1 |
root |
1.1 |
=head1 NAME |
2 |
|
|
|
3 |
root |
1.5 |
OpenCL - Open Computing Language Bindings |
4 |
root |
1.1 |
|
5 |
|
|
=head1 SYNOPSIS |
6 |
|
|
|
7 |
|
|
use OpenCL; |
8 |
|
|
|
9 |
|
|
=head1 DESCRIPTION |
10 |
|
|
|
11 |
root |
1.7 |
This is an early release which might be useful, but hasn't seen much testing. |
12 |
root |
1.1 |
|
13 |
root |
1.9 |
=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 |
root |
1.11 |
From a platform and some device(s), you create an OpenCL::Context, which is |
24 |
root |
1.9 |
a very central object in OpenCL: Once you have a context you can create |
25 |
|
|
most other objects: |
26 |
|
|
|
27 |
root |
1.11 |
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 |
root |
1.9 |
|
32 |
root |
1.20 |
OpenCL::Memory objects of various flavours: OpenCL::Buffer objects (flat |
33 |
root |
1.16 |
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 |
root |
1.9 |
|
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 |
root |
1.3 |
|
48 |
root |
1.71 |
The OpenCL specs used to develop this module - download these and keept |
49 |
|
|
hema round, they are required reference material: |
50 |
root |
1.3 |
|
51 |
|
|
http://www.khronos.org/registry/cl/specs/opencl-1.1.pdf |
52 |
root |
1.57 |
http://www.khronos.org/registry/cl/specs/opencl-1.2.pdf |
53 |
|
|
http://www.khronos.org/registry/cl/specs/opencl-1.2-extensions.pdf |
54 |
root |
1.3 |
|
55 |
|
|
OpenCL manpages: |
56 |
|
|
|
57 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/ |
58 |
root |
1.57 |
http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/ |
59 |
root |
1.3 |
|
60 |
root |
1.18 |
If you are into UML class diagrams, the following diagram might help - if |
61 |
root |
1.57 |
not, it will be mildly confusing (also, the class hierarchy of this module |
62 |
|
|
is much more fine-grained): |
63 |
root |
1.18 |
|
64 |
root |
1.57 |
http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/classDiagram.html |
65 |
root |
1.18 |
|
66 |
root |
1.16 |
Here's a tutorial from AMD (very AMD-centric, too), not sure how useful it |
67 |
|
|
is, 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 |
|
|
|
71 |
root |
1.18 |
And here's NVIDIA's OpenCL Best Practises Guide: |
72 |
root |
1.16 |
|
73 |
root |
1.18 |
http://developer.download.nvidia.com/compute/cuda/3_2/toolkit/docs/OpenCL_Best_Practices_Guide.pdf |
74 |
root |
1.16 |
|
75 |
root |
1.9 |
=head1 BASIC WORKFLOW |
76 |
|
|
|
77 |
root |
1.11 |
To get something done, you basically have to do this once (refer to the |
78 |
|
|
examples below for actual code, this is just a high-level description): |
79 |
root |
1.9 |
|
80 |
root |
1.11 |
Find some platform (e.g. the first one) and some device(s) (e.g. the first |
81 |
|
|
device of the platform), and create a context from those. |
82 |
root |
1.9 |
|
83 |
root |
1.11 |
Create program objects from your OpenCL source code, then build (compile) |
84 |
|
|
the programs for each device you want to run them on. |
85 |
root |
1.9 |
|
86 |
root |
1.11 |
Create kernel objects for all kernels you want to use (surprisingly, these |
87 |
|
|
are not device-specific). |
88 |
root |
1.9 |
|
89 |
root |
1.11 |
Then, to execute stuff, you repeat these steps, possibly resuing or |
90 |
|
|
sharing some buffers: |
91 |
root |
1.9 |
|
92 |
root |
1.11 |
Create some input and output buffers from your context. Set these as |
93 |
|
|
arguments to your kernel. |
94 |
|
|
|
95 |
|
|
Enqueue buffer writes to initialise your input buffers (when not |
96 |
|
|
initialised at creation time). |
97 |
root |
1.9 |
|
98 |
|
|
Enqueue the kernel execution. |
99 |
|
|
|
100 |
|
|
Enqueue buffer reads for your output buffer to read results. |
101 |
|
|
|
102 |
root |
1.3 |
=head1 EXAMPLES |
103 |
|
|
|
104 |
root |
1.5 |
=head2 Enumerate all devices and get contexts for them. |
105 |
root |
1.1 |
|
106 |
root |
1.11 |
Best run this once to get a feel for the platforms and devices in your |
107 |
|
|
system. |
108 |
|
|
|
109 |
root |
1.1 |
for my $platform (OpenCL::platforms) { |
110 |
root |
1.24 |
printf "platform: %s\n", $platform->name; |
111 |
|
|
printf "extensions: %s\n", $platform->extensions; |
112 |
root |
1.1 |
for my $device ($platform->devices) { |
113 |
root |
1.24 |
printf "+ device: %s\n", $device->name; |
114 |
root |
1.29 |
my $ctx = $platform->context (undef, [$device]); |
115 |
root |
1.1 |
# do stuff |
116 |
|
|
} |
117 |
|
|
} |
118 |
|
|
|
119 |
root |
1.5 |
=head2 Get a useful context and a command queue. |
120 |
root |
1.1 |
|
121 |
root |
1.11 |
This is a useful boilerplate for any OpenCL program that only wants to use |
122 |
|
|
one device, |
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 |
128 |
root |
1.1 |
|
129 |
root |
1.5 |
=head2 Print all supported image formats of a context. |
130 |
|
|
|
131 |
root |
1.11 |
Best run this once for your context, to see whats available and how to |
132 |
|
|
gather information. |
133 |
|
|
|
134 |
root |
1.5 |
for my $type (OpenCL::MEM_OBJECT_IMAGE2D, OpenCL::MEM_OBJECT_IMAGE3D) { |
135 |
root |
1.10 |
print "supported image formats for ", OpenCL::enum2str $type, "\n"; |
136 |
root |
1.5 |
|
137 |
|
|
for my $f ($ctx->supported_image_formats (0, $type)) { |
138 |
|
|
printf " %-10s %-20s\n", OpenCL::enum2str $f->[0], OpenCL::enum2str $f->[1]; |
139 |
|
|
} |
140 |
|
|
} |
141 |
|
|
|
142 |
|
|
=head2 Create a buffer with some predefined data, read it back synchronously, |
143 |
|
|
then asynchronously. |
144 |
root |
1.3 |
|
145 |
|
|
my $buf = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, "helmut"); |
146 |
|
|
|
147 |
root |
1.59 |
$queue->read_buffer ($buf, 1, 1, 3, my $data); |
148 |
root |
1.10 |
print "$data\n"; |
149 |
root |
1.3 |
|
150 |
root |
1.59 |
my $ev = $queue->read_buffer ($buf, 0, 1, 3, my $data); |
151 |
root |
1.3 |
$ev->wait; |
152 |
root |
1.10 |
print "$data\n"; # prints "elm" |
153 |
root |
1.3 |
|
154 |
root |
1.5 |
=head2 Create and build a program, then create a kernel out of one of its |
155 |
|
|
functions. |
156 |
root |
1.3 |
|
157 |
|
|
my $src = ' |
158 |
root |
1.31 |
kernel void |
159 |
|
|
squareit (global float *input, global float *output) |
160 |
root |
1.3 |
{ |
161 |
root |
1.15 |
$id = get_global_id (0); |
162 |
root |
1.3 |
output [id] = input [id] * input [id]; |
163 |
|
|
} |
164 |
|
|
'; |
165 |
|
|
|
166 |
root |
1.51 |
my $prog = $ctx->build_program ($src); |
167 |
root |
1.3 |
my $kernel = $prog->kernel ("squareit"); |
168 |
|
|
|
169 |
root |
1.11 |
=head2 Create some input and output float buffers, then call the |
170 |
|
|
'squareit' kernel on them. |
171 |
root |
1.4 |
|
172 |
|
|
my $input = $ctx->buffer_sv (OpenCL::MEM_COPY_HOST_PTR, pack "f*", 1, 2, 3, 4.5); |
173 |
|
|
my $output = $ctx->buffer (0, OpenCL::SIZEOF_FLOAT * 5); |
174 |
|
|
|
175 |
|
|
# set buffer |
176 |
|
|
$kernel->set_buffer (0, $input); |
177 |
|
|
$kernel->set_buffer (1, $output); |
178 |
|
|
|
179 |
|
|
# execute it for all 4 numbers |
180 |
root |
1.59 |
$queue->nd_range_kernel ($kernel, undef, [4], undef); |
181 |
root |
1.4 |
|
182 |
root |
1.5 |
# enqueue a synchronous read |
183 |
root |
1.59 |
$queue->read_buffer ($output, 1, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); |
184 |
root |
1.5 |
|
185 |
|
|
# print the results: |
186 |
root |
1.10 |
printf "%s\n", join ", ", unpack "f*", $data; |
187 |
root |
1.5 |
|
188 |
|
|
=head2 The same enqueue operations as before, but assuming an out-of-order queue, |
189 |
|
|
showing off barriers. |
190 |
|
|
|
191 |
|
|
# execute it for all 4 numbers |
192 |
root |
1.59 |
$queue->nd_range_kernel ($kernel, undef, [4], undef); |
193 |
root |
1.5 |
|
194 |
|
|
# enqueue a barrier to ensure in-order execution |
195 |
root |
1.59 |
$queue->barrier; |
196 |
root |
1.4 |
|
197 |
root |
1.5 |
# enqueue an async read |
198 |
root |
1.59 |
$queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data); |
199 |
root |
1.5 |
|
200 |
|
|
# wait for all requests to finish |
201 |
|
|
$queue->finish; |
202 |
|
|
|
203 |
|
|
=head2 The same enqueue operations as before, but assuming an out-of-order queue, |
204 |
|
|
showing off event objects and wait lists. |
205 |
|
|
|
206 |
|
|
# execute it for all 4 numbers |
207 |
root |
1.59 |
my $ev = $queue->nd_range_kernel ($kernel, undef, [4], undef); |
208 |
root |
1.5 |
|
209 |
|
|
# enqueue an async read |
210 |
root |
1.59 |
$ev = $queue->read_buffer ($output, 0, 0, OpenCL::SIZEOF_FLOAT * 4, my $data, $ev); |
211 |
root |
1.5 |
|
212 |
|
|
# wait for the last event to complete |
213 |
root |
1.4 |
$ev->wait; |
214 |
|
|
|
215 |
root |
1.38 |
=head2 Use the OpenGL module to share a texture between OpenCL and OpenGL and draw some julia |
216 |
root |
1.71 |
set flight effect. |
217 |
root |
1.38 |
|
218 |
root |
1.64 |
This is quite a long example to get you going - you can download it from |
219 |
|
|
L<http://cvs.schmorp.de/OpenCL/examples/juliaflight>. |
220 |
root |
1.38 |
|
221 |
|
|
use OpenGL ":all"; |
222 |
|
|
use OpenCL; |
223 |
|
|
|
224 |
root |
1.64 |
my $S = $ARGV[0] || 256; # window/texture size, smaller is faster |
225 |
|
|
|
226 |
root |
1.38 |
# open a window and create a gl texture |
227 |
root |
1.64 |
OpenGL::glpOpenWindow width => $S, height => $S; |
228 |
root |
1.38 |
my $texid = glGenTextures_p 1; |
229 |
|
|
glBindTexture GL_TEXTURE_2D, $texid; |
230 |
root |
1.64 |
glTexImage2D_c GL_TEXTURE_2D, 0, GL_RGBA8, $S, $S, 0, GL_RGBA, GL_UNSIGNED_BYTE, 0; |
231 |
root |
1.38 |
|
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 |
root |
1.64 |
int2 xy = (int2)(get_global_id (0), get_global_id (1)); |
260 |
|
|
float2 p = convert_float2 (xy) / $S.f * 2.f - 1.f; |
261 |
root |
1.38 |
|
262 |
root |
1.64 |
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 |
root |
1.38 |
|
265 |
|
|
float2 z = m; |
266 |
root |
1.64 |
float2 c = (float2)(sin (time * 0.01133f), cos (time * 0.02521f)); |
267 |
root |
1.38 |
|
268 |
root |
1.64 |
for (int i = 0; i < 25 && dot (z, z) < 4.f; ++i) // standard julia |
269 |
root |
1.38 |
z = (float2)(z.x * z.x - z.y * z.y, 2.f * z.x * z.y) + c; |
270 |
|
|
|
271 |
root |
1.64 |
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 |
root |
1.38 |
} |
274 |
|
|
EOF |
275 |
|
|
|
276 |
root |
1.51 |
my $prog = $ctx->build_program ($src); |
277 |
root |
1.38 |
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 |
root |
1.59 |
$queue->acquire_gl_objects ([$tex]); |
284 |
root |
1.38 |
|
285 |
|
|
# configure and run our kernel |
286 |
root |
1.64 |
$kernel->setf ("mf", $tex, $time*2); # mf = memory object, float |
287 |
|
|
$queue->nd_range_kernel ($kernel, undef, [$S, $S], undef); |
288 |
root |
1.38 |
|
289 |
|
|
# release objects to opengl again |
290 |
root |
1.59 |
$queue->release_gl_objects ([$tex]); |
291 |
root |
1.38 |
|
292 |
|
|
# wait |
293 |
root |
1.40 |
$queue->finish; |
294 |
root |
1.38 |
|
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 |
root |
1.66 |
=head2 How to modify the previous example to not rely on GL sharing. |
312 |
root |
1.65 |
|
313 |
|
|
For those poor souls with only a sucky CPU OpenCL implementation, you |
314 |
|
|
currently have to read the image into some perl scalar, and then modify a |
315 |
|
|
texture or use glDrawPixels or so). |
316 |
|
|
|
317 |
|
|
First, when you don't need gl sharing, you can create the context much simpler: |
318 |
|
|
|
319 |
|
|
$ctx = $platform->context (undef, [$dev]) |
320 |
|
|
|
321 |
|
|
To use a texture, you would modify the above example by creating an |
322 |
|
|
OpenCL::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 |
|
|
|
326 |
root |
1.71 |
And in the draw loop, intead of acquire_gl_objects/release_gl_objects, you |
327 |
root |
1.65 |
would 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 |
|
|
|
331 |
|
|
And 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 |
|
|
|
335 |
|
|
The fully modified example can be found at |
336 |
|
|
L<http://cvs.schmorp.de/OpenCL/examples/juliaflight-nosharing>. |
337 |
|
|
|
338 |
root |
1.71 |
=head2 Julia sets look soooo 80ies. |
339 |
|
|
|
340 |
|
|
Then colour them differently, e.g. using orbit traps! Replace the loop and |
341 |
|
|
colour 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 |
|
|
|
353 |
|
|
Also try C<-10.f> instead of C<-1.f>. |
354 |
|
|
|
355 |
root |
1.5 |
=head1 DOCUMENTATION |
356 |
|
|
|
357 |
|
|
=head2 BASIC CONVENTIONS |
358 |
|
|
|
359 |
root |
1.14 |
This is not a one-to-one C-style translation of OpenCL to Perl - instead |
360 |
|
|
I attempted to make the interface as type-safe as possible by introducing |
361 |
root |
1.5 |
