Lists: | pgsql-general |
---|
From: | Glenn Sullivan <glenn(dot)sullivan(at)varianinc(dot)com> |
---|---|
To: | pgsql-general <pgsql-general(at)postgresql(dot)org> |
Subject: | shared buffers |
Date: | 2004-10-25 20:31:24 |
Message-ID: | 417D629C.7000805@varianinc.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-general |
Hi,
In the 7.4.5 version, the code is now trying to use a much larger
value for shared_buffers. I can certainly set this to a lower number
with the -B option. However, my guestion is:
What is the performance issue with setting shared_buffers to something like 45?
In doing some timing on my system, I cannot tell any difference with 45 versus 1000.
Perhaps I am looking at the wrong performance criteria.
Thanks,
Glenn
From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Glenn Sullivan <glenn(dot)sullivan(at)varianinc(dot)com> |
Cc: | pgsql-general <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: shared buffers |
Date: | 2004-10-25 21:08:49 |
Message-ID: | 28074.1098738529@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-general |
Glenn Sullivan <glenn(dot)sullivan(at)varianinc(dot)com> writes:
> What is the performance issue with setting shared_buffers to something like 45?
> In doing some timing on my system, I cannot tell any difference with 45 versus 1000.
What are you timing exactly? Almost every benchmark I've ever seen is
much happier with shared_buffers of at least a few hundred. Otherwise
you spend too much time copying data between kernel and user space.
regards, tom lane