) (e.g. the first |
73 |
device of the platform), and create a context from those. |
74 |
|
75 |
Create program objects from your OpenCL source code, then build (compile) |
76 |
the programs for each device you want to run them on. |
77 |
|
78 |
Create kernel objects for all kernels you want to use (surprisingly, these |
79 |
are not device-specific). |
80 |
|
81 |
Then, to execute stuff, you repeat these steps, possibly resuing or |
82 |
sharing some buffers: |
83 |
|
84 |
Create some input and output buffers from your context. Set these as |
85 |
arguments to your kernel. |
86 |
|
87 |
Enqueue buffer writes to initialise your input buffers (when not |
88 |
initialised at creation time). |
89 |
|
90 |
Enqueue the kernel execution. |
91 |
|
92 |
Enqueue buffer reads for your output buffer to read results. |
93 |
|
94 |
=head1 EXAMPLES |
95 |
|
96 |
=head2 Enumerate all devices and get contexts for them. |
97 |
|
98 |
Best run this once to get a feel for the platforms and devices in your |
99 |
system. |
100 |
|
101 |
for my $platform (OpenCL::platforms) { |
102 |
printf "platform: %s\n", $platform->info (OpenCL::PLATFORM_NAME); |
103 |
printf "extensions: %s\n", $platform->info (OpenCL::PLATFORM_EXTENSIONS); |
104 |
for my $device ($platform->devices) { |
105 |
printf "+ device: %s\n", $device->info (OpenCL::DEVICE_NAME); |
106 |
my $ctx = $device->context; |
107 |
# do stuff |
108 |
} |
109 |
} |
110 |
|
111 |
=head2 Get a useful context and a command queue. |
112 |
|
113 |
This is a useful boilerplate for any OpenCL program that only wants to use |
114 |
one device, |
115 |
|
116 |
my ($platform) = OpenCL::platforms; # find first platform |
117 |
my ($dev) = $platform->devices; # find first device of platform |
118 |
my $ctx = $platform->context (undef, [$dev]); # create context out of those |
119 |
my $queue = $ctx->queue ($dev); # create a command queue for the device |
120 |
|
121 |
=head2 Print all supported image formats of a context. |
122 |
|
123 |
Best run this once for your context, to see whats available and how to |
124 |
gather information. |
125 |
|
126 |
for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) { |
127 |
print "supported image formats for ", OpenCL::enum2str $type, "\n"; |
128 |
|
129 |
for my $f ($ctx->supported_image_formats (0, $type)) { |
130 |
printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1]; |
131 |
} |
132 |
} |
133 |
|
134 |
=head2 Create a buffer with some predefined data, read it back synchronously, |
135 |
then asynchronously. |
136 |
|
137 |
my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut"); |
138 |
|
139 |
$queue->enqueue_read_buffer ($buf, 1, 1, 3, my $data); |
140 |
print "$data\n"; |
141 |
|
142 |
my $ev = $queue->enqueue_read_buffer ($buf, 0, 1, 3, my $data); |
143 |
$ev->wait; |
144 |
print "$data\n"; # prints "elm" |
145 |
|
146 |
=head2 Create and build a program, then create a kernel out of one of its |
147 |
functions. |
148 |
|
149 |
my $src = ' |
150 |
__kernel void |
151 |
squareit (__global float *input, __global float *output) |
152 |
{ |
153 |
$id = get_global_id (0); |
154 |
output [id] = input [id] * input [id]; |
155 |
} |
156 |
'; |
157 |
|
158 |
my $prog = $ctx->program_with_source ($src); |
159 |
|
160 |
# build croaks on compile errors, so catch it and print the compile errors |
161 |
eval { $prog->build ($dev); 1 } |
162 |
or die $prog->build_info ($dev, OpenCL::PROGRAM_BUILD_LOG); |
163 |
|
164 |
my $kernel = $prog->kernel ("squareit"); |
165 |
|
166 |
=head2 Create some input and output float buffers, then call the |
167 |
'squareit' kernel on them. |
168 |
|
169 |
my $input = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, pack "f*", 1, 2, 3, 4.5); |
170 |
my $output = $ctx->buffer (0, OpenCL::SIZEOF_FLOAT * 5); |
171 |
|
172 |
# set buffer |
173 |
$kernel->set_buffer (0, $input); |
174 |
$kernel->set_buffer (1, $output); |
175 |
|
176 |
# execute it for all 4 numbers |
177 |
$queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); |
178 |
|
179 |
# enqueue a synchronous read |
180 |
$queue->enqueue_read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); |
181 |
|
182 |
# print the results: |
183 |
printf "%s\n", join ", ", unpack "f*", $data; |
184 |
|
185 |
=head2 The same enqueue operations as before, but assuming an out-of-order queue, |
186 |
showing off barriers. |
187 |
|
188 |
# execute it for all 4 numbers |
189 |
$queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); |
190 |
|
191 |
# enqueue a barrier to ensure in-order execution |
192 |
$queue->enqueue_barrier; |
193 |
|
194 |
# enqueue an async read |
195 |
$queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); |
196 |
|
197 |
# wait for all requests to finish |
198 |
$queue->finish; |
199 |
|
200 |
=head2 The same enqueue operations as before, but assuming an out-of-order queue, |
201 |
showing off event objects and wait lists. |
202 |
|
203 |
# execute it for all 4 numbers |
204 |
my $ev = $queue->enqueue_nd_range_kernel ($kernel, undef, [4], undef); |
205 |
|
206 |
# enqueue an async read |
207 |
$ev = $queue->enqueue_read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev); |
208 |
|
209 |
# wait for the last event to complete |
210 |
$ev->wait; |
211 |
|
212 |
=head1 DOCUMENTATION |
213 |
|
214 |
=head2 BASIC CONVENTIONS |
215 |
|
216 |
This is not a one-to-one C-style translation of OpenCL to Perl - instead |
217 |
I attempted to make the interface as type-safe as possible by introducing |
218 |
object syntax where it makes sense. There are a number of important |
219 |
differences between the OpenCL C API and this module: |
220 |
|
221 |
=over 4 |
222 |
|
223 |
=item * Object lifetime managament is automatic - there is no need |
224 |
to free objects explicitly (C<clReleaseXXX>), the release function |
225 |
is called automatically once all Perl references to it go away. |
226 |
|
227 |
=item * OpenCL uses CamelCase for function names (C<clGetPlatformInfo>), |
228 |
while this module uses underscores as word separator and often leaves out |
229 |
prefixes (C<< $platform->info >>). |
230 |
|
231 |
=item * OpenCL often specifies fixed vector function arguments as short |
232 |
arrays (C<$origin[3]>), while this module explicitly expects the |
233 |
components as separate arguments- |
234 |
|
235 |
=item * Structures are often specified with their components, and returned |
236 |
as arrayrefs. |
237 |
|
238 |
=item * Where possible, one of the pitch values is calculated from the |
239 |
perl scalar length and need not be specified. |
240 |
|
241 |
=item * When enqueuing commands, the wait list is specified by adding |
242 |
extra arguments to the function - anywhere a C<$wait_events...> argument |
243 |
is documented this can be any number of event objects. |
244 |
|
245 |
=item * When enqueuing commands, if the enqueue method is called in void |
246 |
context, no event is created. In all other contexts an event is returned |
247 |
by the method. |
248 |
|
249 |
=item * This module expects all functions to return C<CL_SUCCESS>. If any |
250 |
other status is returned the function will throw an exception, so you |
251 |
don't normally have to to any error checking. |
252 |
|
253 |
=back |
254 |
|
255 |
=head2 PERL AND OPENCL TYPES |
256 |
|
257 |
This handy(?) table lists OpenCL types and their perl, PDL and pack/unpack |
258 |
format equivalents: |
259 |
|
260 |
OpenCL perl PDL pack/unpack |
261 |
char IV - c |
262 |
uchar IV byte C |
263 |
short IV short s |
264 |
ushort IV ushort S |
265 |
int IV long? l |
266 |
uint IV - L |
267 |
long IV longlong q |
268 |
ulong IV - Q |
269 |
float NV float f |
270 |
half IV ushort S |
271 |
double NV double d |
272 |
|
273 |
=head2 THE OpenCL PACKAGE |
274 |
|
275 |
=over 4 |
276 |
|
277 |
=item $int = OpenCL::errno |
278 |
|
279 |
The last error returned by a function - it's only valid after an error occured |
280 |
and before calling another OpenCL function. |
281 |
|
282 |
=item $str = OpenCL::err2str $errval |
283 |
|
284 |
Comverts an error value into a human readable string. |
285 |
|
286 |
=item $str = OpenCL::enum2str $enum |
287 |
|
288 |
Converts most enum values (inof parameter names, image format constants, |
289 |
object types, addressing and filter modes, command types etc.) into a |
290 |
human readbale string. When confronted with some random integer it can be |
291 |
very helpful to pass it through this function to maybe get some readable |
292 |
string out of it. |
293 |
|
294 |
=item @platforms = OpenCL::platforms |
295 |
|
296 |
Returns all available OpenCL::Platform objects. |
297 |
|
298 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html> |
299 |
|
300 |
=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $notify = undef |
301 |
|
302 |
Tries to create a context from a default device and platform - never worked for me. |
