Re: [Jack-Devel] jackd2 and high sample rates

On Tuesday 11 February 2014 21:45:28 Matt Flax wrote:
> > It is very likely that you are the first one who ever tried such high
> > rates with JACK. So nobody can tell you for certain whether there will
> > be any problems. However as far as I can see it right now, as long as
> > your period duration is reasonable high enough (i.e. > 1 ms), then I
> > currently don't see a reason why it should not work reliably with JACK2.
> > 
> > Interesting test case though.
> 
> Sure, Is there a dummy driver example I can use in Jack2 ?

You mean an arbitrary small driver as template for writing this new driver?

> The IIO subsystem uses a mmap data acquisition approach. In that
> respect, blocking ioctl calls to dequeue new data from the kernel
> subsystem are used in the 'read' methods of the class.

Mmmm, the mechanism behind page faults involved in mmaps could cause a 
performance penalty. But since you said the driver is already running fine 
without RT issues, it might be OK.

CU
Christian

1392153408.4038_0.ltw:2,a <201402112117.53314.schoenebeck at crudebyte dot com>