Hello, First of all, you can loo-kup long-term latency plot for PREEMPT_RT on different systems. I doubt Raspberry Pi 5 differs from Celeron/Atom, in my experience and from looking at the plots:, but YMMD. https://www.osadl.org/Latency-plot-of-system-in-rack-2-slot.qa-latencyplot-r2s7.0.html?shadow=0 https://www.osadl.org/Long-term-latency-plot-of-system-in-rack.qa-3d-latencyplot-rds3.0.html?shadow=0 You could consider using isolated CPU cores, which should give you much lower latencies, if you have full control over the application logic. Regards, Leon. -- Leon Woestenberg leon@xxxxxxxxxxxxxx T: +31 40 711 42 76 M: +31 6 472 30 372 Sidebranch Embedded Systems Eindhoven, The Netherlands http://www.sidebranch.com On Fri, Dec 20, 2024 at 4:58 PM Florian Paul Schmidt <mista.tapas@xxxxxxx> wrote: > > Hi John! Sorry, forgot to hit "reply all". So here it goes again :) > With a little clarification as well.. > > On 20/12/2024 15:05, John Ogness wrote: > > There are various kernel configurations that affect latency. But I am > > curious... Why do you think a maximum latency of 80us on an $80 board is > > large? > > Simply because 80 us are an "eternity" 😉 That's like 192000 cycles at > 2.4Ghz. And also because other slow systems (like that Celeron j4125 > board I have) behave much better even at similar clock speeds (EDIT: > IIRC). I don't think the price point has anything to do with it really. > > EDIT: Don't get me wrong though. 80 us are OK for my use case (audio > processing) - I just wonder if it might be even better. > > > What maximum latencies are you expecting? And why do you think it > > is not "under control"? > > I would _wish_ for latencies in the low tens of microseconds 🙂 I do > not expect anything. > > Kind regards, > FPS > >
