[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4578F982.70506@homeurl.co.uk>
Date: Fri, 08 Dec 2006 13:34:58 +0800
From: Bob <spam@...eurl.co.uk>
To: linux-kernel@...r.kernel.org
Subject: Re: Kernel 2.6 SMP very slow with ServerWorks LE Chipset
Alan wrote:
>> As a test of raw CPU power I've been decompressing the kernel tree, with
>> a UP 2.6 kernel this takes about 1m 15s, I don't know if bz2 is
>> multithreaded but even if it's not I would expect a slight speed increase
>> but in fact with a SMP 2.6 kernel it take 13 ~ 26m, with a SMP 2.4
>> kernel it takes 1m 28s and with a 2.4 UP 1m 35s.
>
> The 2.4 numbers look correct (slightly slower), the 2.6 numbers do not.
>
>
> Nothing obviously wrong from the traces however. If you pin the bzip to a
> given processor do you get different results according to which CPU ?
>
> (see man taskset for info on the commands)
>
> If you get very different times on the two processors that will be very
> useful information.
Mmm CPU 0 in 1m 32s, but I was running things like ps and taskset in
another terminal to verify affinity, CPU 1 still hadn't finished in 40m
when I killed it. I've downloaded the LFDK mentioned in the other post
and will post the results when I have them.
Thanks for replying,
This is the script I'm using to time decompression
nas:~# cat ./cputest.sh
date
taskset -c 1 nice -n -8 tar -xjf /root/linux-2.6.18.1.tar.bz2
date
uname -a
rm -r /root/linux-2.6.18.1/
nas:~#
both the tar and bzip2 processes have their affinity and priority set by
the above command.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists