Hi, I would like to use the chance at the next Plumbers to discuss the present challenges related to ML accelerators in mainline.
I'm myself more oriented towards edge-oriented deployments, and don't know enough about how these accelerators are being used in the cloud (and maybe desktop?) to tell if there is enough overlap to warrant a common BoF. In any case, these are the topics I would like to discuss, some probably more relevant to the edge than to the cloud or desktop: * What is stopping vendors from mainlining their drivers? * How could we make it easier for them? * Userspace API: how close are we from a common API that we can ask userspace drivers to implement? What can be done to further this goal? * Automated testing: DRM CI can be used, but would be good to have a common test suite to run there. This is probably dependent on a common userspace API. * Other shared userspace infrastructure (compiler, execution, synchronization, virtualization, ...) * Firmware-mediated IP: what should we do about it, if anything? * Any standing issues in DRM infra (GEM, gpu scheduler, DMABuf, etc) that are hurting accel drivers? What do people think, should we have a drivers/accel-wide BoF at Plumbers? If so, what other topics should we have in the agenda? Cheers, Tomeu
