> I've just read that the next generation AMIGA will use Linux as the core OS,
> instead of QNX. ( http://www.amiga.com/diary/executive/linux-e.html )
> Amiga is a multimedia platform, and therefore needs realtime capatibilites.
> Unfortunately actually Linux doesn't allow to play any realtime
> (low-latency) mulimedia content (audio/video,etc.) in a reliable fashion
> , during high system load, especially disk I/O.
> I was very disappointed when I benchmarked a 130ms max scheduling latency of a
> realtime scheduled audio process (playing from RAM) under high disk I/O load on
> my PII400 + 16GB UDMA harddisk. (for infos see my latency benchmark page)
> Even Win98 is able to get 17ms audio latency on my machine (during
> heavy disk I/O) without many problems.
> It's a bit disappointing that under Linux we get latencies which are
> 50-100times the realistic (<5ms) latencies, we are FAR AWAY from the hardware
> limits.
> I was shocked when disabling DMA on my HD, my realtime process was
> not rescheduled for *4000ms* = 4sec !
> (a 100m sprinter is able to run 35m in the same time :-) )
> Not only the audio layer is affected but many other subsystems too.
> For example benchmarking usleep() I noticed a max 100-150ms scheduling
> latency during high disk I/O load.
> I don't understand why a process should block occasionally up to 150ms , when
> usleep()ing only a few ms, because of the high disk I/O.
> Actually there seems not a big interest to cure these deficiencies, just
> because many see Linux as a server OS , and not a desktop OS,
> on a web-server or fileserver you don't need guaranteed timing.
> (Or is there already some unoffical work underway to cure the scheduling
> problems of Linux ?)
> Who cares if his webpage shows up 150ms earlier or later ? ,
> or if your netscape executable which comes from the NFS server, is loaded
> 150ms later ? No one, I think.
> Does anyone know if the disk device driver (or device drivers in general) can be
> preempted from regular processes or is there no way to do this on Linux.
> Do we really need a microkernel to handle realtime multimedia ?
> Hopefully Amiga Inc, will help us to optimize these Linux deficiencies, through
> collaborating with the kernel/audio/video developement folks.
> comments ?
Are you tested all this stuff with `hdparm -u 1' or not ? If not then it
can be expected. Plus you can always use RTLinux if you wish ...
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/