303 |
|
304 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> |
305 |
|
306 |
=item OpenCL::wait_for_events $wait_events... |
307 |
|
308 |
Waits for all events to complete. |
309 |
|
310 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> |
311 |
|
312 |
=back |
313 |
|
314 |
=head2 THE OpenCL::Platform CLASS |
315 |
|
316 |
=over 4 |
317 |
|
318 |
=item $packed_value = $platform->info ($name) |
319 |
|
320 |
Calls C<clGetPlatformInfo> and returns the packed, raw value - for |
321 |
strings, this will be the string, for other values you probably need to |
322 |
use the correct C<unpack>. This might get improved in the future. Hopefully. |
323 |
|
324 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html> |
325 |
|
326 |
=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL) |
327 |
|
328 |
Returns a list of matching OpenCL::Device objects. |
329 |
|
330 |
=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $notify = undef) |
331 |
|
332 |
Tries to create a context. Never worked for me, and you need devices explitly anyway. |
333 |
|
334 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> |
335 |
|
336 |
=item $ctx = $device->context ($properties = undef, @$devices, $notify = undef) |
337 |
|
338 |
Create a new OpenCL::Context object using the given device object(s)- a |
339 |
CL_CONTEXT_PLATFORM property is supplied automatically. |
340 |
|
341 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html> |
342 |
|
343 |
=back |
344 |
|
345 |
=head2 THE OpenCL::Device CLASS |
346 |
|
347 |
=over 4 |
348 |
|
349 |
=item $packed_value = $device->info ($name) |
350 |
|
351 |
See C<< $platform->info >> for details. |
352 |
|
353 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html> |
354 |
|
355 |
=back |
356 |
|
357 |
=head2 THE OpenCL::Context CLASS |
358 |
|
359 |
=over 4 |
360 |
|
361 |
=item $packed_value = $ctx->info ($name) |
362 |
|
363 |
See C<< $platform->info >> for details. |
364 |
|
365 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html> |
366 |
|
367 |
=item $queue = $ctx->queue ($device, $properties) |
368 |
|
369 |
Create a new OpenCL::Queue object from the context and the given device. |
370 |
|
371 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html> |
372 |
|
373 |
=item $ev = $ctx->user_event |
374 |
|
375 |
Creates a new OpenCL::UserEvent object. |
376 |
|
377 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html> |
378 |
|
379 |
=item $buf = $ctx->buffer ($flags, $len) |
380 |
|
381 |
Creates a new OpenCL::Buffer object with the given flags and octet-size. |
382 |
|
383 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html> |
384 |
|
385 |
=item $buf = $ctx->buffer_sv ($flags, $data) |
386 |
|
387 |
Creates a new OpenCL::Buffer object and initialise it with the given data values. |
388 |
|
389 |
=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $data) |
390 |
|
391 |
Creates a new OpenCL::Image2D object and optionally initialises it with the given data values. |
392 |
|
393 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html> |
394 |
|
395 |
=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $slice_pitch, $data) |
396 |
|
397 |
Creates a new OpenCL::Image3D object and optionally initialises it with the given data values. |
398 |
|
399 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html> |
400 |
|
401 |
=item @formats = $ctx->supported_image_formats ($flags, $image_type) |
402 |
|
403 |
Returns a list of matching image formats - each format is an arrayref with |
404 |
two values, $channel_order and $channel_type, in it. |
405 |
|
406 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSupportedImageFormats.html> |
407 |
|
408 |
=item $sampler = $ctx->sampler ($normalized_coords, $addressing_mode, $filter_mode) |
409 |
|
410 |
Creates a new OpenCL::Sampler object. |
411 |
|
412 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSampler.html> |
413 |
|
414 |
=item $program = $ctx->program_with_source ($string) |
415 |
|
416 |
Creates a new OpenCL::Program object from the given source code. |
417 |
|
418 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html> |
419 |
|
420 |
=back |
421 |
|
422 |
=head2 THE OpenCL::Queue CLASS |
423 |
|
424 |
An OpenCL::Queue represents an execution queue for OpenCL. You execute |
425 |
requests by calling their respective C<enqueue_xxx> method and waitinf for |
