That's nothing. I once got an upper bound of 35000ms (that's more than
enough time for *me* to finish a 100m sprint ;-) out of a PII-300 with a
7200RPM IDE UDMA drive doing video capture. Delays of 5-6 seconds were
more common. SCSI would give me delays of 3-4 seconds, which is better,
but not quite good enough to be useful (I had enough RAM to cope with
~3300ms of I/O latency at the time).
The weird thing that I've noticed is that for multimedia work
Pentium-class machines running Linux 2.0 have better (or at least less
bursty) scheduling than PIII-450's running Linux 2.2. Sure, the scheduler
would deadlock from time to time in 2.0, but at least while the machine
was running, it _worked_, damnit!
I dispute the assertion that this is due to optimization away from
realtime scheduling toward time sharing, because it's simply bad
time sharing as well. The CPU is not executing some other process or
reorganizing the buffer cache or blocking on disk I/O; it's simply idle,
while there are runnable tasks waiting for a CPU.
-- Zygo Blaxell, Linux Engineer, Corel Corporation. zygob@corel.ca (work) or zblaxell@furryterror.org (play). Opinions above are my own, not Corel's. Size of 'diff -Nurw [...] winehq corel' as of Thu Jun 3 13:14:00 EDT 1999 Lines/files: In 20094 / 98, Out 17319 / 152, Both 12093 / 142- 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/