--H+4ONPRPur6+Ovig
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Jan 07, 2004 at 09:41:27AM +1000, George Michaelson wrote:
> On Tue, 6 Jan 2004 15:31:03 +0100 (MET) Paul Kranenburg <pk@cs.few.eur.nl>
> wrote:
>=20
> > i.e. to around 5% of physical memory.  To verify that this is the case,
> > could you do
> >=20
> > 	`sysctl -w vm.bufcache=3D5'
> >=20
> > and see if that makes the symptoms go away?
>=20
> Yes. Many thanks.=20
>=20
> So with the default I got on boot 30, and 5 demonstrating a marked improv=
ement,
> whats a rational choice? I'm tempted to try binary chop on the space betw=
een for
> a while. 30 must have been predicated on some belief about 'more is bette=
r' but
> it clashed head-on with interactive response on laptops (at least).=20

I think 30 was just chosen as something that worked. We haven't put a lot=
=20
of effort into tuning, and it could be that we just need to re-tune to go=
=20
with the new caching code.

Take care,

Bill

--H+4ONPRPur6+Ovig
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQE/+1eMWz+3JHUci9cRArg0AJ94dpBaAJ2WAHY1ga+JTWt8gKVLCQCfQcGb
0gjmQdo0D7B56WEE0zxfZyY=
=jwv/
-----END PGP SIGNATURE-----

--H+4ONPRPur6+Ovig--