On 8/11/2013 5:00 AM, Schooner wrote: > Sebs kernel is ahead of xenomai on this machine, with rt-preempt > notionally just behind that. > In straight latency tests rt-preempt was not far off, but I am running > it full time currently and have had some realtime errors when running a > sim, > which is a bit alarming given there is no base thread.
Both xenomai and rtai give much more consistent results across hardware and do not rely on the linux kernel code to 'behave'. rt-preempt relies on the kernel code (and any hardware drivers) to be cleanly written for high-performance SMP, and latency can be impacted by any poorly written code. That's why this works so much better on the x86 (where a lot of work has focused on good SMP performance) than on other architectures. The good thing about rt-preempt is it is being mainlined, and on x86 the latencies are easily good enough to run a hardware assisted machine w/o a base thread, and on some hardware even a base thread. I did my first LinuxCNC 3D print with rt-preempt and software stepping, but I had to tweak my bios and kernel command line settings to get good latency performance. -- Charles Steinkuehler [email protected]
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------------ Get 100% visibility into Java/.NET code with AppDynamics Lite! It's a free troubleshooting tool designed for production. Get down to code-level detail for bottlenecks, with <2% overhead. Download for free and get started troubleshooting in minutes. http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
