>> If Rik or Juan can assure me that streaming I/O performance is a
>> high priority for them, I might feel better, but I see more
>> concern about boundary cases caused by quite different
>> application patterns.
>
>Before I start hacking on this, could you tell me if 2.4.0-ac17+
>gives you the performance you want?
>
>(or at least, if it reaches 80% of 2.3.51 performance, the rest
>seems related to the new elevator code and not to MM)
I'm about try booting ac17/18 in the next day or two. I'm a little
disturbed by some of the bug reports (the aic7xxx breakage in the
early ac series would have been fatal for me), but I'll give it a
shot.
Note that I always run with the elevator tuned to 10000/10000/128. Its
the only way to get even a baseline level of performance on pre-8.
Just as a footnote: I would really welcome a way to tell the kernel
"don't try to do anything smart with the elevator, the VM system or
anything else because I know what I'm doing". In my case, this is
really true: I have data carefully (well, as best as I can do with
ext2) laid out on a disk to match my access patterns, I am the only
process of any importance, i called mlockall(MCL_CURRENT|MCL_FUTURE),
I am running SCHED_FIFO and I basically want the operating system to
get the hell out of my way! OTOH, being able to read email while
recording 26 channels of 24 bit audio is quite a nice demo for
potential users :)
--p
-
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/
This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:33 EST