Re: Non-linear Performance - Mailing list pgsql-general

From Bruce Momjian
Subject Re: Non-linear Performance
Date
Msg-id 200206072354.g57NsR408142@candle.pha.pa.us
Whole thread Raw
In response to Re: Non-linear Performance  ("Peter A. Daly" <petedaly@ix.netcom.com>)
Responses Re: Non-linear Performance  (Curt Sampson <cjs@cynic.net>)
List pgsql-general
Peter A. Daly wrote:
> >
> >
> >>What is the most amount of SORT_MEM it makes sense to allocate?
> >>
> >
> >I've never experimented with it, but certainly the standard default
> >(512K = 0.5M) is pretty small for modern machines.  In a 2G machine
> >I might try settings around 100M-500M to see what works best.  (Note
> >this is just for a one-off btree creation --- for ordinary queries you
> >need to allow for multiple sorts going on in parallel, which is one
> >reason the default sort_mem is not very large.)
> >
> I will run some benchmarks and let the list know the results.  If this
> can speed it up a large amount, I can get another 6 gig of RAM into this
> machine which I hope can let me leave the SORT_MEM at a high enough
> amount to speed up our huge nightly batch database reload.
>
> It's a Dual Xeon 500Mhz Machine.

I am curious how you are going to address >4 gig of RAM on a 32-bit
system, especially if you want to address it all from the same process.

--
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

pgsql-general by date:

Previous
From: Andrew Sullivan
Date:
Subject: Re: Are globally defined constants possible at all ?
Next
From: Gregory Seidman
Date:
Subject: more on timestamps