426 |
it to complete in some way. |
427 |
|
428 |
All the enqueue methods return an event object that can be used to wait |
429 |
for completion, unless the method is called in void context, in which case |
430 |
no event object is created. |
431 |
|
432 |
They also allow you to specify any number of other event objects that this |
433 |
request has to wait for before it starts executing, by simply passing the |
434 |
event objects as extra parameters to the enqueue methods. |
435 |
|
436 |
Queues execute in-order by default, without any parallelism, so in most |
437 |
cases (i.e. you use only one queue) it's not necessary to wait for or |
438 |
create event objects. |
439 |
|
440 |
=over 4 |
441 |
|
442 |
=item $packed_value = $ctx->info ($name) |
443 |
|
444 |
See C<< $platform->info >> for details. |
445 |
|
446 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetCommandQueueInfo.html> |
447 |
|
448 |
=item $ev = $queue->enqueue_read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...) |
449 |
|
450 |
Reads data from buffer into the given string. |
451 |
|
452 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html> |
453 |
|
454 |
=item $ev = $queue->enqueue_write_buffer ($buffer, $blocking, $offset, $data, $wait_events...) |
455 |
|
456 |
Writes data to buffer from the given string. |
457 |
|
458 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html> |
459 |
|
460 |
=item $ev = $queue->enqueue_copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...) |
461 |
|
462 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html> |
463 |
|
464 |
=item $ev = $queue->enqueue_read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) |
465 |
|
466 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html> |
467 |
|
468 |
=item $ev = $queue->enqueue_write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) |
469 |
|
470 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html> |
471 |
|
472 |
=item $ev = $queue->enqueue_copy_buffer_rect ($src, $dst, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $src_row_pitch, $src_slice_pitch, $dst_row_pitch, $dst_slice_pitch, $wait_event...) |
473 |
|
474 |
Yeah. |
475 |
|
476 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferRect.html> |
477 |
|
478 |
=item $ev = $queue->enqueue_copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...) |
479 |
|
480 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>. |
481 |
|
482 |
=item $ev = $queue->enqueue_copy_image ($src_image, $dst_image, $src_x, $src_y, $src_z, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...) |
483 |
|
484 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html> |
485 |
|
486 |
=item $ev = $queue->enqueue_copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...) |
487 |
|
488 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html> |
489 |
|
490 |
=item $ev = $queue->enqueue_task ($kernel, $wait_events...) |
491 |
|
492 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html> |
493 |
|
494 |
=item $ev = $queue->enqueue_nd_range_kernel ($kernel, @$global_work_offset, @$global_work_size, @$local_work_size, $wait_events...) |
495 |
|
496 |
Enqueues a kernel execution. |
497 |
|
498 |
@$global_work_size must be specified as a reference to an array of |
499 |
integers specifying the work sizes (element counts). |
500 |
|
501 |
@$global_work_offset must be either C<undef> (in which case all offsets |
502 |
are C<0>), or a reference to an array of work offsets, with the same number |
503 |
of elements as @$global_work_size. |
504 |
|
505 |
@$local_work_size must be either C<undef> (in which case the |
506 |
implementation is supposed to choose good local work sizes), or a |
507 |
reference to an array of local work sizes, with the same number of |
508 |
elements as @$global_work_size. |
509 |
|
510 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html> |
511 |
|
512 |
=item $ev = $queue->enqueue_marker |
513 |
|
514 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMarker.html> |
515 |
|
516 |
=item $ev = $queue->enqueue_wait_for_events ($wait_events...) |
517 |
|
518 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html> |
519 |
|
520 |
=item $queue->enqueue_barrier |
521 |
|
522 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueBarrier.html> |
523 |
|
524 |
=item $queue->flush |
525 |
|
526 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html> |
527 |
|
528 |
