Date: Sun, 3 Aug 2003 17:20:03 -0700 From: David Schultz <das@FreeBSD.ORG> To: "Dag-Erling =?us-ascii:iso-8859-1?Q?Sm=F8rgrav?=" <des@des.no> Cc: arch@FreeBSD.ORG Subject: Re: headsup: swap_pager.c Message-ID: <20030804002003.GA5823@HAL9000.homeunix.com> In-Reply-To: <xzpel0568cn.fsf@dwp.des.no> References: <6955.1059728599@critter.freebsd.dk> <xzpel0568cn.fsf@dwp.des.no>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Aug 01, 2003, Dag-Erling Smrgrav wrote:
> "Poul-Henning Kamp" <phk@phk.freebsd.dk> writes:
> > The thing you overlook is that often when things gets paged out, the
> > system is short on memory and therefore more likely to not do anything
> > productive, whereas when things gets paged in, there are a better chance
> > of some other process being able to use the CPU time productively.
> > If we did predictive pageouts like some of the "serious" mainfram OS's
> > this would be less true.
>
> How hard would it be to get the kernel to write the pages "most likely
> to be swapped out" to swap in the idle loop, to save time if / when
> they actually need to be swapped out later?
>
> I thought we already did this to some extent (ref. FAQ 16.1), but
> apparently I was wrong?
FreeBSD already does that. ;-) You can control the number of
clean pages that it keeps around with the sysctls
vm.v_cache_{min,max}, but you shouldn't need to tune anything to
get good performance. FWIW, the stuff phk is working on is
in a different area; it has to do with what swap device your
pages wind up on after the VM system has already decided to
write them out.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030804002003.GA5823>