Bug#795262: nvidia-cuda-toolkit: Reboot, after package installation, do not unload the Nouveau drivers

[François Legendre]

Hello,

I understand your point of view. Could you please update the documentation :

This package contains the nvcc compiler etc.

This package contains the nvcc compiler etc. and enables the build of
CUDA applications.  To run CUDA applications on nvidia graphics cards,
please install the nvidia-driver package.

Regards and a lot of thanks for the interaction,
----
François

2015-08-21 12:28 GMT+02:00 Luca Boccassi <luca.boccassi at gmail.com>:
> On 21 August 2015 at 11:15, François Legendre <F.Legendre at u-pec.fr> wrote:
>> Hello,
>>
>> I thought IMHO that the maintainers of nvidia-cuda-toolkit should set
>> a dependency : nvidia-cuda-toolkit depends on nvidia-driver as nouveau
>> drivers are not able to run a CUDA application.
>
> Hello François,
>
> Thanks for sharing your view.
>
> In my opinion that would not be desirable, because as far as I
> understand the toolkit can be used to just build the applications, and
> it should be entirely possible to do so on a headless machine
> designated as a build server with no graphics stack installed. If we
> added a dependency on the driver, this would no longer be possible,
> and the toolkit would then depend on the whole X stack.
>
> I think it's good that there is a clear separation between the "build"
> stack and the "runtime" stack, as it is often the case.
>
> The toolkit dependency chain already has a "suggests" on the
> nvidia-driver package (there might even be a "recommends" along the
> way, can't remember), and on top of that the Nvidia documentation
> makes clear that the driver is required to run the programs.
>
> In my opinion, this should be enough.
>
> Vincent, Graham, Andreas, thoughts?
>
> Kind regards,
> Luca Boccassi