Date: Fri, 2 Apr 2004 14:25:35 +0200 (MET DST) From: Helge Oldach <helge.oldach@atosorigin.com> To: thierry@herbelot.com Cc: bzeeb-lists@lists.zabbadoz.net Subject: Re: implications of SMP kernel on UP Message-ID: <200404021225.OAA28162@galaxy.hbg.de.ao-srv.com> In-Reply-To: <200404011829.04221.thierry@herbelot.com> from Thierry Herbelot at "Apr 1, 2004 6:29: 4 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Thierry Herbelot:
>Le Thursday 01 April 2004 09:10, Bjoern A. Zeeb a écrit :
>> what are the implications on running an SMP enabled kernel on a UP
>> machine ?
>>
>> I first thought of things like:
>> - performence (most likely not worth the discussion ?)
>
>I got an improvement with a factor of ten between an SMP and a UP
>kernel on a HTT-enabled P4/2,6GHz/800MHz FSB on network transfers (with
>gigabit Ethernet boards: SMP gives about 6MB/s for FTP transfer rate,
>and UP gives up to 75MB/s)
>
>So: as long as the network stack is not fully locked (this is coming -
>perhaps for 5.3), a server should definitely run a UP kernel.
Ooops, that is indeed dramatic.
Does the same locking impact also affect other, interrupt-bound
subsystems, such as the disk subsystem (in particular SCSI)?
I am asking because I have a HTT machine that walks some 35.000
files every few minutes with an aggreated disk I/O load of only 5
MBytes/second, while the SCSI hardware is able to do SCSI-3/160...
It appears that a lot of processes are indeed disk bound ("D").
Helge
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200404021225.OAA28162>