object syntax where it makes sense. There are a number of important |
362 |
|
|
differences between the OpenCL C API and this module: |
363 |
|
|
|
364 |
|
|
=over 4 |
365 |
|
|
|
366 |
|
|
=item * Object lifetime managament is automatic - there is no need |
367 |
|
|
to free objects explicitly (C<clReleaseXXX>), the release function |
368 |
|
|
is called automatically once all Perl references to it go away. |
369 |
|
|
|
370 |
root |
1.20 |
=item * OpenCL uses CamelCase for function names |
371 |
|
|
(e.g. C<clGetPlatformIDs>, C<clGetPlatformInfo>), while this module |
372 |
|
|
uses underscores as word separator and often leaves out prefixes |
373 |
|
|
(C<OpenCL::platforms>, C<< $platform->info >>). |
374 |
root |
1.5 |
|
375 |
|
|
=item * OpenCL often specifies fixed vector function arguments as short |
376 |
root |
1.19 |
arrays (C<size_t origin[3]>), while this module explicitly expects the |
377 |
|
|
components as separate arguments (C<$orig_x, $orig_y, $orig_z>) in |
378 |
|
|
function calls. |
379 |
root |
1.5 |
|
380 |
root |
1.19 |
=item * Structures are often specified by flattening out their components |
381 |
|
|
as with short vectors, and returned as arrayrefs. |
382 |
root |
1.5 |
|
383 |
|
|
=item * When enqueuing commands, the wait list is specified by adding |
384 |
root |
1.9 |
extra arguments to the function - anywhere a C<$wait_events...> argument |
385 |
root |
1.44 |
is documented this can be any number of event objects. As an extsnion |
386 |
|
|
implemented by this module, C<undef> values will be ignored in the event |
387 |
|
|
list. |
388 |
root |
1.5 |
|
389 |
|
|
=item * When enqueuing commands, if the enqueue method is called in void |
390 |
|
|
context, no event is created. In all other contexts an event is returned |
391 |
|
|
by the method. |
392 |
|
|
|
393 |
|
|
=item * This module expects all functions to return C<CL_SUCCESS>. If any |
394 |
|
|
other status is returned the function will throw an exception, so you |
395 |
|
|
don't normally have to to any error checking. |
396 |
|
|
|
397 |
|
|
=back |
398 |
|
|
|
399 |
root |
1.7 |
=head2 PERL AND OPENCL TYPES |
400 |
|
|
|
401 |
root |
1.8 |
This handy(?) table lists OpenCL types and their perl, PDL and pack/unpack |
402 |
root |
1.7 |
format equivalents: |
403 |
|
|
|
404 |
root |
1.8 |
OpenCL perl PDL pack/unpack |
405 |
|
|
char IV - c |
406 |
|
|
uchar IV byte C |
407 |
|
|
short IV short s |
408 |
|
|
ushort IV ushort S |
409 |
|
|
int IV long? l |
410 |
|
|
uint IV - L |
411 |
|
|
long IV longlong q |
412 |
|
|
ulong IV - Q |
413 |
|
|
float NV float f |
414 |
|
|
half IV ushort S |
415 |
|
|
double NV double d |
416 |
root |
1.7 |
|
417 |
root |
1.36 |
=head2 GLX SUPPORT |
418 |
|
|
|
419 |
|
|
Due to the sad state that OpenGL support is in in Perl (mostly the OpenGL |
420 |
|
|
module, which has little to no documentation and has little to no support |
421 |
root |
1.38 |
for glX), this module, as a special extension, treats context creation |
422 |
root |
1.36 |
properties C<OpenCL::GLX_DISPLAY_KHR> and C<OpenCL::GL_CONTEXT_KHR> |
423 |
|
|
specially: If either or both of these are C<undef>, then the OpenCL |
424 |
root |
1.38 |
module tries to dynamically resolve C<glXGetCurrentDisplay> and |
425 |
|
|
C<glXGetCurrentContext>, call these functions and use their return values |
426 |
root |
1.36 |
instead. |
427 |
|
|
|
428 |
|
|
For this to work, the OpenGL library must be loaded, a GLX context must |
429 |
|
|
have been created and be made current, and C<dlsym> must be available and |
430 |
|
|
capable of finding the function via C<RTLD_DEFAULT>. |
431 |
|
|
|
432 |
root |
1.55 |
=head2 EVENT SYSTEM |
433 |
|
|
|
434 |
|
|
OpenCL can generate a number of (potentially) asynchronous events, for |
435 |
|
|
example, after compiling a program, to signal a context-related error or, |
436 |
|
|
perhaps most important, to signal completion of queued jobs (by setting |
437 |
|
|
callbacks on OpenCL::Event objects). |
438 |
|
|
|
439 |
|
|
To facilitate this, this module maintains an event queue - each |
440 |
|
|
time an asynchronous event happens, it is queued, and perl will be |
441 |
|
|
interrupted. This is implemented via the L<Async::Interrupt> module. In |
442 |
|
|
addition, this module has L<AnyEvent> support, so it can seamlessly |
443 |
|
|
integrate itself into many event loops. |
444 |
|
|
|
445 |
|
|
Since this module is a bit hard to understand, here are some case examples: |
446 |
|
|
|
447 |
|
|
=head3 Don't use callbacks. |
448 |
|
|
|
449 |
|
|
When your program never uses any callbacks, then there will never be any |
450 |
|
|
notifications you need to take care of, and therefore no need to worry |
451 |
|
|
about all this. |
452 |
|
|
|
453 |
|
|
You can achieve a great deal by explicitly waiting for events, or using |
454 |
|
|
barriers and flush calls. In many programs, there is no need at all to |
455 |
|
|
tinker with asynchronous events. |
456 |
|
|
|
457 |
|
|
=head3 Use AnyEvent |
458 |
|
|
|
459 |
|
|
This module automatically registers a watcher that invokes all outstanding |
460 |
|
|
event callbacks when AnyEvent is initialised (and block asynchronous |
461 |
|
|
interruptions). Using this mode of operations is the safest and most |
462 |
|
|
recommended one. |
463 |
|
|
|
464 |
|
|
To use this, simply use AnyEvent and this module normally, make sure you |
465 |
|
|
have an event loop running: |
466 |
|
|
|
467 |
|
|
use Gtk2 -init; |
468 |
|
|
use AnyEvent; |
469 |
|
|
|
470 |
|
|
# initialise AnyEvent, by creating a watcher, or: |
471 |
|
|
AnyEvent::detect; |
472 |
|
|
|
473 |
root |
1.59 |
my $e = $queue->marker; |
474 |
root |
1.55 |
$e->cb (sub { |
475 |
|
|
warn "opencl is finished\n"; |
476 |
|
|
}) |
477 |
|
|
|
478 |
|
|
main Gtk2; |
479 |
|
|
|
480 |
|
|
Note that this module will not initialise AnyEvent for you. Before |
481 |
|
|
AnyEvent is initialised, the module will asynchronously interrupt perl |
482 |
|
|
instead. To avoid any surprises, it's best to explicitly initialise |
483 |
|
|
AnyEvent. |
484 |
|
|
|
485 |
|
|
You can temporarily enable asynchronous interruptions (see next paragraph) |
486 |
|
|
by calling C<$OpenCL::INTERRUPT->unblock> and disable them again by |
487 |
|
|
calling C<$OpenCL::INTERRUPT->block>. |
488 |
|
|
|
489 |
|
|
=head3 Let yourself be interrupted at any time |
490 |
|
|
|
491 |
|
|
This mode is the default unless AnyEvent is loaded and initialised. In |
492 |
|
|
this mode, OpenCL asynchronously interrupts a running perl program. The |
493 |
|
|
emphasis is on both I<asynchronously> and I<running> here. |
494 |
|
|
|
495 |
|
|
Asynchronously means that perl might execute your callbacks at any |
496 |
|
|
time. For example, in the following code (I<THAT YOU SHOULD NOT COPY>), |
497 |
|
|
the C<until> loop following the marker call will be interrupted by the |
498 |
|
|
callback: |
499 |
|
|
|
500 |
root |
1.59 |
my $e = $queue->marker; |
501 |
root |
1.55 |
my $flag; |
502 |
|
|
$e->cb (sub { $flag = 1 }); |
503 |
|
|
1 until $flag; |
504 |
|
|
# $flag is now 1 |
505 |
|
|
|
506 |
|
|
The reason why you shouldn't blindly copy the above code is that |
507 |
|
|
busy waiting is a really really bad thing, and really really bad for |
508 |
|
|
performance. |
509 |
|
|
|
510 |
|
|
While at first this asynchronous business might look exciting, it can be |
511 |
|
|
really hard, because you need to be prepared for the callback code to be |
512 |
|
|
executed at any time, which limits the amount of things the callback code |
513 |
|
|
can do safely. |
514 |
|
|
|
515 |
|
|
This can be mitigated somewhat by using C<< |
516 |
|
|
$OpenCL::INTERRUPT->scope_block >> (see the L<Async::Interrupt> |
517 |
|
|
documentation for details). |
518 |
|
|
|
519 |
|
|
The other problem is that your program must be actively I<running> to be |
520 |
|
|
interrupted. When you calculate stuff, your program is running. When you |
521 |
|
|
hang in some C functions or other block execution (by calling C<sleep>, |
522 |
|
|
C<select>, running an event loop and so on), your program is waiting, not |
523 |
|
|
running. |
524 |
|
|
|
525 |
|
|
One way around that would be to attach a read watcher to your event loop, |
526 |
|
|
listening for events on C<< $OpenCL::INTERRUPT->pipe_fileno >>, using a |
527 |
|
|
dummy callback (C<sub { }>) to temporarily execute some perl code. |
528 |
|
|
|
529 |
|
|
That is then awfully close to using the built-in AnyEvent support above, |
530 |
|
|
though, so consider that one instead. |
531 |
|
|
|
532 |
|
|
=head3 Be creative |
533 |
|
|
|
534 |
|
|
OpenCL exports the L<Async::Interrupt> object it uses in the global |
535 |
|
|
variable C<$OpenCL::INTERRUPT>. You can configure it in any way you like. |
536 |
|
|
|
537 |
|
|
So if you want to feel like a real pro, err, wait, if you feel no risk |
538 |
|
|
menas no fun, you can experiment by implementing your own mode of |
539 |
|
|
operations. |
540 |
|
|
|
541 |
root |
1.52 |
=cut |
542 |
|
|
|
543 |
|
|
package OpenCL; |
544 |
|
|
|
545 |
|
|
use common::sense; |
546 |
root |
1.62 |
use Carp (); |
547 |
root |
1.55 |
use Async::Interrupt (); |
548 |
|
|
|
549 |
|
|
our $POLL_FUNC; # set by XS |
550 |
root |
1.52 |
|
551 |
|
|
BEGIN { |
552 |
root |
1.70 |
our $VERSION = '0.99'; |
553 |
root |
1.52 |
|
554 |
|
|
require XSLoader; |
555 |
|
|
XSLoader::load (__PACKAGE__, $VERSION); |
556 |
|
|
|
557 |
|
|
@OpenCL::Platform::ISA = |
558 |
|
|
@OpenCL::Device::ISA = |
559 |
|
|
@OpenCL::Context::ISA = |
560 |
|
|
@OpenCL::Queue::ISA = |
561 |
|
|
@OpenCL::Memory::ISA = |
562 |
|
|
@OpenCL::Sampler::ISA = |
563 |
|
|
@OpenCL::Program::ISA = |
564 |
|
|
@OpenCL::Kernel::ISA = |
565 |
|
|
@OpenCL::Event::ISA = OpenCL::Object::; |
566 |
|
|
|
567 |
root |
1.71 |
@OpenCL::SubDevice::ISA = OpenCL::Device::; |
568 |
|
|
|
569 |
root |
1.52 |
@OpenCL::Buffer::ISA = |
570 |
|
|
@OpenCL::Image::ISA = OpenCL::Memory::; |
571 |
|
|
|
572 |
|
|
@OpenCL::BufferObj::ISA = OpenCL::Buffer::; |
573 |
|
|
|
574 |
|
|
@OpenCL::Image2D::ISA = |
575 |
|
|
@OpenCL::Image3D::ISA = |
576 |
|
|
@OpenCL::Image2DArray::ISA = |
577 |
|
|
@OpenCL::Image1D::ISA = |
578 |
|
|
@OpenCL::Image1DArray::ISA = |
579 |
|
|
@OpenCL::Image1DBuffer::ISA = OpenCL::Image::; |
580 |
|
|
|
581 |
|
|
@OpenCL::UserEvent::ISA = OpenCL::Event::; |
582 |
root |
1.66 |
|
583 |
root |
1.67 |
@OpenCL::MappedBuffer::ISA = |
584 |
|
|
@OpenCL::MappedImage::ISA = OpenCL::Mapped::; |
585 |
root |
1.52 |
} |
586 |
|
|
|
587 |
root |
1.5 |
=head2 THE OpenCL PACKAGE |
588 |
|
|
|
589 |
|
|
=over 4 |
590 |
|
|
|
591 |
|
|
=item $int = OpenCL::errno |
592 |
|
|
|
593 |
root |
1.11 |
The last error returned by a function - it's only valid after an error occured |
594 |
|
|
and before calling another OpenCL function. |
595 |
root |
1.5 |
|
596 |
root |
1.61 |
=item $str = OpenCL::err2str [$errval] |
597 |
root |
1.5 |
|
598 |
root |
1.61 |
Converts an error value into a human readable string. IF no error value is |
599 |
|
|
given, then the last error will be used (as returned by OpenCL::errno). |
600 |
root |
1.5 |
|
601 |
root |
1.9 |
=item $str = OpenCL::enum2str $enum |
602 |
root |
1.5 |
|
603 |
root |
1.30 |
Converts most enum values (of parameter names, image format constants, |
604 |
root |
1.5 |
object types, addressing and filter modes, command types etc.) into a |
605 |
root |
1.30 |
human readable string. When confronted with some random integer it can be |
606 |
root |
1.5 |
very helpful to pass it through this function to maybe get some readable |
607 |
|
|
string out of it. |
608 |
|
|
|
609 |
|
|
=item @platforms = OpenCL::platforms |
610 |
|
|
|
611 |
|
|
Returns all available OpenCL::Platform objects. |
612 |
|
|
|
613 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformIDs.html> |
614 |
|
|
|
615 |
root |
1.56 |
=item $ctx = OpenCL::context_from_type $properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr |
616 |
root |
1.5 |
|
617 |
root |
1.56 |
Tries to create a context from a default device and platform type - never worked for me. |
618 |
root |
1.5 |
|
619 |
root |
1.71 |
type: OpenCL::DEVICE_TYPE_DEFAULT, OpenCL::DEVICE_TYPE_CPU, OpenCL::DEVICE_TYPE_GPU, |
620 |
|
|
OpenCL::DEVICE_TYPE_ACCELERATOR, OpenCL::DEVICE_TYPE_CUSTOM, OpenCL::DEVICE_TYPE_ALL. |
621 |
|
|
|
622 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> |
623 |
|
|
|
624 |
root |
1.56 |
=item $ctx = OpenCL::context $properties, \@devices, $callback->($err, $pvt) = $print_stderr) |
625 |
|
|
|
626 |
|
|
Create a new OpenCL::Context object using the given device object(s). This |
627 |
|
|
