EBo wrote: > This is only a little OT, but... If the modern kernels work out of the box on > the Cortex-A8, then maybe real-time preemptive would work until RTAI has a > port ready. IIRC, RT preemtion was running ~10us compared to ~4us on the same > hardware. Those tests though are a few years old, so it might be better now. > It would also be useful for RT applications which do not have as tight a > tolerance. > We also need to be able to use floating point in the RT modules, which some RT patches do not allow. I know rt-linux and RTAI do allow this on X86. Since embedded ARM processors have not had FP hardware until VERY recently, the rt patches didn't have to deal with this. And, of course, saving the full FP state varies with each architecture, so the code has to be specific to the CPU. 10 us would be no problem when EMC is used with a servo interface, but it needs to be pretty good, 50 - 100 us jitter could be acceptable for a 1 KHz servo cycle, but that is the limit, and it has to be TOTALLY reliable, NEVER exceeding some reasonable threshold. As that report from Jeff via Stephen indicates, the preemptive rt didn't make the cut there.
Jon ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