=item $queue->finish |
529 |
|
530 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html> |
531 |
|
532 |
=back |
533 |
|
534 |
=head2 THE OpenCL::Memory CLASS |
535 |
|
536 |
This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image, |
537 |
OpenCL::Image2D and OpenCL::Image3D. The subclasses of this class |
538 |
currently only exist to allow type-checking. |
539 |
|
540 |
=over 4 |
541 |
|
542 |
=item $packed_value = $memory->info ($name) |
543 |
|
544 |
See C<< $platform->info >> for details. |
545 |
|
546 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html> |
547 |
|
548 |
=back |
549 |
|
550 |
=head2 THE OpenCL::Sampler CLASS |
551 |
|
552 |
=over 4 |
553 |
|
554 |
=item $packed_value = $sampler->info ($name) |
555 |
|
556 |
See C<< $platform->info >> for details. |
557 |
|
558 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html> |
559 |
|
560 |
=back |
561 |
|
562 |
=head2 THE OpenCL::Program CLASS |
563 |
|
564 |
=over 4 |
565 |
|
566 |
=item $packed_value = $program->info ($name) |
567 |
|
568 |
See C<< $platform->info >> for details. |
569 |
|
570 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html> |
571 |
|
572 |
=item $program->build ($device, $options = "") |
573 |
|
574 |
Tries to build the program with the givne options. |
575 |
|
576 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html> |
577 |
|
578 |
=item $packed_value = $program->build_info ($device, $name) |
579 |
|
580 |
Similar to C<< $platform->info >>, but returns build info for a previous |
581 |
build attempt for the given device. |
582 |
|
583 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetBuildInfo.html> |
584 |
|
585 |
=item $kernel = $program->kernel ($function_name) |
586 |
|
587 |
Creates an OpenCL::Kernel object out of the named C<__kernel> function in |
588 |
the program. |
589 |
|
590 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html> |
591 |
|
592 |
=back |
593 |
|
594 |
=head2 THE OpenCL::Kernel CLASS |
595 |
|
596 |
=over 4 |
597 |
|
598 |
=item $packed_value = $kernel->info ($name) |
599 |
|
600 |
See C<< $platform->info >> for details. |
601 |
|
602 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html> |
603 |
|
604 |
=item $kernel->set_TYPE ($index, $value) |
605 |
|
606 |
This is a family of methods to set the kernel argument with the number C<$index> to the give C<$value>. |
607 |
|
608 |
TYPE is one of C<char>, C<uchar>, C<short>, C<ushort>, C<int>, C<uint>, |
609 |
C<long>, C<ulong>, C<half>, C<float>, C<double>, C<memory>, C<buffer>, |
610 |
C<image2d>, C<image3d>, C<sampler> or C<event>. |
611 |
|
612 |
Chars and integers (including the half type) are specified as integers, |
613 |
float and double as floating point values, memory/buffer/image2d/image3d |
614 |
must be an object of that type or C<undef>, and sampler and event must be |
615 |
objects of that type. |
616 |
|
617 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html> |
618 |
|
619 |
=back |
620 |
|
621 |
=head2 THE OpenCL::Event CLASS |
622 |
|
623 |
This is the superclass for all event objects (including OpenCL::UserEvent |
624 |
objects). |
625 |
|
626 |
=over 4 |
627 |
|
628 |
=item $packed_value = $ev->info ($name) |
629 |
|
630 |
See C<< $platform->info >> for details. |
631 |
|
632 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html> |
633 |
|
634 |
=item $ev->wait |
635 |
|
636 |
Waits for the event to complete. |
637 |
|
638 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> |
639 |
|
640 |
=back |
641 |
|
642 |
=head2 THE OpenCL::UserEvent CLASS |
643 |
|
644 |
This is a subclass of OpenCL::Event. |
645 |
|
646 |
=over 4 |
647 |
|
648 |
=item $ev->set_status ($execution_status) |
649 |
|
650 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html> |
651 |
|
652 |
=back |
653 |
|
654 |
=cut |
655 |
|
656 |
package OpenCL; |
657 |
|
658 |
use common::sense; |
659 |
|
660 |
BEGIN { |
661 |
our $VERSION = '0.14'; |
662 |
|
663 |
require XSLoader; |
664 |
XSLoader::load (__PACKAGE__, $VERSION); |
665 |
|
666 |
@OpenCL::Buffer::ISA = |
667 |
@OpenCL::Image::ISA = OpenCL::Memory::; |
668 |
|
669 |
@OpenCL::Image2D::ISA = |
670 |
@OpenCL::Image3D::ISA = OpenCL::Image::; |
671 |
|
672 |
@OpenCL::UserEvent::ISA = OpenCL::Event::; |
673 |
} |
674 |
|
675 |
1; |
676 |
|
677 |
=head1 AUTHOR |
678 |
|
679 |
Marc Lehmann <schmorp@schmorp.de> |
680 |
http://home.schmorp.de/ |
681 |
|
682 |
=cut |
683 |
|