function isn't implemented yet, use C<< $platform->context >> instead. |
628 |
|
|
|
629 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html> |
630 |
|
|
|
631 |
root |
1.5 |
=item OpenCL::wait_for_events $wait_events... |
632 |
|
|
|
633 |
|
|
Waits for all events to complete. |
634 |
|
|
|
635 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> |
636 |
|
|
|
637 |
root |
1.55 |
=item OpenCL::poll |
638 |
|
|
|
639 |
|
|
Checks if there are any outstanding events (see L<EVENT SYSTEM>) and |
640 |
|
|
invokes their callbacks. |
641 |
|
|
|
642 |
|
|
=item $OpenCL::INTERRUPT |
643 |
|
|
|
644 |
|
|
The L<Async::Interrupt> object used to signal asynchronous events (see |
645 |
|
|
L<EVENT SYSTEM>). |
646 |
|
|
|
647 |
|
|
=cut |
648 |
|
|
|
649 |
|
|
our $INTERRUPT = new Async::Interrupt c_cb => [$POLL_FUNC, 0]; |
650 |
|
|
|
651 |
|
|
&_eq_initialise ($INTERRUPT->signal_func); |
652 |
|
|
|
653 |
|
|
=item $OpenCL::WATCHER |
654 |
|
|
|
655 |
|
|
The L<AnyEvent> watcher object used to watch for asynchronous events (see |
656 |
|
|
L<EVENT SYSTEM>). This variable is C<undef> until L<AnyEvent> has been |
657 |
|
|
loaded I<and> initialised (e.g. by calling C<AnyEvent::detect>). |
658 |
|
|
|
659 |
|
|
=cut |
660 |
|
|
|
661 |
|
|
our $WATCHER; |
662 |
|
|
|
663 |
|
|
sub _init_anyevent { |
664 |
|
|
$INTERRUPT->block; |
665 |
|
|
$WATCHER = AE::io ($INTERRUPT->pipe_fileno, 0, sub { $INTERRUPT->handle }); |
666 |
|
|
} |
667 |
|
|
|
668 |
|
|
if (defined $AnyEvent::MODEL) { |
669 |
|
|
_init_anyevent; |
670 |
|
|
} else { |
671 |
|
|
push @AnyEvent::post_detect, \&_init_anyevent; |
672 |
|
|
} |
673 |
|
|
|
674 |
root |
1.5 |
=back |
675 |
|
|
|
676 |
root |
1.52 |
=head2 THE OpenCL::Object CLASS |
677 |
|
|
|
678 |
|
|
This is the base class for all objects in the OpenCL module. The only |
679 |
|
|
method it implements is the C<id> method, which is only useful if you want |
680 |
|
|
to interface to OpenCL on the C level. |
681 |
|
|
|
682 |
|
|
=over 4 |
683 |
|
|
|
684 |
|
|
=item $iv = $obj->id |
685 |
|
|
|
686 |
|
|
OpenCL objects are represented by pointers or integers on the C level. If |
687 |
|
|
you want to interface to an OpenCL object directly on the C level, then |
688 |
|
|
you need this value, which is returned by this method. You should use an |
689 |
|
|
C<IV> type in your code and cast that to the correct type. |
690 |
|
|
|
691 |
|
|
=cut |
692 |
|
|
|
693 |
|
|
sub OpenCL::Object::id { |
694 |
root |
1.55 |
ref $_[0] eq "SCALAR" |
695 |
|
|
? ${ $_[0] } |
696 |
|
|
: $_[0][0] |
697 |
root |
1.52 |
} |
698 |
|
|
|
699 |
|
|
=back |
700 |
|
|
|
701 |
root |
1.5 |
=head2 THE OpenCL::Platform CLASS |
702 |
|
|
|
703 |
|
|
=over 4 |
704 |
|
|
|
705 |
|
|
=item @devices = $platform->devices ($type = OpenCL::DEVICE_TYPE_ALL) |
706 |
|
|
|
707 |
|
|
Returns a list of matching OpenCL::Device objects. |
708 |
|
|
|
709 |
root |
1.56 |
=item $ctx = $platform->context_from_type ($properties, $type = OpenCL::DEVICE_TYPE_DEFAULT, $callback->($err, $pvt) = $print_stderr) |
710 |
root |
1.5 |
|
711 |
root |
1.22 |
Tries to create a context. Never worked for me, and you need devices explicitly anyway. |
712 |
root |
1.5 |
|
713 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContextFromType.html> |
714 |
|
|
|
715 |
root |
1.56 |
=item $ctx = $platform->context ($properties, \@devices, $callback->($err, $pvt) = $print_stderr) |
716 |
root |
1.11 |
|
717 |
|
|
Create a new OpenCL::Context object using the given device object(s)- a |
718 |
|
|
CL_CONTEXT_PLATFORM property is supplied automatically. |
719 |
|
|
|
720 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateContext.html> |
721 |
|
|
|
722 |
root |
1.20 |
=item $packed_value = $platform->info ($name) |
723 |
|
|
|
724 |
|
|
Calls C<clGetPlatformInfo> and returns the packed, raw value - for |
725 |
root |
1.22 |
strings, this will be the string (possibly including terminating \0), for |
726 |
|
|
other values you probably need to use the correct C<unpack>. |
727 |
root |
1.20 |
|
728 |
root |
1.22 |
It's best to avoid this method and use one of the following convenience |
729 |
|
|
wrappers. |
730 |
root |
1.20 |
|
731 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetPlatformInfo.html> |
732 |
|
|
|
733 |
root |
1.50 |
=item $platform->unload_compiler |
734 |
|
|
|
735 |
|
|
Attempts to unload the compiler for this platform, for endless |
736 |
|
|
profit. Does nothing on OpenCL 1.1. |
737 |
|
|
|
738 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clUnloadPlatformCompiler.html> |
739 |
|
|
|
740 |
root |
1.20 |
=for gengetinfo begin platform |
741 |
|
|
|
742 |
|
|
=item $string = $platform->profile |
743 |
|
|
|
744 |
root |
1.71 |
Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_PROFILE> and returns the result. |
745 |
root |
1.20 |
|
746 |
|
|
=item $string = $platform->version |
747 |
|
|
|
748 |
root |
1.71 |
Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_VERSION> and returns the result. |
749 |
root |
1.20 |
|
750 |
|
|
=item $string = $platform->name |
751 |
|
|
|
752 |
root |
1.71 |
Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_NAME> and returns the result. |
753 |
root |
1.20 |
|
754 |
|
|
=item $string = $platform->vendor |
755 |
|
|
|
756 |
root |
1.71 |
Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_VENDOR> and returns the result. |
757 |
root |
1.20 |
|
758 |
|
|
=item $string = $platform->extensions |
759 |
|
|
|
760 |
root |
1.71 |
Calls C<clGetPlatformInfo> with C<OpenCL::PLATFORM_EXTENSIONS> and returns the result. |
761 |
root |
1.21 |
|
762 |
root |
1.20 |
=for gengetinfo end platform |
763 |
|
|
|
764 |
root |
1.5 |
=back |
765 |
|
|
|
766 |
|
|
=head2 THE OpenCL::Device CLASS |
767 |
|
|
|
768 |
|
|
=over 4 |
769 |
|
|
|
770 |
|
|
=item $packed_value = $device->info ($name) |
771 |
|
|
|
772 |
|
|
See C<< $platform->info >> for details. |
773 |
|
|
|
774 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetDeviceInfo.html> |
775 |
|
|
|
776 |
root |
1.71 |
=item @devices = $device->sub_devices (\@properties) |
777 |
|
|
|
778 |
|
|
Creates OpencL::SubDevice objects by partitioning an existing device. |
779 |
|
|
|
780 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateSubDevices.html> |
781 |
|
|
|
782 |
root |
1.21 |
=for gengetinfo begin device |
783 |
|
|
|
784 |
|
|
=item $device_type = $device->type |
785 |
|
|
|
786 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_TYPE> and returns the result. |
787 |
root |
1.21 |
|
788 |
|
|
=item $uint = $device->vendor_id |
789 |
|
|
|
790 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VENDOR_ID> and returns the result. |
791 |
root |
1.21 |
|
792 |
|
|
=item $uint = $device->max_compute_units |
793 |
|
|
|
794 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_COMPUTE_UNITS> and returns the result. |
795 |
root |
1.21 |
|
796 |
|
|
=item $uint = $device->max_work_item_dimensions |
797 |
|
|
|
798 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_ITEM_DIMENSIONS> and returns the result. |
799 |
root |
1.21 |
|
800 |
|
|
=item $int = $device->max_work_group_size |
801 |
|
|
|
802 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_GROUP_SIZE> and returns the result. |
803 |
root |
1.21 |
|
804 |
|
|
=item @ints = $device->max_work_item_sizes |
805 |
|
|
|
806 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WORK_ITEM_SIZES> and returns the result. |
807 |
root |
1.21 |
|
808 |
|
|
=item $uint = $device->preferred_vector_width_char |
809 |
|
|
|
810 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_CHAR> and returns the result. |
811 |
root |
1.21 |
|
812 |
|
|
=item $uint = $device->preferred_vector_width_short |
813 |
|
|
|
814 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_SHORT> and returns the result. |
815 |
root |
1.21 |
|
816 |
|
|
=item $uint = $device->preferred_vector_width_int |
817 |
|
|
|
818 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_INT> and returns the result. |
819 |
root |
1.21 |
|
820 |
|
|
=item $uint = $device->preferred_vector_width_long |
821 |
|
|
|
822 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_LONG> and returns the result. |
823 |
root |
1.21 |
|
824 |
|
|
=item $uint = $device->preferred_vector_width_float |
825 |
|
|
|
826 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_FLOAT> and returns the result. |
827 |
root |
1.21 |
|
828 |
|
|
=item $uint = $device->preferred_vector_width_double |
829 |
|
|
|
830 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_DOUBLE> and returns the result. |
831 |
root |
1.21 |
|
832 |
|
|
=item $uint = $device->max_clock_frequency |
833 |
|
|
|
834 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CLOCK_FREQUENCY> and returns the result. |
835 |
root |
1.21 |
|
836 |
|
|
=item $bitfield = $device->address_bits |
837 |
|
|
|
838 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ADDRESS_BITS> and returns the result. |
839 |
root |
1.21 |
|
840 |
|
|
=item $uint = $device->max_read_image_args |
841 |
|
|
|
842 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_READ_IMAGE_ARGS> and returns the result. |
843 |
root |
1.21 |
|
844 |
|
|
=item $uint = $device->max_write_image_args |
845 |
|
|
|
846 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_WRITE_IMAGE_ARGS> and returns the result. |
847 |
root |
1.21 |
|
848 |
|
|
=item $ulong = $device->max_mem_alloc_size |
849 |
|
|
|
850 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_MEM_ALLOC_SIZE> and returns the result. |
851 |
root |
1.21 |
|
852 |
|
|
=item $int = $device->image2d_max_width |
853 |
|
|
|
854 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE2D_MAX_WIDTH> and returns the result. |
855 |
root |
1.21 |
|
856 |
|
|
=item $int = $device->image2d_max_height |
857 |
|
|
|
858 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE2D_MAX_HEIGHT> and returns the result. |
859 |
root |
1.21 |
|
860 |
|
|
=item $int = $device->image3d_max_width |
861 |
|
|
|
862 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_WIDTH> and returns the result. |
863 |
root |
1.21 |
|
864 |
|
|
=item $int = $device->image3d_max_height |
865 |
|
|
|
866 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_HEIGHT> and returns the result. |
867 |
root |
1.21 |
|
868 |
|
|
=item $int = $device->image3d_max_depth |
869 |
|
|
|
870 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE3D_MAX_DEPTH> and returns the result. |
871 |
root |
1.21 |
|
872 |
|
|
=item $uint = $device->image_support |
873 |
|
|
|
874 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_IMAGE_SUPPORT> and returns the result. |
875 |
root |
1.21 |
|
876 |
|
|
=item $int = $device->max_parameter_size |
877 |
|
|
|
878 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_PARAMETER_SIZE> and returns the result. |
879 |
root |
1.21 |
|
880 |
|
|
=item $uint = $device->max_samplers |
881 |
|
|
|
882 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_SAMPLERS> and returns the result. |
883 |
root |
1.21 |
|
884 |
|
|
=item $uint = $device->mem_base_addr_align |
885 |
|
|
|
886 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MEM_BASE_ADDR_ALIGN> and returns the result. |
887 |
root |
1.21 |
|
888 |
|
|
=item $uint = $device->min_data_type_align_size |
889 |
|
|
|
890 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MIN_DATA_TYPE_ALIGN_SIZE> and returns the result. |
891 |
root |
1.21 |
|
892 |
|
|
=item $device_fp_config = $device->single_fp_config |
893 |
|
|
|
894 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_SINGLE_FP_CONFIG> and returns the result. |
895 |
root |
1.21 |
|
896 |
|
|
=item $device_mem_cache_type = $device->global_mem_cache_type |
897 |
|
|
|
898 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHE_TYPE> and returns the result. |
899 |
root |
1.21 |
|
900 |
|
|
=item $uint = $device->global_mem_cacheline_size |
901 |
|
|
|
902 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHELINE_SIZE> and returns the result. |
903 |
root |
1.21 |
|
904 |
|
|
=item $ulong = $device->global_mem_cache_size |
905 |
|
|
|
906 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_CACHE_SIZE> and returns the result. |
907 |
root |
1.21 |
|
908 |
|
|
=item $ulong = $device->global_mem_size |
909 |
|
|
|
910 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_GLOBAL_MEM_SIZE> and returns the result. |
911 |
root |
1.21 |
|
912 |
|
|
=item $ulong = $device->max_constant_buffer_size |
913 |
|
|
|
914 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CONSTANT_BUFFER_SIZE> and returns the result. |
915 |
root |
1.21 |
|
916 |
|
|
=item $uint = $device->max_constant_args |
917 |
|
|
|
