enabling CUDA support

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi all,

I've following bugreport from a BOINC user:
https://bugzilla.redhat.com/show_bug.cgi?id=487981

Basically, CUDA is a Nvidia technology which enables the GPU to be used for various complex scientific computations (which are then even faster than on CPU).

I was about to close the bug as WONTFIX as the whole CUDA is not open source, but then I found out that BOINC (which recently added support for CUDA applications) needs only the single libcudart.so library and that this prebuilt library coming from Nvidia has been already included in the source tarball (but not packaged as far).

Now I have a question: as it principally enables the hardware to be controlled by some end-user applications, would it be possible to ship the libcudart.so library in a subpackage as "Redistributable, no modification permitted" like a firmware?

Actually this is what the Nvidia's EULA is saying about the Linux part of CUDA:
http://developer.download.nvidia.com/compute/cuda/2_1/toolkit/CUDA_Toolkit_EULA_081215.pdf
(see section 2.1.3)

Although I guess we can't do it in this way (I'm afraid that same arguments could then be used for e.g. all closed-source modules), I rather ask before definitely closing the bugreport.

Regards,
Milos

_______________________________________________
Fedora-legal-list mailing list
Fedora-legal-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-legal-list

[Index of Archives]     [Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [Gnome Users]     [KDE Users]

  Powered by Linux