soundcheck wrote: > Look guys. > > This is a development thread. And not a Kindergarden. > Only quite knowledgeable folks have a clue what's being talked about and > how to apply it to SoA. There's IMO pretty much no risk associated to > my posts. > I wouldn't post my overclocking settings though. These might cause some > real damage. I do strongly recommend to have a look at overclocking. > You can run a PI at much higher rates without causing damage. Even the > power consumption stays pretty low. Several OC settings are even > supported by the PI Foundation. > A combined ARM based server/player IMO can't afford to let this or that > load percentage go. It's not the average load causing issues. It's the > peakloads and the bottlenecks. > > All optimizations that I posted are pretty much known at other places. > The info is spread all over the web. I'm not reinventing the wheel > here. > And I never said what would be the best for our purposes! > In the past I've had several occasions offline with Triode. Several of > my inputs - sometimes slightly adapted - and that includes compiler > options - went into the images. > > > Regarding Triode. > Triode goes the easy way (from an optimization perspective). He doesn't > put much effort on optimizing his OS. He makes everything going. > And he keeps it as generic as possible. He has to. Otherwise he gets > stuck. Doing what he does is a full time (24h) job on its own. There's > no space left for fancies. > The outcome is more then sufficient for the vast majority of the current > squeeze community. Fair enough. I do not belong to that group. > There are several other Linux OSes out there, which do have a focus on > optimizing and advancing the projects, e.g. Volumio, Rune, Moode, Voyage > you name it. > These folks do share the market between them. I don't see SoA on the > same level yet. > > Let's keep the fingers crossed that the SoA project won't die before it > really started. > > Unfortunately the community is not supporting Triode very much in > maintaining and advancing SoA as it seems to me (that includes, beside > optimizations, documentation/website etc.). > That keeps SoA in a state where it still is - a one-man-show. And a kind > of work-in-progress development project - which is a pity. > > > > @Krisbee > > Your changes are OK to me. No issue. At least that's some constructive > feedback. We should find a way to feed that on Triodes git page.
My views on the points you raised are: 1.OPTIMIZATIONS I doubt we would agree on what optimizations are worth pursuing as I remain highly sceptical of the kind things that have long been circulating on the web: e.g. RT kernels, cpu process affinity, in memory stuff, etc. etc. I know OS ( Operating System) is a shorthand term, but we are really talking about examples of Linux Embedded Systems for specific audio applications. Comparing SOA to Volumio, or any of the others you mentioned, is like comparing apples to oranges. Apart from the the fact they are all based on MPD, except for Voyage (now Mubox), they need to be concerned about optimization as they are far more demanding of resources because of their web interfaces. They need to focus on how best to code this area. A recent announcement at Volumio is tacit acceptance that the project was stagnating without a major re-write. See here: https://volumio.org/introducing-volumio-2/ In contrast, Triode has used his skills to build an extremely Lightweight and flexible system for LMS/squeezelite users. (It runs on x86 and a range of ARM devices, and AFAIK uniquely offers system installers.) Even with its obvious hardware limitations, the single core RPi can work as a combined server/player using SOA. Triode's installers create systems that already contain useful optimizations: udev rules to match best fit scheduler to storage medium; sysctl tweaks to ipv4 network tcp settings; setting systemd journal storage to volatile and a systemd service unit file that prioritises squeezelite. Squeezelite, and other core apps, are built with links to librt were appropriate. Triode also provides a patched kernel for both the cubietruck and wandboard. 2. SOA WHAT NEXT? Without Triode the show is most definitely not on the road. Yet, Triode may consider that the project's aims are, for the most part, already met and I don't imagine for one minutes market share was of interest to him. There has been considerable feedback in this long running thread, but it's hard to tell exactly what uptake there has been of SOA. As to anyone stepping into his shoes, who would be willing and have the skills to fork the project even for maintenance purposes? For the moment, SOA is at a standstill, and this thread is for mutual self help. ------------------------------------------------------------------------ Krisbee's Profile: http://forums.slimdevices.com/member.php?userid=59080 View this thread: http://forums.slimdevices.com/showthread.php?t=101624
_______________________________________________ unix mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/unix