918 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_MAX_CONSTANT_ARGS> and returns the result. |
919 |
root |
1.21 |
|
920 |
|
|
=item $device_local_mem_type = $device->local_mem_type |
921 |
|
|
|
922 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_LOCAL_MEM_TYPE> and returns the result. |
923 |
root |
1.21 |
|
924 |
|
|
=item $ulong = $device->local_mem_size |
925 |
|
|
|
926 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_LOCAL_MEM_SIZE> and returns the result. |
927 |
root |
1.21 |
|
928 |
|
|
=item $boolean = $device->error_correction_support |
929 |
|
|
|
930 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ERROR_CORRECTION_SUPPORT> and returns the result. |
931 |
root |
1.21 |
|
932 |
|
|
=item $int = $device->profiling_timer_resolution |
933 |
|
|
|
934 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PROFILING_TIMER_RESOLUTION> and returns the result. |
935 |
root |
1.21 |
|
936 |
|
|
=item $boolean = $device->endian_little |
937 |
|
|
|
938 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_ENDIAN_LITTLE> and returns the result. |
939 |
root |
1.21 |
|
940 |
|
|
=item $boolean = $device->available |
941 |
|
|
|
942 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_AVAILABLE> and returns the result. |
943 |
root |
1.21 |
|
944 |
|
|
=item $boolean = $device->compiler_available |
945 |
|
|
|
946 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_COMPILER_AVAILABLE> and returns the result. |
947 |
root |
1.21 |
|
948 |
|
|
=item $device_exec_capabilities = $device->execution_capabilities |
949 |
|
|
|
950 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_EXECUTION_CAPABILITIES> and returns the result. |
951 |
root |
1.21 |
|
952 |
|
|
=item $command_queue_properties = $device->properties |
953 |
|
|
|
954 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_QUEUE_PROPERTIES> and returns the result. |
955 |
root |
1.21 |
|
956 |
|
|
=item $ = $device->platform |
957 |
|
|
|
958 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PLATFORM> and returns the result. |
959 |
root |
1.21 |
|
960 |
|
|
=item $string = $device->name |
961 |
|
|
|
962 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NAME> and returns the result. |
963 |
root |
1.21 |
|
964 |
|
|
=item $string = $device->vendor |
965 |
|
|
|
966 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VENDOR> and returns the result. |
967 |
root |
1.21 |
|
968 |
|
|
=item $string = $device->driver_version |
969 |
|
|
|
970 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DRIVER_VERSION> and returns the result. |
971 |
root |
1.21 |
|
972 |
|
|
=item $string = $device->profile |
973 |
|
|
|
974 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PROFILE> and returns the result. |
975 |
root |
1.21 |
|
976 |
|
|
=item $string = $device->version |
977 |
|
|
|
978 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_VERSION> and returns the result. |
979 |
root |
1.21 |
|
980 |
|
|
=item $string = $device->extensions |
981 |
|
|
|
982 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_EXTENSIONS> and returns the result. |
983 |
root |
1.21 |
|
984 |
|
|
=item $uint = $device->preferred_vector_width_half |
985 |
|
|
|
986 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PREFERRED_VECTOR_WIDTH_HALF> and returns the result. |
987 |
root |
1.21 |
|
988 |
|
|
=item $uint = $device->native_vector_width_char |
989 |
|
|
|
990 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_CHAR> and returns the result. |
991 |
root |
1.21 |
|
992 |
|
|
=item $uint = $device->native_vector_width_short |
993 |
|
|
|
994 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_SHORT> and returns the result. |
995 |
root |
1.21 |
|
996 |
|
|
=item $uint = $device->native_vector_width_int |
997 |
|
|
|
998 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_INT> and returns the result. |
999 |
root |
1.21 |
|
1000 |
|
|
=item $uint = $device->native_vector_width_long |
1001 |
|
|
|
1002 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_LONG> and returns the result. |
1003 |
root |
1.21 |
|
1004 |
|
|
=item $uint = $device->native_vector_width_float |
1005 |
|
|
|
1006 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_FLOAT> and returns the result. |
1007 |
root |
1.21 |
|
1008 |
|
|
=item $uint = $device->native_vector_width_double |
1009 |
|
|
|
1010 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_DOUBLE> and returns the result. |
1011 |
root |
1.21 |
|
1012 |
|
|
=item $uint = $device->native_vector_width_half |
1013 |
|
|
|
1014 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_NATIVE_VECTOR_WIDTH_HALF> and returns the result. |
1015 |
root |
1.21 |
|
1016 |
|
|
=item $device_fp_config = $device->double_fp_config |
1017 |
|
|
|
1018 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_DOUBLE_FP_CONFIG> and returns the result. |
1019 |
root |
1.21 |
|
1020 |
|
|
=item $device_fp_config = $device->half_fp_config |
1021 |
|
|
|
1022 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_HALF_FP_CONFIG> and returns the result. |
1023 |
root |
1.21 |
|
1024 |
|
|
=item $boolean = $device->host_unified_memory |
1025 |
|
|
|
1026 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_HOST_UNIFIED_MEMORY> and returns the result. |
1027 |
root |
1.21 |
|
1028 |
|
|
=item $device = $device->parent_device_ext |
1029 |
|
|
|
1030 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARENT_DEVICE_EXT> and returns the result. |
1031 |
root |
1.21 |
|
1032 |
|
|
=item @device_partition_property_exts = $device->partition_types_ext |
1033 |
|
|
|
1034 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARTITION_TYPES_EXT> and returns the result. |
1035 |
root |
1.21 |
|
1036 |
|
|
=item @device_partition_property_exts = $device->affinity_domains_ext |
1037 |
|
|
|
1038 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_AFFINITY_DOMAINS_EXT> and returns the result. |
1039 |
root |
1.21 |
|
1040 |
root |
1.45 |
=item $uint = $device->reference_count_ext |
1041 |
root |
1.21 |
|
1042 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_REFERENCE_COUNT_EXT> and returns the result. |
1043 |
root |
1.21 |
|
1044 |
|
|
=item @device_partition_property_exts = $device->partition_style_ext |
1045 |
|
|
|
1046 |
root |
1.71 |
Calls C<clGetDeviceInfo> with C<OpenCL::DEVICE_PARTITION_STYLE_EXT> and returns the result. |
1047 |
root |
1.21 |
|
1048 |
|
|
=for gengetinfo end device |
1049 |
|
|
|
1050 |
root |
1.5 |
=back |
1051 |
|
|
|
1052 |
|
|
=head2 THE OpenCL::Context CLASS |
1053 |
|
|
|
1054 |
root |
1.71 |
An OpenCL::Context is basically a container, or manager, for a number of |
1055 |
|
|
devices of a platform. It is used to create all sorts of secondary objects |
1056 |
|
|
such as buffers, queues, programs and so on. |
1057 |
|
|
|
1058 |
|
|
All context creation functions and methods take a list of properties |
1059 |
|
|
(type-value pairs). All property values can be specified as integers - |
1060 |
|
|
some additionally support other types: |
1061 |
|
|
|
1062 |
|
|
=over 4 |
1063 |
|
|
|
1064 |
|
|
=item OpenCL::CONTEXT_PLATFORM |
1065 |
|
|
|
1066 |
|
|
Also accepts OpenCL::Platform objects. |
1067 |
|
|
|
1068 |
|
|
=item OpenCL::GLX_DISPLAY_KHR |
1069 |
|
|
|
1070 |
|
|
Also accepts C<undef>, in which case a deep and troubling hack is engaged |
1071 |
|
|
to find the current glx display (see L<GLX SUPPORT>). |
1072 |
|
|
|
1073 |
|
|
=item OpenCL::GL_CONTEXT_KHR |
1074 |
|
|
|
1075 |
|
|
Also accepts C<undef>, in which case a deep and troubling hack is engaged |
1076 |
|
|
to find the current glx context (see L<GLX SUPPORT>). |
1077 |
|
|
|
1078 |
|
|
=back |
1079 |
|
|
|
1080 |
root |
1.5 |
=over 4 |
1081 |
|
|
|
1082 |
root |
1.51 |
=item $prog = $ctx->build_program ($program, $options = "") |
1083 |
|
|
|
1084 |
|
|
This convenience function tries to build the program on all devices in |
1085 |
|
|
the context. If the build fails, then the function will C<croak> with the |
1086 |
|
|
build log. Otherwise ti returns the program object. |
1087 |
|
|
|
1088 |
|
|
The C<$program> can either be a C<OpenCL::Program> object or a string |
1089 |
|
|
containing the program. In the latter case, a program objetc will be |
1090 |
|
|
created automatically. |
1091 |
|
|
|
1092 |
|
|
=cut |
1093 |
|
|
|
1094 |
|
|
sub OpenCL::Context::build_program { |
1095 |
|
|
my ($self, $prog, $options) = @_; |
1096 |
|
|
|
1097 |
|
|
$prog = $self->program_with_source ($prog) |
1098 |
|
|
unless ref $prog; |
1099 |
|
|
|
1100 |
root |
1.61 |
eval { $prog->build (undef, $options); 1 } |
1101 |
|
|
or errno == BUILD_PROGRAM_FAILURE |
1102 |
root |
1.63 |
or errno == INVALID_BINARY # workaround nvidia bug |
1103 |
root |
1.61 |
or Carp::croak "OpenCL::Context->build_program: " . err2str; |
1104 |
|
|
|
1105 |
|
|
# we check status for all devices |
1106 |
root |
1.51 |
for my $dev ($self->devices) { |
1107 |
root |
1.61 |
$prog->build_status ($dev) == BUILD_SUCCESS |
1108 |
root |
1.62 |
or Carp::croak "Building OpenCL program for device '" . $dev->name . "' failed:\n" |
1109 |
|
|
. $prog->build_log ($dev); |
1110 |
root |
1.51 |
} |
1111 |
|
|
|
1112 |
|
|
$prog |
1113 |
|
|
} |
1114 |
|
|
|
1115 |
root |
1.9 |
=item $queue = $ctx->queue ($device, $properties) |
1116 |
root |
1.5 |
|
1117 |
root |
1.9 |
Create a new OpenCL::Queue object from the context and the given device. |
1118 |
root |
1.5 |
|
1119 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateCommandQueue.html> |
1120 |
|
|
|
1121 |
root |
1.45 |
Example: create an out-of-order queue. |
1122 |
|
|
|
1123 |
|
|
$queue = $ctx->queue ($device, OpenCL::QUEUE_OUT_OF_ORDER_EXEC_MODE_ENABLE); |
1124 |
|
|
|
1125 |
root |
1.5 |
=item $ev = $ctx->user_event |
1126 |
|
|
|
1127 |
|
|
Creates a new OpenCL::UserEvent object. |
1128 |
|
|
|
1129 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateUserEvent.html> |
1130 |
|
|
|
1131 |
|
|
=item $buf = $ctx->buffer ($flags, $len) |
1132 |
|
|
|
1133 |
root |
1.27 |
Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object with the |
1134 |
|
|
given flags and octet-size. |
1135 |
root |
1.5 |
|
1136 |
root |
1.71 |
flags: OpenCL::MEM_READ_WRITE, OpenCL::MEM_WRITE_ONLY, OpenCL::MEM_READ_ONLY, |
1137 |
|
|
OpenCL::MEM_USE_HOST_PTR, OpenCL::MEM_ALLOC_HOST_PTR, OpenCL::MEM_COPY_HOST_PTR, |
1138 |
|
|
OpenCL::MEM_HOST_WRITE_ONLY, OpenCL::MEM_HOST_READ_ONLY, OpenCL::MEM_HOST_NO_ACCESS. |
1139 |
|
|
|
1140 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateBuffer.html> |
1141 |
|
|
|
1142 |
|
|
=item $buf = $ctx->buffer_sv ($flags, $data) |
1143 |
|
|
|
1144 |
root |
1.27 |
Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object and |
1145 |
|
|
initialise it with the given data values. |
1146 |
root |
1.5 |
|
1147 |
root |
1.59 |
=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) |
1148 |
root |
1.49 |
|
1149 |
|
|
Creates a new OpenCL::Image object and optionally initialises it with |
1150 |
|
|
the given data values. |
1151 |
|
|
|
1152 |
root |
1.71 |
channel_order: OpenCL::R, OpenCL::A, OpenCL::RG, OpenCL::RA, OpenCL::RGB, |
1153 |
|
|
OpenCL::RGBA, OpenCL::BGRA, OpenCL::ARGB, OpenCL::INTENSITY, OpenCL::LUMINANCE, |
1154 |
|
|
OpenCL::Rx, OpenCL::RGx, OpenCL::RGBx. |
1155 |
|
|
|
1156 |
|
|
channel_type: OpenCL::SNORM_INT8, OpenCL::SNORM_INT16, OpenCL::UNORM_INT8, |
1157 |
|
|
OpenCL::UNORM_INT16, OpenCL::UNORM_SHORT_565, OpenCL::UNORM_SHORT_555, |
1158 |
|
|
OpenCL::UNORM_INT_101010, OpenCL::SIGNED_INT8, OpenCL::SIGNED_INT16, |
1159 |
|
|
OpenCL::SIGNED_INT32, OpenCL::UNSIGNED_INT8, OpenCL::UNSIGNED_INT16, |
1160 |
|
|
OpenCL::UNSIGNED_INT32, OpenCL::HALF_FLOAT, OpenCL::FLOAT. |
1161 |
|
|
|
1162 |
|
|
|
1163 |
|
|
type: OpenCL::MEM_OBJECT_BUFFER, OpenCL::MEM_OBJECT_IMAGE2D, |
1164 |
|
|
OpenCL::MEM_OBJECT_IMAGE3D, OpenCL::MEM_OBJECT_IMAGE2D_ARRAY, |
1165 |
|
|
OpenCL::MEM_OBJECT_IMAGE1D, OpenCL::MEM_OBJECT_IMAGE1D_ARRAY, |
1166 |
|
|
OpenCL::MEM_OBJECT_IMAGE1D_BUFFER. |
1167 |
|
|
|
1168 |
root |
1.49 |
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage.html> |
1169 |
|
|
|
1170 |
root |
1.18 |
=item $img = $ctx->image2d ($flags, $channel_order, $channel_type, $width, $height, $row_pitch = 0, $data = undef) |
1171 |
root |
1.5 |
|
1172 |
root |
1.27 |
Creates a new OpenCL::Image2D object and optionally initialises it with |
1173 |
|
|
the given data values. |
1174 |
root |
1.5 |
|
1175 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage2D.html> |
1176 |
|
|
|
1177 |
root |
1.18 |
=item $img = $ctx->image3d ($flags, $channel_order, $channel_type, $width, $height, $depth, $row_pitch = 0, $slice_pitch = 0, $data = undef) |
1178 |
root |
1.5 |
|
1179 |
root |
