From: | "Merlin Moncure" <mmoncure(at)gmail(dot)com> |
---|---|
To: | "Tom Laudeman" <twl8n(at)virginia(dot)edu> |
Cc: | "Michael Fuhr" <mike(at)fuhr(dot)org>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Tuning to speed select |
Date: | 2006-08-10 14:26:11 |
Message-ID: | b42b73150608100726x5ed32799m37fec202594e1b79@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On 8/9/06, Tom Laudeman <twl8n(at)virginia(dot)edu> wrote:
> The speed of the query is (as Michael implies) limited to the rate at which
> the disk can seek and read. I have done experiments with views and cursors;
> there was no improvement in speed. I've also tried only pulling back
> primary keys in the hope that a smaller amount of data would more quickly be
> read into memory. No speed increase. I have also raised all the usual memory
> limits, with the expected results (slight speed improvements).
>
Are your data structures normalized? Performance problems queying a
single giganto table is usually (but not necessirly in your case) a
sign of a poorly designed table structure.
otherwise it's pretty clear you get the most bang for the buck with
hardware. consider upping ram and/or buying better disks. you could
buy cheap sata controller and 4 raptors in raid 0+1 configuration for
<1000$ and you will feel like you have supercomputer relative to what
you have now :)
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | Andy Foster | 2006-08-10 14:37:10 | CURRENT_TIMESTAMP wierd behaviour |
Previous Message | Carl R. Brune | 2006-08-10 14:15:38 | Re: read only transaction, temporary tables |