Subject: Re: EXTERNAL storage and substring on long strings
From: Scott Cain <cain@cshl.org>
To: Joe Conway <mail@joeconway.com>
Cc: pgsql-performance@postgresql.org
In-Reply-To: <3F2980DD.5010608@joeconway.com>
References: <1059679600.1429.41.camel@localhost.localdomain>
	<3F297CBA.1080703@joeconway.com>
	<1059684097.1429.62.camel@localhost.localdomain>
	<3F2980DD.5010608@joeconway.com>
Content-Type: text/plain
Organization: Cold Spring Harbor Lab
Message-Id: <1059685870.1429.73.camel@localhost.localdomain>
Mime-Version: 1.0
Date: 31 Jul 2003 17:11:15 -0400
Content-Transfer-Encoding: 7bit

Joe,

I'm working on the comparison--I think the best way to do it is to
reload the original data into a new database and compare them, so it
will take a while.

I have tuned postgresql.conf according to the page that everybody around
here seems to cite. I'll probably post back tomorrow with another set of
results.

Also, the perl script that did several queries used lengths of 5000,
10,000 and 40,000 because those are the typical lengths I would use
(occasionally shorter).

Thanks,
Scott

On Thu, 2003-07-31 at 16:49, Joe Conway wrote:
> Scott Cain wrote:
> > So it is possible that if I had a fast scsi drive, the performance might
> > be better?
> 
> Faster drives are always better ;-)
> 
> Did you try the comparison with shorter substrings? Also, maybe not 
> related to your specific question, but have you tuned any other 
> postgresql.conf settings?
> 
> Joe
> 
> 
-- 
------------------------------------------------------------------------
Scott Cain, Ph. D.                                         cain@cshl.org
GMOD Coordinator (http://www.gmod.org/)                     216-392-3087
Cold Spring Harbor Laboratory