1.27 |
Creates a new OpenCL::Image3D object and optionally initialises it with |
1180 |
|
|
the given data values. |
1181 |
root |
1.5 |
|
1182 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateImage3D.html> |
1183 |
|
|
|
1184 |
root |
1.33 |
=item $buffer = $ctx->gl_buffer ($flags, $bufobj) |
1185 |
|
|
|
1186 |
|
|
Creates a new OpenCL::Buffer (actually OpenCL::BufferObj) object that refers to the given |
1187 |
|
|
OpenGL buffer object. |
1188 |
|
|
|
1189 |
root |
1.71 |
flags: OpenCL::MEM_READ_WRITE, OpenCL::MEM_READ_ONLY, OpenCL::MEM_WRITE_ONLY. |
1190 |
|
|
|
1191 |
root |
1.33 |
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLBuffer.html |
1192 |
|
|
|
1193 |
root |
1.47 |
=item $img = $ctx->gl_texture ($flags, $target, $miplevel, $texture) |
1194 |
|
|
|
1195 |
|
|
Creates a new OpenCL::Image object that refers to the given OpenGL |
1196 |
|
|
texture object or buffer. |
1197 |
|
|
|
1198 |
root |
1.71 |
target: GL_TEXTURE_1D, GL_TEXTURE_1D_ARRAY, GL_TEXTURE_BUFFER, |
1199 |
|
|
GL_TEXTURE_2D, GL_TEXTURE_2D_ARRAY, GL_TEXTURE_3D, |
1200 |
|
|
GL_TEXTURE_CUBE_MAP_POSITIVE_X, GL_TEXTURE_CUBE_MAP_POSITIVE_Y, |
1201 |
|
|
GL_TEXTURE_CUBE_MAP_POSITIVE_Z, GL_TEXTURE_CUBE_MAP_NEGATIVE_X, |
1202 |
|
|
GL_TEXTURE_CUBE_MAP_NEGATIVE_Y, GL_TEXTURE_CUBE_MAP_NEGATIVE_Z, |
1203 |
|
|
GL_TEXTURE_RECTANGLE/GL_TEXTURE_RECTANGLE_ARB. |
1204 |
|
|
|
1205 |
root |
1.47 |
http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture.html |
1206 |
|
|
|
1207 |
|
|
=item $img = $ctx->gl_texture2d ($flags, $target, $miplevel, $texture) |
1208 |
root |
1.33 |
|
1209 |
|
|
Creates a new OpenCL::Image2D object that refers to the given OpenGL |
1210 |
|
|
2D texture object. |
1211 |
|
|
|
1212 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture2D.html |
1213 |
|
|
|
1214 |
root |
1.47 |
=item $img = $ctx->gl_texture3d ($flags, $target, $miplevel, $texture) |
1215 |
root |
1.33 |
|
1216 |
|
|
Creates a new OpenCL::Image3D object that refers to the given OpenGL |
1217 |
|
|
3D texture object. |
1218 |
|
|
|
1219 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLTexture3D.html |
1220 |
|
|
|
1221 |
|
|
=item $ctx->gl_renderbuffer ($flags, $renderbuffer) |
1222 |
|
|
|
1223 |
|
|
Creates a new OpenCL::Image2D object that refers to the given OpenGL |
1224 |
|
|
render buffer. |
1225 |
|
|
|
1226 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateFromGLRenderbuffer.html |
1227 |
|
|
|
1228 |
root |
1.5 |
=item @formats = $ctx->supported_image_formats ($flags, $image_type) |
1229 |
|
|
|
1230 |
|
|
Returns a list of matching image formats - each format is an arrayref with |
1231 |
|
|
two values, $channel_order and $channel_type, in it. |
1232 |
|
|
|
1233 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSupportedImageFormats.html> |
1234 |
|
|
|
1235 |
|
|
=item $sampler = $ctx->sampler ($normalized_coords, $addressing_mode, $filter_mode) |
1236 |
|
|
|
1237 |
|
|
Creates a new OpenCL::Sampler object. |
1238 |
|
|
|
1239 |
root |
1.71 |
addressing_mode: OpenCL::ADDRESS_NONE, OpenCL::ADDRESS_CLAMP_TO_EDGE, |
1240 |
|
|
OpenCL::ADDRESS_CLAMP, OpenCL::ADDRESS_REPEAT, OpenCL::ADDRESS_MIRRORED_REPEAT. |
1241 |
|
|
|
1242 |
|
|
filter_mode: OpenCL::FILTER_NEAREST, OpenCL::FILTER_LINEAR. |
1243 |
|
|
|
1244 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSampler.html> |
1245 |
|
|
|
1246 |
|
|
=item $program = $ctx->program_with_source ($string) |
1247 |
|
|
|
1248 |
|
|
Creates a new OpenCL::Program object from the given source code. |
1249 |
|
|
|
1250 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithSource.html> |
1251 |
|
|
|
1252 |
root |
1.69 |
=item ($program, \@status) = $ctx->program_with_binary (\@devices, \@binaries) |
1253 |
|
|
|
1254 |
|
|
Creates a new OpenCL::Program object from the given binaries. |
1255 |
|
|
|
1256 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateProgramWithBinary.html> |
1257 |
|
|
|
1258 |
|
|
Example: clone an existing program object that contains a successfully |
1259 |
|
|
compiled program, no matter how useless this is. |
1260 |
|
|
|
1261 |
|
|
my $clone = $ctx->program_with_binary ([$prog->devices], [$prog->binaries]); |
1262 |
|
|
|
1263 |
root |
1.71 |
=item $program = $ctx->program_with_built_in_kernels (\@devices, $kernel_names) |
1264 |
|
|
|
1265 |
|
|
Creates a new OpenCL::Program object from the given built-in kernel names. |
1266 |
|
|
|
1267 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateProgramWithBuiltInKernels.html> |
1268 |
|
|
|
1269 |
root |
1.20 |
=item $packed_value = $ctx->info ($name) |
1270 |
|
|
|
1271 |
|
|
See C<< $platform->info >> for details. |
1272 |
|
|
|
1273 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetContextInfo.html> |
1274 |
|
|
|
1275 |
|
|
=for gengetinfo begin context |
1276 |
|
|
|
1277 |
root |
1.21 |
=item $uint = $context->reference_count |
1278 |
|
|
|
1279 |
root |
1.71 |
Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_REFERENCE_COUNT> and returns the result. |
1280 |
root |
1.21 |
|
1281 |
|
|
=item @devices = $context->devices |
1282 |
|
|
|
1283 |
root |
1.71 |
Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_DEVICES> and returns the result. |
1284 |
root |
1.21 |
|
1285 |
|
|
=item @property_ints = $context->properties |
1286 |
|
|
|
1287 |
root |
1.71 |
Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_PROPERTIES> and returns the result. |
1288 |
root |
1.21 |
|
1289 |
|
|
=item $uint = $context->num_devices |
1290 |
|
|
|
1291 |
root |
1.71 |
Calls C<clGetContextInfo> with C<OpenCL::CONTEXT_NUM_DEVICES> and returns the result. |
1292 |
root |
1.21 |
|
1293 |
root |
1.20 |
=for gengetinfo end context |
1294 |
|
|
|
1295 |
root |
1.5 |
=back |
1296 |
|
|
|
1297 |
|
|
=head2 THE OpenCL::Queue CLASS |
1298 |
|
|
|
1299 |
|
|
An OpenCL::Queue represents an execution queue for OpenCL. You execute |
1300 |
root |
1.59 |
requests by calling their respective method and waiting for it to complete |
1301 |
|
|
in some way. |
1302 |
root |
1.5 |
|
1303 |
root |
1.59 |
Most methods that enqueue some request return an event object that can |
1304 |
|
|
be used to wait for completion (optionally using a callback), unless |
1305 |
|
|
the method is called in void context, in which case no event object is |
1306 |
|
|
created. |
1307 |
root |
1.5 |
|
1308 |
|
|
They also allow you to specify any number of other event objects that this |
1309 |
|
|
request has to wait for before it starts executing, by simply passing the |
1310 |
root |
1.45 |
event objects as extra parameters to the enqueue methods. To simplify |
1311 |
|
|
program design, this module ignores any C<undef> values in the list of |
1312 |
|
|
events. This makes it possible to code operations such as this, without |
1313 |
|
|
having to put a valid event object into C<$event> first: |
1314 |
|
|
|
1315 |
root |
1.59 |
$event = $queue->xxx (..., $event); |
1316 |
root |
1.5 |
|
1317 |
|
|
Queues execute in-order by default, without any parallelism, so in most |
1318 |
root |
1.6 |
cases (i.e. you use only one queue) it's not necessary to wait for or |
1319 |
root |
1.45 |
create event objects, althoguh an our of order queue is often a bit |
1320 |
|
|
faster. |
1321 |
root |
1.5 |
|
1322 |
|
|
=over 4 |
1323 |
|
|
|
1324 |
root |
1.59 |
=item $ev = $queue->read_buffer ($buffer, $blocking, $offset, $len, $data, $wait_events...) |
1325 |
root |
1.5 |
|
1326 |
|
|
Reads data from buffer into the given string. |
1327 |
|
|
|
1328 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBuffer.html> |
1329 |
|
|
|
1330 |
root |
1.59 |
=item $ev = $queue->write_buffer ($buffer, $blocking, $offset, $data, $wait_events...) |
1331 |
root |
1.5 |
|
1332 |
|
|
Writes data to buffer from the given string. |
1333 |
|
|
|
1334 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBuffer.html> |
1335 |
|
|
|
1336 |
root |
1.59 |
=item $ev = $queue->copy_buffer ($src, $dst, $src_offset, $dst_offset, $len, $wait_events...) |
1337 |
root |
1.5 |
|
1338 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBuffer.html> |
1339 |
|
|
|
1340 |
root |
1.59 |
=item $ev = $queue->read_buffer_rect (OpenCL::Memory buf, cl_bool blocking, $buf_x, $buf_y, $buf_z, $host_x, $host_y, $host_z, $width, $height, $depth, $buf_row_pitch, $buf_slice_pitch, $host_row_pitch, $host_slice_pitch, $data, $wait_events...) |
1341 |
root |
1.25 |
|
1342 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadBufferRect.html |
1343 |
|
|
|
1344 |
root |
1.59 |
=item $ev = $queue->write_buffer_rect (OpenCL::Memory buf, cl_bool blocking, $buf_x, $buf_y, $buf_z, $host_x, $host_y, $host_z, $width, $height, $depth, $buf_row_pitch, $buf_slice_pitch, $host_row_pitch, $host_slice_pitch, $data, $wait_events...) |
1345 |
root |
1.25 |
|
1346 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteBufferRect.html |
1347 |
|
|
|
1348 |
root |
1.65 |
=item $ev = $queue->copy_buffer_to_image ($src_buffer, $dst_image, $src_offset, $dst_x, $dst_y, $dst_z, $width, $height, $depth, $wait_events...) |
1349 |
|
|
|
1350 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html> |
1351 |
|
|
|
1352 |
root |
1.59 |
=item $ev = $queue->read_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) |
1353 |
root |
1.5 |
|
1354 |
root |
1.65 |
C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL |
1355 |
|
|
module uses the image width (and height) to supply default values. |
1356 |
root |
1.27 |
|
1357 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReadImage.html> |
1358 |
|
|
|
1359 |
root |
1.59 |
=item $ev = $queue->write_image ($src, $blocking, $x, $y, $z, $width, $height, $depth, $row_pitch, $slice_pitch, $data, $wait_events...) |
1360 |
root |
1.5 |
|
1361 |
root |
1.65 |
C<$row_pitch> (and C<$slice_pitch>) can be C<0>, in which case the OpenCL |
1362 |
|
|
module uses the image width (and height) to supply default values. |
1363 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWriteImage.html> |
1364 |
|
|
|
1365 |
root |
1.59 |
=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...) |
1366 |
root |
1.5 |
|
1367 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImage.html> |
1368 |
|
|
|
1369 |
root |
1.59 |
=item $ev = $queue->copy_image_to_buffer ($src_image, $dst_image, $src_x, $src_y, $src_z, $width, $height, $depth, $dst_offset, $wait_events...) |
1370 |
root |
1.5 |
|
1371 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyImageToBuffer.html> |
1372 |
|
|
|
1373 |
root |
1.59 |
=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...) |
1374 |
root |
1.27 |
|
1375 |
|
|
Yeah. |
1376 |
|
|
|
1377 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueCopyBufferToImage.html>. |
1378 |
|
|
|
1379 |
root |
1.59 |
=item $ev = $queue->fill_buffer ($mem, $pattern, $offset, $size, ...) |
1380 |
root |
1.52 |
|
1381 |
|
|
Fills the given buffer object with repeated applications of C<$pattern>, |
1382 |
|
|
starting at C<$offset> for C<$size> octets. |
1383 |
|
|
|
1384 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillBuffer.html> |
1385 |
|
|
|
1386 |
root |
1.59 |
=item $ev = $queue->fill_image ($img, $r, $g, $b, $a, $x, $y, $z, $width, $height, $depth, ...) |
1387 |
root |
1.52 |
|
1388 |
|
|
Fills the given image area with the given rgba colour components. The |
1389 |
|
|
components are normally floating point values between C<0> and C<1>, |
1390 |
|
|
except when the image channel data type is a signe dor unsigned |
1391 |
|
|
unnormalised format, in which case the range is determined by the format. |
1392 |
|
|
|
1393 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueFillImage.html> |
1394 |
|
|
|
1395 |
root |
1.59 |
=item $ev = $queue->task ($kernel, $wait_events...) |
1396 |
root |
1.5 |
|
1397 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueTask.html> |
1398 |
|
|
|
1399 |
root |
1.59 |
=item $ev = $queue->nd_range_kernel ($kernel, \@global_work_offset, \@global_work_size, \@local_work_size, $wait_events...) |
1400 |
root |
1.5 |
|
1401 |
|
|
Enqueues a kernel execution. |
1402 |
|
|
|
1403 |
root |
1.57 |
\@global_work_size must be specified as a reference to an array of |
1404 |
root |
1.5 |
integers specifying the work sizes (element counts). |
1405 |
|
|
|
1406 |
root |
1.57 |
\@global_work_offset must be either C<undef> (in which case all offsets |
1407 |
root |
1.5 |
are C<0>), or a reference to an array of work offsets, with the same number |
1408 |
root |
1.57 |
of elements as \@global_work_size. |
1409 |
root |
1.5 |
|
1410 |
root |
1.57 |
\@local_work_size must be either C<undef> (in which case the |
1411 |
root |
1.5 |
implementation is supposed to choose good local work sizes), or a |
