The new mali seems to have a hardware sheduler if I'm not misunderstanding.
It would be very neat to be able to not only have compute sources which are hardware accelerated in OpenCL but also schedulers which can be hardware accelerated.
(And to everyone, please let users (and hardware accelerated schedulers) decide about where to execute the calculations/what parts of a device are taxed)
Got the idea from an article:
Unlike OpenCL, where command queues are statically associated to devices, with RenderScript compute it is the runtime that decides where the job will execute.
Wanna have that in OpenCl two, but reading a little further:
Very interesting thus the idea of hardware accelerated schedulers....task dependency coordination is entirely designed into the hardware job manager unit of Mali-T604