On Saturday 09 March 2002 18.30, Martijn Sipkema wrote: [...] > preemtible kernel patch is already in 2.5...
Yes, but that doesn't mean that a timer implementation that will burn CPU cycles on the remainders of the ISA bus on UP systems is going in as well... > > After all, if games, audio applications and other multimedia > > applications are going to fight for the RTC, and then hog the > > scheduler with 1024+ Hz "wakeup rates", we have a problem. > > > > At that point, it would be a lot more efficient if those > > applications could just use high resolution software timers > > (driven by the programmable system timer, something like in KURT, > > AFAIK), programmed to wake threads up *only when required*, as > > opposed to "at an arbitrary rate, just high enough to do the > > job". > > I think IBM had a patch that did this because the overhead of > getting interrupted by the system timer every 10ms was too large > running a lot of linux kernel images on their servers. Yeah, I think I've heard that mentioned somewhere... Any idea how they figured out when would be a good time to consider rescheduling? > > Is this relevant to other applications than "professional MIDI > > applications"? > > I think so. Well, in fact, it doesn't look better than that a serious video driver will need it as well, to simulate the missing retrace sync IRQ. (In that sense, my partly successful attempts to do it without accurate scheduling seem a bit "counterproductive", as they might invalidate one of the few reasons to get highres timers in the kernel... *heh*) > > > I really don't think that's necessary. A <<ms accurate time > > > should be sufficient. > > > > Well, it all depends on your definition of "correct MIDI > > timing"... > > <<ms Well, it doesn't get much better than that, so it seems like we're basically aiming for the same kind of figures. The main difference is probably that I'm a hard real time guy, who prefers when everything is totally deterministic, and things happen exactly as you schedule them, with any significant jitter being eliminated by properly designed I/O hardware. As a result, most current MIDI hardware seems broken to me. :-) //David .- M A I A -------------------------------------------------. | Multimedia Application Integration Architecture | | A Free/Open Source Plugin API for Professional Multimedia | `----------------------> http://www.linuxaudiodev.com/maia -' .- David Olofson -------------------------------------------. | Audio Hacker - Open Source Advocate - Singer - Songwriter | `-------------------------------------> http://olofson.net -'