1412 |
|
|
reference to an array of local work sizes, with the same number of |
1413 |
root |
1.57 |
elements as \@global_work_size. |
1414 |
root |
1.5 |
|
1415 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueNDRangeKernel.html> |
1416 |
|
|
|
1417 |
root |
1.71 |
=item $ev = $queue->migrate_mem_objects (\@mem_objects, $flags, $wait_events...) |
1418 |
|
|
|
1419 |
|
|
Migrates a number of OpenCL::Memory objects to or from the device. |
1420 |
|
|
|
1421 |
|
|
flags: OpenCL::MIGRATE_MEM_OBJECT_HOST, OpenCL::MIGRATE_MEM_OBJECT_CONTENT_UNDEFINED |
1422 |
|
|
|
1423 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMigrateMemObjects.html> |
1424 |
|
|
|
1425 |
root |
1.59 |
=item $ev = $queue->acquire_gl_objects ([object, ...], $wait_events...) |
1426 |
root |
1.35 |
|
1427 |
|
|
Enqueues a list (an array-ref of OpenCL::Memory objects) to be acquired |
1428 |
|
|
for subsequent OpenCL usage. |
1429 |
|
|
|
1430 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueAcquireGLObjects.html> |
1431 |
|
|
|
1432 |
root |
1.59 |
=item $ev = $queue->release_gl_objects ([object, ...], $wait_events...) |
1433 |
root |
1.35 |
|
1434 |
|
|
Enqueues a list (an array-ref of OpenCL::Memory objects) to be released |
1435 |
|
|
for subsequent OpenGL usage. |
1436 |
|
|
|
1437 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueReleaseGLObjects.html> |
1438 |
|
|
|
1439 |
root |
1.59 |
=item $ev = $queue->wait_for_events ($wait_events...) |
1440 |
root |
1.5 |
|
1441 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueWaitForEvents.html> |
1442 |
|
|
|
1443 |
root |
1.59 |
=item $ev = $queue->marker ($wait_events...) |
1444 |
root |
1.46 |
|
1445 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueMarkerWithWaitList.html> |
1446 |
|
|
|
1447 |
root |
1.59 |
=item $ev = $queue->barrier ($wait_events...) |
1448 |
root |
1.5 |
|
1449 |
root |
1.46 |
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clEnqueueBarrierWithWaitList.html> |
1450 |
root |
1.5 |
|
1451 |
|
|
=item $queue->flush |
1452 |
|
|
|
1453 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFlush.html> |
1454 |
|
|
|
1455 |
|
|
=item $queue->finish |
1456 |
|
|
|
1457 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clFinish.html> |
1458 |
|
|
|
1459 |
root |
1.21 |
=item $packed_value = $queue->info ($name) |
1460 |
|
|
|
1461 |
|
|
See C<< $platform->info >> for details. |
1462 |
|
|
|
1463 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetCommandQueueInfo.html> |
1464 |
|
|
|
1465 |
|
|
=for gengetinfo begin command_queue |
1466 |
|
|
|
1467 |
|
|
=item $ctx = $command_queue->context |
1468 |
|
|
|
1469 |
root |
1.71 |
Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_CONTEXT> and returns the result. |
1470 |
root |
1.21 |
|
1471 |
|
|
=item $device = $command_queue->device |
1472 |
|
|
|
1473 |
root |
1.71 |
Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_DEVICE> and returns the result. |
1474 |
root |
1.21 |
|
1475 |
|
|
=item $uint = $command_queue->reference_count |
1476 |
|
|
|
1477 |
root |
1.71 |
Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_REFERENCE_COUNT> and returns the result. |
1478 |
root |
1.21 |
|
1479 |
|
|
=item $command_queue_properties = $command_queue->properties |
1480 |
|
|
|
1481 |
root |
1.71 |
Calls C<clGetCommandQueueInfo> with C<OpenCL::QUEUE_PROPERTIES> and returns the result. |
1482 |
root |
1.21 |
|
1483 |
|
|
=for gengetinfo end command_queue |
1484 |
|
|
|
1485 |
root |
1.5 |
=back |
1486 |
|
|
|
1487 |
root |
1.66 |
=head3 MEMORY MAPPED BUFFERS |
1488 |
|
|
|
1489 |
|
|
OpenCL allows you to map buffers and images to host memory (read: perl |
1490 |
|
|
scalars). This is done much like reading or copying a buffer, by enqueuing |
1491 |
|
|
a map or unmap operation on the command queue. |
1492 |
|
|
|
1493 |
root |
1.69 |
The map operations return an C<OpenCL::Mapped> object - see L<THE |
1494 |
root |
1.66 |
OpenCL::Mapped CLASS> section for details on what to do with these |
1495 |
|
|
objects. |
1496 |
|
|
|
1497 |
|
|
The object will be unmapped automatically when the mapped object is |
1498 |
|
|
destroyed (you can use a barrier to make sure the unmap has finished, |
1499 |
|
|
before using the buffer in a kernel), but you can also enqueue an unmap |
1500 |
|
|
operation manually. |
1501 |
|
|
|
1502 |
|
|
=over 4 |
1503 |
|
|
|
1504 |
root |
1.69 |
=item $mapped_buffer = $queue->map_buffer ($buf, $blocking=1, $map_flags=OpenCL::MAP_READ|OpenCL::MAP_WRITE, $offset=0, $size=undef, $wait_events...) |
1505 |
root |
1.66 |
|
1506 |
root |
1.69 |
Maps the given buffer into host memory and returns an |
1507 |
|
|
C<OpenCL::MappedBuffer> object. If C<$size> is specified as undef, then |
1508 |
|
|
the map will extend to the end of the buffer. |
1509 |
root |
1.66 |
|
1510 |
root |
1.71 |
map_flags: OpenCL::MAP_READ, OpenCL::MAP_WRITE, OpenCL::MAP_WRITE_INVALIDATE_REGION. |
1511 |
|
|
|
1512 |
root |
1.66 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapBuffer.html> |
1513 |
|
|
|
1514 |
root |
1.69 |
Example: map the buffer $buf fully and replace the first 4 bytes by "abcd", then unmap. |
1515 |
|
|
|
1516 |
|
|
{ |
1517 |
|
|
my $mapped = $queue->map_buffer ($buf, 1, OpenCL::MAP_WRITE); |
1518 |
|
|
substr $$mapped, 0, 4, "abcd"; |
1519 |
|
|
} # asynchronously unmap because $mapped is destroyed |
1520 |
root |
1.66 |
|
1521 |
root |
1.69 |
=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...) |
1522 |
|
|
|
1523 |
|
|
Maps the given image area into host memory and return an |
1524 |
|
|
C<OpenCL::MappedImage> object. |
1525 |
|
|
|
1526 |
|
|
If any of C<$width>, C<$height> and/or C<$depth> are C<undef> then they |
1527 |
|
|
will be replaced by the maximum possible value. |
1528 |
root |
1.66 |
|
1529 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clEnqueueMapImage.html> |
1530 |
|
|
|
1531 |
root |
1.69 |
Example: map an image (with OpenCL::UNSIGNED_INT8 channel type) and set |
1532 |
|
|
the first channel of the leftmost column to 5, then explicitly unmap |
1533 |
|
|
it. You are not necessarily meant to do it this way, this example just |
1534 |
|
|
shows you the accessors to use :) |
1535 |
|
|
|
1536 |
|
|
my $mapped = $queue->map_image ($image, 1, OpenCL::MAP_WRITE); |
1537 |
|
|
|
1538 |
|
|
$mapped->set ($_ * $mapped->row_pitch, pack "C", 5) |
1539 |
|
|
for 0..$image->height; |
1540 |
|
|
|
1541 |
|
|
$mapped->unmap;. |
1542 |
|
|
$mapped->wait; # only needed for out of order queues normally |
1543 |
|
|
|
1544 |
root |
1.66 |
=item $ev = $queue->unmap ($mapped, $wait_events...) |
1545 |
|
|
|
1546 |
|
|
Unmaps the data from host memory. You must not call any methods that |
1547 |
|
|
modify the data, or modify the data scalar directly, after calling this |
1548 |
|
|
method. |
1549 |
|
|
|
1550 |
|
|
The mapped event object will always be passed as part of the |
1551 |
|
|
$wait_events. The mapped event object will be replaced by the new event |
1552 |
|
|
object that this request creates. |
1553 |
|
|
|
1554 |
|
|
=back |
1555 |
|
|
|
1556 |
root |
1.5 |
=head2 THE OpenCL::Memory CLASS |
1557 |
|
|
|
1558 |
|
|
This the superclass of all memory objects - OpenCL::Buffer, OpenCL::Image, |
1559 |
root |
1.21 |
OpenCL::Image2D and OpenCL::Image3D. |
1560 |
root |
1.5 |
|
1561 |
|
|
=over 4 |
1562 |
|
|
|
1563 |
|
|
=item $packed_value = $memory->info ($name) |
1564 |
|
|
|
1565 |
|
|
See C<< $platform->info >> for details. |
1566 |
|
|
|
1567 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetMemObjectInfo.html> |
1568 |
|
|
|
1569 |
root |
1.21 |
=for gengetinfo begin mem |
1570 |
|
|
|
1571 |
|
|
=item $mem_object_type = $mem->type |
1572 |
|
|
|
1573 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_TYPE> and returns the result. |
1574 |
root |
1.21 |
|
1575 |
|
|
=item $mem_flags = $mem->flags |
1576 |
|
|
|
1577 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_FLAGS> and returns the result. |
1578 |
root |
1.21 |
|
1579 |
|
|
=item $int = $mem->size |
1580 |
|
|
|
1581 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_SIZE> and returns the result. |
1582 |
root |
1.21 |
|
1583 |
|
|
=item $ptr_value = $mem->host_ptr |
1584 |
|
|
|
1585 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_HOST_PTR> and returns the result. |
1586 |
root |
1.21 |
|
1587 |
|
|
=item $uint = $mem->map_count |
1588 |
|
|
|
1589 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_MAP_COUNT> and returns the result. |
1590 |
root |
1.21 |
|
1591 |
|
|
=item $uint = $mem->reference_count |
1592 |
|
|
|
1593 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_REFERENCE_COUNT> and returns the result. |
1594 |
root |
1.21 |
|
1595 |
|
|
=item $ctx = $mem->context |
1596 |
|
|
|
1597 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_CONTEXT> and returns the result. |
1598 |
root |
1.21 |
|
1599 |
|
|
=item $mem = $mem->associated_memobject |
1600 |
|
|
|
1601 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_ASSOCIATED_MEMOBJECT> and returns the result. |
1602 |
root |
1.21 |
|
1603 |
|
|
=item $int = $mem->offset |
1604 |
|
|
|
1605 |
root |
1.71 |
Calls C<clGetMemObjectInfo> with C<OpenCL::MEM_OFFSET> and returns the result. |
1606 |
root |
1.21 |
|
1607 |
|
|
=for gengetinfo end mem |
1608 |
|
|
|
1609 |
root |
1.34 |
=item ($type, $name) = $mem->gl_object_info |
1610 |
|
|
|
1611 |
|
|
Returns the OpenGL object type (e.g. OpenCL::GL_OBJECT_TEXTURE2D) and the |
1612 |
|
|
object "name" (e.g. the texture name) used to create this memory object. |
1613 |
|
|
|
1614 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetGLObjectInfo.html> |
1615 |
|
|
|
1616 |
root |
1.5 |
=back |
1617 |
|
|
|
1618 |
root |
1.27 |
=head2 THE OpenCL::Buffer CLASS |
1619 |
|
|
|
1620 |
|
|
This is a subclass of OpenCL::Memory, and the superclass of |
1621 |
|
|
OpenCL::BufferObj. Its purpose is simply to distinguish between buffers |
1622 |
|
|
and sub-buffers. |
1623 |
|
|
|
1624 |
|
|
=head2 THE OpenCL::BufferObj CLASS |
1625 |
|
|
|
1626 |
|
|
This is a subclass of OpenCL::Buffer and thus OpenCL::Memory. It exists |
1627 |
|
|
because one cna create sub buffers of OpenLC::BufferObj objects, but not |
1628 |
|
|
sub buffers from these sub buffers. |
1629 |
|
|
|
1630 |
|
|
=over 4 |
1631 |
|
|
|
1632 |
|
|
=item $subbuf = $buf_obj->sub_buffer_region ($flags, $origin, $size) |
1633 |
|
|
|
1634 |
|
|
Creates an OpenCL::Buffer objects from this buffer and returns it. The |
1635 |
|
|
C<buffer_create_type> is assumed to be C<CL_BUFFER_CREATE_TYPE_REGION>. |
1636 |
|
|
|
1637 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateSubBuffer.html> |
1638 |
|
|
|
1639 |
|
|
=back |
1640 |
|
|
|
1641 |
root |
1.20 |
=head2 THE OpenCL::Image CLASS |
1642 |
|
|
|
1643 |
root |
1.47 |
This is the superclass of all image objects - OpenCL::Image1D, |
1644 |
|
|
OpenCL::Image1DArray, OpenCL::Image1DBuffer, OpenCL::Image2D, |
1645 |
|
|
OpenCL::Image2DArray and OpenCL::Image3D. |
1646 |
root |
1.20 |
|
1647 |
|
|
=over 4 |
1648 |
|
|
|
1649 |
root |
1.53 |
=item $packed_value = $image->image_info ($name) |
1650 |
root |
1.20 |
|
1651 |
|
|
See C<< $platform->info >> for details. |
1652 |
|
|
|
1653 |
|
|
The reason this method is not called C<info> is that there already is an |
1654 |
|
|
C<< ->info >> method inherited from C<OpenCL::Memory>. |
1655 |
|
|
|
1656 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetImageInfo.html> |
1657 |
|
|
|
1658 |
root |
1.53 |
=item ($channel_order, $channel_data_type) = $image->format |
1659 |
|
|
|
1660 |
|
|
Returns the channel order and type used to create the image by calling |
1661 |
|
|
C<clGetImageInfo> with C<CL_IMAGE_FORMAT>. |
1662 |
|
|
|
1663 |
root |
1.21 |
=for gengetinfo begin image |
1664 |
|
|
|
1665 |
|
|
=item $int = $image->element_size |
1666 |
|
|
|
1667 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_ELEMENT_SIZE> and returns the result. |
1668 |
root |
1.21 |
|
1669 |
|
|
=item $int = $image->row_pitch |
1670 |
|
|
|
1671 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_ROW_PITCH> and returns the result. |
1672 |
root |
1.21 |
|
1673 |
|
|
=item $int = $image->slice_pitch |
1674 |
|
|
|
1675 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_SLICE_PITCH> and returns the result. |
1676 |
root |
1.21 |
|
1677 |
|
|
=item $int = $image->width |
1678 |
|
|
|
1679 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_WIDTH> and returns the result. |
1680 |
root |
1.21 |
|
1681 |
|
|
=item $int = $image->height |
1682 |
|
|
|
1683 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_HEIGHT> and returns the result. |
1684 |
