On Fri, 19 Mar 2004, Greg Troxel wrote:

> "Steven M. Bellovin" <smb@research.att.com> writes:
>
> > Has anyone revalidated those numbers for modern disks?  When FFS first
> > came out, there was a lot of talk about rotational optimizations,
> > staying within the cylinder, etc.  Nowadays, of course, that's all
> > mythical.
>
> I believe rotational optimization is mythical due to whole-track
> buffering (as well as not knowing the layout), but staying 'close' in
> cylinder space probably still helps, in terms of small seeks vs
> whole-disk seeks.
>
> It should be possible to figure out the real geometry with timing
> experiments, but it seems no one has written a spiffy program to do
> this.  It would require considerable thought to avoid getting fooled
> by the disk's cache; perhaps it could be done with writes with WC
> disabled.

	Mmmm Traxtents

http://www.usenix.org/publications/library/proceedings/fast02/schindler/schindler_html/

-- 
		David/absolute          -- www.netbsd.org: No hype required --