Subject: Re: pagedaemon: measuring performance
To: Sumantra Kundu <sumantra@gmail.com>
From: Bill Stouder-Studenmund <wrstuden@netbsd.org>
List: tech-kern
Date: 05/30/2007 13:26:50
--wac7ysb48OaltWcw
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, May 28, 2007 at 12:33:57AM -0500, Sumantra Kundu wrote:
> Hi,
>        Currently I am tweaking the performance of the page daemon by
> modifying the decision thresholds based on which the daemon moves
> pages from active  to inactive list.

I'm confused. I thought the active list contained pages that were in use,=
=20
and the inactive ones that weren't. Thus how does anything other than the=
=20
actual user of pages make this change?

> Questions: What is the most appropriate way (measurement parameters)
> to understand the change in performance of the page daemon?

To be honest, this depends.

It depends on what you are doing.

I think a good thing to try is pick a class of workload, such as=20
mostly-sequential access. Then pick a test tool that generates this load.=
=20
Then monitor its performance metrics as you adjust things.

Take care,

Bill

--wac7ysb48OaltWcw
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (NetBSD)

iD8DBQFGXd4KWz+3JHUci9cRAnzqAJ9myev3sx8dW0IrQRKJ1NmBRPsxYgCgi2ns
4/+4/nB5ZR/L9YzgHHsjlfk=
=IeNC
-----END PGP SIGNATURE-----

--wac7ysb48OaltWcw--