root |
1.21 |
|
1685 |
|
|
=item $int = $image->depth |
1686 |
|
|
|
1687 |
root |
1.71 |
Calls C<clGetImageInfo> with C<OpenCL::IMAGE_DEPTH> and returns the result. |
1688 |
root |
1.21 |
|
1689 |
|
|
=for gengetinfo end image |
1690 |
|
|
|
1691 |
root |
1.34 |
=for gengetinfo begin gl_texture |
1692 |
|
|
|
1693 |
|
|
=item $GLenum = $gl_texture->target |
1694 |
|
|
|
1695 |
root |
1.71 |
Calls C<clGetGLTextureInfo> with C<OpenCL::GL_TEXTURE_TARGET> and returns the result. |
1696 |
root |
1.34 |
|
1697 |
|
|
=item $GLint = $gl_texture->gl_mipmap_level |
1698 |
|
|
|
1699 |
root |
1.71 |
Calls C<clGetGLTextureInfo> with C<OpenCL::GL_MIPMAP_LEVEL> and returns the result. |
1700 |
root |
1.34 |
|
1701 |
|
|
=for gengetinfo end gl_texture |
1702 |
|
|
|
1703 |
root |
1.20 |
=back |
1704 |
|
|
|
1705 |
root |
1.5 |
=head2 THE OpenCL::Sampler CLASS |
1706 |
|
|
|
1707 |
|
|
=over 4 |
1708 |
|
|
|
1709 |
|
|
=item $packed_value = $sampler->info ($name) |
1710 |
|
|
|
1711 |
|
|
See C<< $platform->info >> for details. |
1712 |
|
|
|
1713 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetSamplerInfo.html> |
1714 |
|
|
|
1715 |
root |
1.21 |
=for gengetinfo begin sampler |
1716 |
|
|
|
1717 |
|
|
=item $uint = $sampler->reference_count |
1718 |
|
|
|
1719 |
root |
1.71 |
Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_REFERENCE_COUNT> and returns the result. |
1720 |
root |
1.21 |
|
1721 |
|
|
=item $ctx = $sampler->context |
1722 |
|
|
|
1723 |
root |
1.71 |
Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_CONTEXT> and returns the result. |
1724 |
root |
1.21 |
|
1725 |
|
|
=item $addressing_mode = $sampler->normalized_coords |
1726 |
|
|
|
1727 |
root |
1.71 |
Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_NORMALIZED_COORDS> and returns the result. |
1728 |
root |
1.21 |
|
1729 |
|
|
=item $filter_mode = $sampler->addressing_mode |
1730 |
|
|
|
1731 |
root |
1.71 |
Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_ADDRESSING_MODE> and returns the result. |
1732 |
root |
1.21 |
|
1733 |
|
|
=item $boolean = $sampler->filter_mode |
1734 |
|
|
|
1735 |
root |
1.71 |
Calls C<clGetSamplerInfo> with C<OpenCL::SAMPLER_FILTER_MODE> and returns the result. |
1736 |
root |
1.21 |
|
1737 |
|
|
=for gengetinfo end sampler |
1738 |
|
|
|
1739 |
root |
1.5 |
=back |
1740 |
|
|
|
1741 |
|
|
=head2 THE OpenCL::Program CLASS |
1742 |
|
|
|
1743 |
|
|
=over 4 |
1744 |
|
|
|
1745 |
root |
1.55 |
=item $program->build (\@devices = undef, $options = "", $cb->($program) = undef) |
1746 |
root |
1.5 |
|
1747 |
root |
1.51 |
Tries to build the program with the given options. See also the |
1748 |
|
|
C<$ctx->build> convenience function. |
1749 |
root |
1.5 |
|
1750 |
root |
1.55 |
If a callback is specified, then it will be called when compilation is |
1751 |
|
|
finished. Note that many OpenCL implementations block your program while |
1752 |
|
|
compiling whether you use a callback or not. See C<build_async> if you |
1753 |
|
|
want to make sure the build is done in the background. |
1754 |
|
|
|
1755 |
root |
1.63 |
Note that some OpenCL implementations act up badly, and don't call the |
1756 |
root |
1.55 |
callback in some error cases (but call it in others). This implementation |
1757 |
|
|
assumes the callback will always be called, and leaks memory if this is |
1758 |
|
|
not so. So best make sure you don't pass in invalid values. |
1759 |
|
|
|
1760 |
root |
1.63 |
Some implementations fail with C<OpenCL::INVALID_BINARY> when the |
1761 |
|
|
compilation state is successful but some later stage fails. |
1762 |
|
|
|
1763 |
root |
1.71 |
options: C<-D name>, C<-D name=definition>, C<-I dir>, |
1764 |
|
|
C<-cl-single-precision-constant>, C<-cl-denorms-are-zero>, |
1765 |
|
|
C<-cl-fp32-correctly-rounded-divide-sqrt>, C<-cl-opt-disable>, |
1766 |
|
|
C<-cl-mad-enable>, C<-cl-no-signed-zeros>, C<-cl-unsafe-math-optimizations>, |
1767 |
|
|
C<-cl-finite-math-only>, C<-cl-fast-relaxed-math>, |
1768 |
|
|
C<-w>, C<-Werror>, C<-cl-std=CL1.1/CL1.2>, C<-cl-kernel-arg-info>, |
1769 |
|
|
C<-create-library>, C<-enable-link-options>. |
1770 |
|
|
|
1771 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clBuildProgram.html> |
1772 |
|
|
|
1773 |
root |
1.55 |
=item $program->build_async (\@devices = undef, $options = "", $cb->($program) = undef) |
1774 |
|
|
|
1775 |
|
|
Similar to C<< ->build >>, except it starts a thread, and never fails (you |
1776 |
|
|
need to check the compilation status form the callback, or by polling). |
1777 |
|
|
|
1778 |
root |
1.72 |
build_status: OpenCL::BUILD_SUCCESS, OpenCL::BUILD_NONE, |
1779 |
|
|
OpenCL::BUILD_ERROR, OpenCL::BUILD_IN_PROGRESS. |
1780 |
|
|
|
1781 |
root |
1.5 |
=item $packed_value = $program->build_info ($device, $name) |
1782 |
|
|
|
1783 |
|
|
Similar to C<< $platform->info >>, but returns build info for a previous |
1784 |
|
|
build attempt for the given device. |
1785 |
|
|
|
1786 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetBuildInfo.html> |
1787 |
|
|
|
1788 |
|
|
=item $kernel = $program->kernel ($function_name) |
1789 |
|
|
|
1790 |
|
|
Creates an OpenCL::Kernel object out of the named C<__kernel> function in |
1791 |
|
|
the program. |
1792 |
|
|
|
1793 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernel.html> |
1794 |
|
|
|
1795 |
root |
1.50 |
=item @kernels = $program->kernels_in_program |
1796 |
|
|
|
1797 |
|
|
Returns all kernels successfully compiled for all devices in program. |
1798 |
|
|
|
1799 |
|
|
http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clCreateKernelsInProgram.html |
1800 |
|
|
|
1801 |
root |
1.21 |
=for gengetinfo begin program_build |
1802 |
|
|
|
1803 |
|
|
=item $build_status = $program->build_status ($device) |
1804 |
|
|
|
1805 |
root |
1.71 |
Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_STATUS> and returns the result. |
1806 |
|
|
|
1807 |
root |
1.21 |
=item $string = $program->build_options ($device) |
1808 |
|
|
|
1809 |
root |
1.71 |
Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_OPTIONS> and returns the result. |
1810 |
root |
1.21 |
|
1811 |
|
|
=item $string = $program->build_log ($device) |
1812 |
|
|
|
1813 |
root |
1.71 |
Calls C<clGetProgramBuildInfo> with C<OpenCL::PROGRAM_BUILD_LOG> and returns the result. |
1814 |
root |
1.21 |
|
1815 |
|
|
=for gengetinfo end program_build |
1816 |
|
|
|
1817 |
|
|
=item $packed_value = $program->info ($name) |
1818 |
|
|
|
1819 |
|
|
See C<< $platform->info >> for details. |
1820 |
|
|
|
1821 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html> |
1822 |
|
|
|
1823 |
|
|
=for gengetinfo begin program |
1824 |
|
|
|
1825 |
|
|
=item $uint = $program->reference_count |
1826 |
|
|
|
1827 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_REFERENCE_COUNT> and returns the result. |
1828 |
root |
1.21 |
|
1829 |
|
|
=item $ctx = $program->context |
1830 |
|
|
|
1831 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_CONTEXT> and returns the result. |
1832 |
root |
1.21 |
|
1833 |
|
|
=item $uint = $program->num_devices |
1834 |
|
|
|
1835 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_NUM_DEVICES> and returns the result. |
1836 |
root |
1.21 |
|
1837 |
|
|
=item @devices = $program->devices |
1838 |
|
|
|
1839 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_DEVICES> and returns the result. |
1840 |
root |
1.21 |
|
1841 |
|
|
=item $string = $program->source |
1842 |
|
|
|
1843 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_SOURCE> and returns the result. |
1844 |
root |
1.21 |
|
1845 |
|
|
=item @ints = $program->binary_sizes |
1846 |
|
|
|
1847 |
root |
1.71 |
Calls C<clGetProgramInfo> with C<OpenCL::PROGRAM_BINARY_SIZES> and returns the result. |
1848 |
root |
1.21 |
|
1849 |
|
|
=for gengetinfo end program |
1850 |
|
|
|
1851 |
root |
1.23 |
=item @blobs = $program->binaries |
1852 |
|
|
|
1853 |
|
|
Returns a string for the compiled binary for every device associated with |
1854 |
|
|
the program, empty strings indicate missing programs, and an empty result |
1855 |
|
|
means no program binaries are available. |
1856 |
|
|
|
1857 |
|
|
These "binaries" are often, in fact, informative low-level assembly |
1858 |
|
|
sources. |
1859 |
|
|
|
1860 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProgramInfo.html> |
1861 |
|
|
|
1862 |
root |
1.5 |
=back |
1863 |
|
|
|
1864 |
|
|
=head2 THE OpenCL::Kernel CLASS |
1865 |
|
|
|
1866 |
|
|
=over 4 |
1867 |
|
|
|
1868 |
|
|
=item $packed_value = $kernel->info ($name) |
1869 |
|
|
|
1870 |
|
|
See C<< $platform->info >> for details. |
1871 |
|
|
|
1872 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelInfo.html> |
1873 |
|
|
|
1874 |
root |
1.21 |
=for gengetinfo begin kernel |
1875 |
|
|
|
1876 |
|
|
=item $string = $kernel->function_name |
1877 |
|
|
|
1878 |
root |
1.71 |
Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_FUNCTION_NAME> and returns the result. |
1879 |
root |
1.21 |
|
1880 |
|
|
=item $uint = $kernel->num_args |
1881 |
|
|
|
1882 |
root |
1.71 |
Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_NUM_ARGS> and returns the result. |
1883 |
root |
1.21 |
|
1884 |
|
|
=item $uint = $kernel->reference_count |
1885 |
|
|
|
1886 |
root |
1.71 |
Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_REFERENCE_COUNT> and returns the result. |
1887 |
root |
1.21 |
|
1888 |
|
|
=item $ctx = $kernel->context |
1889 |
|
|
|
1890 |
root |
1.71 |
Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_CONTEXT> and returns the result. |
1891 |
root |
1.21 |
|
1892 |
|
|
=item $program = $kernel->program |
1893 |
|
|
|
1894 |
root |
1.71 |
Calls C<clGetKernelInfo> with C<OpenCL::KERNEL_PROGRAM> and returns the result. |
1895 |
root |
1.21 |
|
1896 |
|
|
=for gengetinfo end kernel |
1897 |
|
|
|
1898 |
root |
1.20 |
=item $packed_value = $kernel->work_group_info ($device, $name) |
1899 |
|
|
|
1900 |
|
|
See C<< $platform->info >> for details. |
1901 |
|
|
|
1902 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetKernelWorkGroupInfo.html> |
1903 |
|
|
|
1904 |
root |
1.21 |
=for gengetinfo begin kernel_work_group |
1905 |
|
|
|
1906 |
|
|
=item $int = $kernel->work_group_size ($device) |
1907 |
|
|
|
1908 |
root |
1.71 |
Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_WORK_GROUP_SIZE> and returns the result. |
1909 |
root |
1.21 |
|
1910 |
|
|
=item @ints = $kernel->compile_work_group_size ($device) |
1911 |
|
|
|
1912 |
root |
1.71 |
Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_COMPILE_WORK_GROUP_SIZE> and returns the result. |
1913 |
root |
1.21 |
|
1914 |
|
|
=item $ulong = $kernel->local_mem_size ($device) |
1915 |
|
|
|
1916 |
root |
1.71 |
Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_LOCAL_MEM_SIZE> and returns the result. |
1917 |
root |
1.21 |
|
1918 |
|
|
=item $int = $kernel->preferred_work_group_size_multiple ($device) |
1919 |
|
|
|
1920 |
root |
1.71 |
Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_PREFERRED_WORK_GROUP_SIZE_MULTIPLE> and returns the result. |
1921 |
root |
1.21 |
|
1922 |
|
|
=item $ulong = $kernel->private_mem_size ($device) |
1923 |
|
|
|
1924 |
root |
1.71 |
Calls C<clGetKernelWorkGroupInfo> with C<OpenCL::KERNEL_PRIVATE_MEM_SIZE> and returns the result. |
1925 |
root |
1.21 |
|
1926 |
|
|
=for gengetinfo end kernel_work_group |
1927 |
|
|
|
1928 |
root |
1.73 |
=item $packed_value = $kernel->arg_info ($idx, $name) |
1929 |
|
|
|
1930 |
|
|
See C<< $platform->info >> for details. |
1931 |
|
|
|
1932 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clGetKernelArgInfo.html> |
1933 |
|
|
|
1934 |
|
|
=for gengetinfo begin kernel_arg |
1935 |
|
|
|
1936 |
|
|
=item $kernel_arg_address_qualifier = $kernel->arg_address_qualifier ($idx) |
1937 |
|
|
|
1938 |
|
|
Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_ADDRESS_QUALIFIER> and returns the result. |
1939 |
|
|
|
1940 |
|
|
=item $kernel_arg_access_qualifier = $kernel->arg_access_qualifier ($idx) |
1941 |
|
|
|
1942 |
|
|
Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_ACCESS_QUALIFIER> and returns the result. |
1943 |
|
|
|
1944 |
|
|
=item $string = $kernel->arg_type_name ($idx) |
1945 |
|
|
|
1946 |
|
|
Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_TYPE_NAME> and returns the result. |
1947 |
|
|
|
1948 |
|
|
=item $kernel_arg_type_qualifier = $kernel->arg_type_qualifier ($idx) |
1949 |
|
|
|
1950 |
|
|
Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_TYPE_QUALIFIER> and returns the result. |
1951 |
|
|
|
1952 |
|
|
=item $string = $kernel->arg_name ($idx) |
1953 |
|
|
|
1954 |
|
|
Calls C<clGetKernelArgInfo> with C<OpenCL::KERNEL_ARG_NAME> and returns the result. |
