> On 28 Feb 2019, at 20:24, [email protected] wrote: > > Yes, sorry for the hijack. It was just an out-of-the-blue idea which popped > in my head in reaction to Schooner's post. > > And yes, I meant that commit. > > I remember from my LinuxCNC days that CPU hogging, as you aptly named, used > be the thing. (Not so sure now as I don't follow that project closely, only > the forum.) And I quite liked the multicore, multi-instance granularity of > Machinekit. > > However, isolating one(/more) CPU core for the special process of software > step generation doesn't seem so bad.
indeed, as long as you still have cpu’s left for the rest of your system. CPU hogging seems really silly, you buy a PC and only use 50% (or 25% depending on how many cpu’s you bought) > Zultron tested it on J1900 (I think) which is board everybody is composing > odes about (at least on LinuxCNC forum) but it's little old and harder to > get. So it would be interesting to try it on J5005 or J40-something based > boards, which are around 100 euros. hmm, I’m not so sure about that. It’s widely used in industrial setups. I’ve used a couple, and although they are not the fastest. I ran a mesa 5i25+7i76 so there was not a real issue w.r.t. latency. (this was before the cgroups option). I had no bad experiences which ultimately means that it’s a good experience :) https://www.logicsupply.com/catalogsearch/result/?q=J1900 <https://www.logicsupply.com/catalogsearch/result/?q=J1900> I think there are other brands/options available. Personally I prefer to spend some extra money and buy something that (is known to) work. When you run into trouble/unexpected then the hours working on it and billing that to a customer or chucking up those hours for yourself is not worth the “savings”. And generally, if I have something is good enough and works, I tend to stick with it. (Just my take) Bas > > Cern. > > Dne čtvrtek 28. února 2019 19:43:34 UTC+1 Bas de Bruijn napsal(a): > So as not to hijack the thread, a new thread > > On 28 Feb 2019, at 19:20, ce...@ <>tuta.io <http://tuta.io/> wrote: > >> BTW, did someone else (than Zultron) try to use the core isolation feature >> to "turn on" back the step signal generation in software? My electrotrash >> which I use for Machinekitdoes not have cores to spare. > > Yes, you’re probably referring to this? > https://github.com/machinekit/machinekit/pull/1426 > <https://github.com/machinekit/machinekit/pull/1426> > > Although i dont use software step generation. The latency obtained is very > good (6-7 us). > > What was done here is to force the thread to run on designated cpu’s. > Because these cpu’s were isolated during startup (isolcpus= kernel parameter) > there are no other processes running on these cpu’s. > These cpu’s must be the pair that share their L2 cache, or hyperthreading > should be disabled and that cpu used. > > A setup that was historically used with iolcpus= kernel parameters was to run > everything on 1 cpu (disabling all but one), so that the cpu was hogged and > that got good latency results. > > This cgroups setup is the other way around, you’ll only run HAL on the > isolated cpu’s, the /RT cpuset. > > I’m no expert on this so there might be some more clarification needed from > the experts :) > > Bas > -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. Visit this group at https://groups.google.com/group/machinekit. For more options, visit https://groups.google.com/d/optout.
