From: urlichs@smurf.sub.org (Matthias Urlichs) Subject: Re: SCSI Performance (Yet Again) Date: 25 Aug 1993 17:37:51 +0200
In comp.os.linux, article <1993Aug23.161232.1118@kf8nh.wariat.org>,
bsa@kf8nh.wariat.org (Brandon S. Allbery) writes:
> In article <1993Aug22.200138.15349@hip-hop.suvl.ca.us> root@hip-hop.suvl.ca.us (Remco Treffkorn) writes:
>
> >There are some critical assumtions here:
> > You should do an fsync after writing and include the time till
> > fsync returns. If you don't then the read performance is too high
> > and write performance too high. BUT between the two you will see
> > a glimmer of reality.
>
> Another critical assumption is that the cache is flushed all at once. If it
> does it a block at a time to free up space for the next buffered write, (e.g.
> one physical write per buffered write), performance will suffer because of
> buffer management overhead as well as because of the physical write time.
>
You should do a sync() _before_ starting your measurements, to make sure
that no writes of unrelated blocks are included.
-- Matthias Urlichs -- urlichs@smurf.sub.org -- Phone: NONE; use email or lose. Schleiermacherstrasse 12 -- 90491 Nuernberg -- Germany || Linux+Mac Consulting