1955 |
root |
1.72 |
|
1956 |
root |
1.73 |
=for gengetinfo end kernel_arg |
1957 |
root |
1.72 |
|
1958 |
root |
1.60 |
=item $kernel->setf ($format, ...) |
1959 |
|
|
|
1960 |
|
|
Sets the arguments of a kernel. Since OpenCL 1.1 doesn't have a generic |
1961 |
|
|
way to set arguments (and with OpenCL 1.2 it might be rather slow), you |
1962 |
|
|
need to specify a format argument, much as with C<printf>, to tell OpenCL |
1963 |
|
|
what type of argument it is. |
1964 |
|
|
|
1965 |
|
|
The format arguments are single letters: |
1966 |
|
|
|
1967 |
|
|
c char |
1968 |
|
|
C unsigned char |
1969 |
|
|
s short |
1970 |
|
|
S unsigned short |
1971 |
|
|
i int |
1972 |
|
|
I unsigned int |
1973 |
|
|
l long |
1974 |
|
|
L unsigned long |
1975 |
|
|
|
1976 |
|
|
h half float (0..65535) |
1977 |
|
|
f float |
1978 |
|
|
d double |
1979 |
|
|
|
1980 |
|
|
z local (octet size) |
1981 |
|
|
|
1982 |
|
|
m memory object (buffer or image) |
1983 |
|
|
a sampler |
1984 |
|
|
e event |
1985 |
|
|
|
1986 |
|
|
Space characters in the format string are ignored. |
1987 |
|
|
|
1988 |
|
|
Example: set the arguments for a kernel that expects an int, two floats, a buffer and an image. |
1989 |
|
|
|
1990 |
|
|
$kernel->setf ("i ff mm", 5, 0.5, 3, $buffer, $image); |
1991 |
|
|
|
1992 |
root |
1.58 |
=item $kernel->set_TYPE ($index, $value) |
1993 |
root |
1.5 |
|
1994 |
root |
1.58 |
=item $kernel->set_char ($index, $value) |
1995 |
root |
1.5 |
|
1996 |
root |
1.58 |
=item $kernel->set_uchar ($index, $value) |
1997 |
|
|
|
1998 |
|
|
=item $kernel->set_short ($index, $value) |
1999 |
|
|
|
2000 |
|
|
=item $kernel->set_ushort ($index, $value) |
2001 |
|
|
|
2002 |
|
|
=item $kernel->set_int ($index, $value) |
2003 |
|
|
|
2004 |
|
|
=item $kernel->set_uint ($index, $value) |
2005 |
|
|
|
2006 |
|
|
=item $kernel->set_long ($index, $value) |
2007 |
|
|
|
2008 |
|
|
=item $kernel->set_ulong ($index, $value) |
2009 |
|
|
|
2010 |
|
|
=item $kernel->set_half ($index, $value) |
2011 |
|
|
|
2012 |
|
|
=item $kernel->set_float ($index, $value) |
2013 |
|
|
|
2014 |
|
|
=item $kernel->set_double ($index, $value) |
2015 |
|
|
|
2016 |
|
|
=item $kernel->set_memory ($index, $value) |
2017 |
|
|
|
2018 |
|
|
=item $kernel->set_buffer ($index, $value) |
2019 |
|
|
|
2020 |
|
|
=item $kernel->set_image ($index, $value) |
2021 |
|
|
|
2022 |
|
|
=item $kernel->set_sampler ($index, $value) |
2023 |
|
|
|
2024 |
|
|
=item $kernel->set_local ($index, $value) |
2025 |
|
|
|
2026 |
|
|
=item $kernel->set_event ($index, $value) |
2027 |
|
|
|
2028 |
|
|
This is a family of methods to set the kernel argument with the number |
2029 |
|
|
C<$index> to the give C<$value>. |
2030 |
root |
1.5 |
|
2031 |
|
|
Chars and integers (including the half type) are specified as integers, |
2032 |
root |
1.58 |
float and double as floating point values, memory/buffer/image must be |
2033 |
|
|
an object of that type or C<undef>, local-memory arguments are set by |
2034 |
|
|
specifying the size, and sampler and event must be objects of that type. |
2035 |
|
|
|
2036 |
|
|
Note that C<set_memory> works for all memory objects (all types of buffers |
2037 |
|
|
and images) - the main purpose of the more specific C<set_TYPE> functions |
2038 |
|
|
is type checking. |
2039 |
root |
1.5 |
|
2040 |
root |
1.50 |
Setting an argument for a kernel does NOT keep a reference to the object - |
2041 |
|
|
for example, if you set an argument to some image object, free the image, |
2042 |
|
|
and call the kernel, you will run into undefined behaviour. |
2043 |
|
|
|
2044 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetKernelArg.html> |
2045 |
|
|
|
2046 |
|
|
=back |
2047 |
|
|
|
2048 |
|
|
=head2 THE OpenCL::Event CLASS |
2049 |
|
|
|
2050 |
|
|
This is the superclass for all event objects (including OpenCL::UserEvent |
2051 |
|
|
objects). |
2052 |
|
|
|
2053 |
|
|
=over 4 |
2054 |
|
|
|
2055 |
root |
1.21 |
=item $ev->wait |
2056 |
|
|
|
2057 |
|
|
Waits for the event to complete. |
2058 |
|
|
|
2059 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clWaitForEvents.html> |
2060 |
|
|
|
2061 |
root |
1.55 |
=item $ev->cb ($exec_callback_type, $callback->($event, $event_command_exec_status)) |
2062 |
|
|
|
2063 |
|
|
Adds a callback to the callback stack for the given event type. There is |
2064 |
|
|
no way to remove a callback again. |
2065 |
|
|
|
2066 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetEventCallback.html> |
2067 |
|
|
|
2068 |
root |
1.5 |
=item $packed_value = $ev->info ($name) |
2069 |
|
|
|
2070 |
|
|
See C<< $platform->info >> for details. |
2071 |
|
|
|
2072 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetEventInfo.html> |
2073 |
|
|
|
2074 |
root |
1.21 |
=for gengetinfo begin event |
2075 |
|
|
|
2076 |
|
|
=item $queue = $event->command_queue |
2077 |
|
|
|
2078 |
root |
1.71 |
Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_QUEUE> and returns the result. |
2079 |
root |
1.21 |
|
2080 |
|
|
=item $command_type = $event->command_type |
2081 |
|
|
|
2082 |
root |
1.71 |
Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_TYPE> and returns the result. |
2083 |
root |
1.21 |
|
2084 |
|
|
=item $uint = $event->reference_count |
2085 |
|
|
|
2086 |
root |
1.71 |
Calls C<clGetEventInfo> with C<OpenCL::EVENT_REFERENCE_COUNT> and returns the result. |
2087 |
root |
1.21 |
|
2088 |
|
|
=item $uint = $event->command_execution_status |
2089 |
|
|
|
2090 |
root |
1.71 |
Calls C<clGetEventInfo> with C<OpenCL::EVENT_COMMAND_EXECUTION_STATUS> and returns the result. |
2091 |
root |
1.21 |
|
2092 |
|
|
=item $ctx = $event->context |
2093 |
|
|
|
2094 |
root |
1.71 |
Calls C<clGetEventInfo> with C<OpenCL::EVENT_CONTEXT> and returns the result. |
2095 |
root |
1.21 |
|
2096 |
|
|
=for gengetinfo end event |
2097 |
|
|
|
2098 |
root |
1.20 |
=item $packed_value = $ev->profiling_info ($name) |
2099 |
|
|
|
2100 |
|
|
See C<< $platform->info >> for details. |
2101 |
|
|
|
2102 |
|
|
The reason this method is not called C<info> is that there already is an |
2103 |
|
|
C<< ->info >> method. |
2104 |
|
|
|
2105 |
|
|
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clGetProfilingInfo.html> |
2106 |
|
|
|
2107 |
root |
1.21 |
=for gengetinfo begin profiling |
2108 |
|
|
|
2109 |
|
|
=item $ulong = $event->profiling_command_queued |
2110 |
|
|
|
2111 |
root |
1.71 |
Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_QUEUED> and returns the result. |
2112 |
root |
1.21 |
|
2113 |
|
|
=item $ulong = $event->profiling_command_submit |
2114 |
|
|
|
2115 |
root |
1.71 |
Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_SUBMIT> and returns the result. |
2116 |
root |
1.21 |
|
2117 |
|
|
=item $ulong = $event->profiling_command_start |
2118 |
|
|
|
2119 |
root |
1.71 |
Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_START> and returns the result. |
2120 |
root |
1.21 |
|
2121 |
|
|
=item $ulong = $event->profiling_command_end |
2122 |
root |
1.5 |
|
2123 |
root |
1.71 |
Calls C<clGetEventProfilingInfo> with C<OpenCL::PROFILING_COMMAND_END> and returns the result. |
2124 |
root |
1.5 |
|
2125 |
root |
1.21 |
=for gengetinfo end profiling |
2126 |
root |
1.5 |
|
2127 |
|
|
=back |
2128 |
|
|
|
2129 |
|
|
=head2 THE OpenCL::UserEvent CLASS |
2130 |
|
|
|
2131 |
|
|
This is a subclass of OpenCL::Event. |
2132 |
root |
1.4 |
|
2133 |
root |
1.1 |
=over 4 |
2134 |
|
|
|
2135 |
root |
1.5 |
=item $ev->set_status ($execution_status) |
2136 |
|
|
|
2137 |
root |
1.55 |
Sets the execution status of the user event. Can only be called once, |
2138 |
|
|
either with OpenCL::COMPLETE or a negative number as status. |
2139 |
|
|
|
2140 |
root |
1.71 |
execution_status: OpenCL::COMPLETE or a negative integer. |
2141 |
|
|
|
2142 |
root |
1.5 |
L<http://www.khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/clSetUserEventStatus.html> |
2143 |
|
|
|
2144 |
|
|
=back |
2145 |
|
|
|
2146 |
root |
1.66 |
=head2 THE OpenCL::Mapped CLASS |
2147 |
|
|
|
2148 |
|
|
This class represents objects mapped into host memory. They are |
2149 |
|
|
represented by a blessed string scalar. The string data is the mapped |
2150 |
|
|
memory area, that is, if you read or write it, then the mapped object is |
2151 |
|
|
accessed directly. |
2152 |
|
|
|
2153 |
|
|
You must only ever use operations that modify the string in-place - for |
2154 |
|
|
example, a C<substr> that doesn't change the length, or maybe a regex that |
2155 |
|
|
doesn't change the length. Any other operation might cause the data to be |
2156 |
|
|
copied. |
2157 |
|
|
|
2158 |
|
|
When the object is destroyed it will enqueue an implicit unmap operation |
2159 |
|
|
on the queue that was used to create it. |
2160 |
|
|
|
2161 |
root |
1.67 |
Keep in mind that you I<need> to unmap (or destroy) mapped objects before |
2162 |
|
|
OpenCL sees the changes, even if some implementations don't need this |
2163 |
|
|
sometimes. |
2164 |
|
|
|
2165 |
root |
1.66 |
Example, replace the first two floats in the mapped buffer by 1 and 2. |
2166 |
|
|
|
2167 |
|
|
my $mapped = $queue->map_buffer ($buf, ... |
2168 |
|
|
$mapped->event->wait; # make sure it's there |
2169 |
|
|
|
2170 |
|
|
# now replace first 8 bytes by new data, which is exactly 8 bytes long |
2171 |
|
|
# we blindly assume device endianness to equal host endianness |
2172 |
|
|
# (and of course, we assume iee 754 single precision floats :) |
2173 |
|
|
substr $$mapped, 0, 8, pack "f*", 1, 2; |
2174 |
|
|
|
2175 |
|
|
=over 4 |
2176 |
|
|
|
2177 |
root |
1.67 |
=item $ev = $mapped->unmap ($wait_events...) |
2178 |
|
|
|
2179 |
|
|
Unmaps the mapped memory object, using the queue originally used to create |
2180 |
|
|
it, quite similarly to C<< $queue->unmap ($mapped, ...) >>. |
2181 |
|
|
|
2182 |
root |
1.66 |
=item $bool = $mapped->mapped |
2183 |
|
|
|
2184 |
|
|
Returns whether the object is still mapped - true before an C<unmap> is |
2185 |
|
|
enqueued, false afterwards. |
2186 |
|
|
|
2187 |
|
|
=item $ev = $mapped->event |
2188 |
|
|
|
2189 |
|
|
Return the event object associated with the mapped object. Initially, this |
2190 |
|
|
will be the event object created when mapping the object, and after an |
2191 |
|
|
unmap, this will be the event object that the unmap operation created. |
2192 |
|
|
|
2193 |
|
|
=item $mapped->wait |
2194 |
|
|
|
2195 |
|
|
Same as C<< $mapped->event->wait >> - makes sure no operations on this |
2196 |
|
|
mapped object are outstanding. |
2197 |
|
|
|
2198 |
|
|
=item $bytes = $mapped->size |
2199 |
|
|
|
2200 |
|
|
Returns the size of the mapped area, in bytes. Same as C<length $$mapped>. |
2201 |
|
|
|
2202 |
|
|
=item $ptr = $mapped->ptr |
2203 |
|
|
|
2204 |
root |
1.68 |
Returns the raw memory address of the mapped area. |
2205 |
root |
1.66 |
|
2206 |
root |
1.67 |
=item $mapped->set ($offset, $data) |
2207 |
|
|
|
2208 |
|
|
Replaces the data at the given C<$offset> in the memory area by the new |
2209 |
root |
1.68 |
C<$data>. This method is safer than direct manipulation of C<$mapped> |
2210 |
|
|
because it does bounds-checking, but also slower. |
2211 |
root |
1.67 |
|
2212 |
|
|
=item $data = $mapped->get ($offset, $length) |
2213 |
|
|
|
2214 |
|
|
Returns (without copying) a scalar representing the data at the given |
2215 |
|
|
C<$offset> and C<$length> in the mapped memory area. This is the same as |
2216 |
root |
1.68 |
the following substr, except much slower; |
2217 |
root |
1.67 |
|
2218 |
|
|
$data = substr $$mapped, $offset, $length |
2219 |
|
|
|
2220 |
|
|
=cut |
2221 |
|
|
|
2222 |
root |
1.68 |
sub OpenCL::Mapped::get { |
2223 |
root |
1.67 |
substr ${$_[0]}, $_[1], $_[2] |
2224 |
|
|
} |
2225 |
|
|
|
2226 |
root |
1.66 |
=back |
2227 |
|
|
|
2228 |
|
|
=head2 THE OpenCL::MappedBuffer CLASS |
2229 |
|
|
|
2230 |
|
|
This is a subclass of OpenCL::Mapped, representing mapped buffers. |
2231 |
|
|
|
2232 |
|
|
=head2 THE OpenCL::MappedImage CLASS |
2233 |
|
|
|
2234 |
|
|
This is a subclass of OpenCL::Mapped, representing mapped images. |
2235 |
|
|
|
2236 |
|
|
=over 4 |
2237 |
|
|
|
2238 |
root |
1.67 |
=item $bytes = $mapped->row_pitch |
2239 |
|
|
|
2240 |
|
|
=item $bytes = $mapped->slice_pitch |
2241 |
|
|
|
2242 |
|
|
Return the row or slice pitch of the image that has been mapped. |
2243 |
|
|
|
2244 |
root |
1.66 |
=back |
2245 |
|
|
|
2246 |
|
|
|
2247 |
root |
1.1 |
=cut |
2248 |
|
|
|
2249 |
|
|
1; |
2250 |
|
|
|
2251 |
|
|
=head1 AUTHOR |
2252 |
|
|
|
2253 |
|
|
Marc Lehmann <schmorp@schmorp.de> |
2254 |
|
|
http://home.schmorp.de/ |
2255 |
|
|
|
2256 |
|
|
=cut |
2257 |
|
|
|