From: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
---|---|
To: | Andreas Pflug <pgadmin(at)pse-consulting(dot)de> |
Cc: | Merlin Moncure <merlin(dot)moncure(at)rcsonline(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Timing of 'SELECT 1' |
Date: | 2004-03-10 19:18:17 |
Message-ID: | 200403101918.i2AJIIf13951@candle.pha.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Andreas Pflug wrote:
> Bruce Momjian wrote:
>
> >>There seems to be a 'PostgreSQL ping' time of about 1-2 ms in best case
> >>conditions which limits the amount of queries you can fire off in 1
> >>second, no matter how simple. In certain rare cases this is something
> >>of a bottleneck. In my personal case it would be nice to see that time
> >>lower because converted COBOL applications tend to generate a lot of
> >>'simple' queries.
> >>
> >>
> >
> >Yes, most of that might be network time. I am using log_duration, which
> >I think just tests backend time, not network transfer time, but I might
> >be wrong. I want to look into this as it seems no one knows the answer.
> >
> >
> >
> That's easy to verify with standard ping. In my switched 100MBit
> network, roundtrip for small packets is about 0.2 ms, and 0.5ms for 1kb
> packets. How about context switch latency?
I am on a dual Xeon. I wouldn't think there was that much of a context
switch overhead, except for kernel calls.
--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
From | Date | Subject | |
---|---|---|---|
Next Message | Bruce Momjian | 2004-03-10 19:24:18 | Re: selective statement logging |
Previous Message | Andreas Pflug | 2004-03-10 19:06:38 | Re: Timing of 'SELECT 1' |