Re: Serious performance problem - Mailing list pgsql-hackers

From Zeugswetter Andreas SB SD
Subject Re: Serious performance problem
Date
Msg-id 46C15C39FEB2C44BA555E356FBCD6FA41EB3DD@m0114.s-mxs.net
Whole thread Raw
In response to Serious performance problem  ("Tille, Andreas" <TilleA@rki.de>)
Responses Re: Serious performance problem
List pgsql-hackers
> I also
> enforced the index scan and compared with forbidding the index scan.
The
> result was on my more realistic examples that both versions performed
quite
> the same.  There was no *real* difference. For sure in this simple
query there
> is a difference but the real examples showed only 2% - 5% speed
increase
> (if not slower with enforcing index scans!).

You could somewhat speed up the query if you avoid that the sort
hits the disk. A simple test here showed, that you need somewhere
near sort_mem = 15000 in postgresql.conf.

Andreas


pgsql-hackers by date:

Previous
From: Lee Kindness
Date:
Subject: Re: Compiling on Solaris with Sun compiler
Next
From: Brent Verner
Date:
Subject: Re: Serious performance problem