@Ionut, Lukas: I support this proposal too and I'd agree that only the libOpenCL.so (the ICD loader) needs to be split off nvidia-utils. Otherwise it could just stay intact as far as I can tell... Also, it would probably be wise to settle for dependency names once for all, by dependency names I mean what the pkgs provide - right now, all OpenCL pkgs provide both 'opencl' and 'libcl'. So following your proposal, 'libcl' could be provided by the loader, while 'opencl' could be provided by opencl implementations. Or something along those lines... @Nicolas: That looks great. If you released that as opensource, it'd probably be a better choice over the clinfo by Amd. ~kralyk Dne 7. července 2011 18:58 Lukáš Jirkovský <l.jirkovsky@gmail.com> napsal(a):
On 7 July 2011 18:34, Ionut Biru <ibiru@archlinux.org> wrote:
On 07/07/2011 05:46 PM, Nicolas Bigaouette wrote:
<snip>
OKKKKKKK after talking with Lukas i now have an idea about what you guys talking about.
Right now, what can i do is to take care about nvidia's opencl implementation and loader.
What I understand from Nicolas is that we can use any loader, either form nvidia, ati, intel. It will work with any implementation.
My propose is to split opencl loader and implementation out from nvidia-utils like this:
libcl - containing /usr/lib/libOpenCL.so*. This can be easy replaced with an opensource loader once is available in mesa or other that is better. libcl-nvidia - depending on libcl containing /etc/OpenCL/vendors/nvidia.icd and libcuda.so and maybe nvidia module(need confirmation)
for you guys: libcl-intel - depending on libcl libcl-amd - idem
How does it sound for you?
-- Ionuț
I support this proposal.
First I was not convinced that splitting nvidia-utils to three packages is necessary and that two packages would be enough. However OpenCL implementation is independent on OpenGL and it's possible to use multiple devices at once it actually makes sense.
Lukas