pg.ddx.io pgsql-performance@postgresql.org mailing list archive
help / color / mirror / Atom feedMulti-processor question
2+ messages / 2 participants
[nested] [flat]
* Multi-processor question
@ 2006-09-25 08:30 Kjell Tore Fossbakk <kjelltore@gmail.com>
2006-09-25 08:47 ` Re: Multi-processor question Markus Schaber <schabi@logix-tt.com>
0 siblings, 1 reply; 2+ messages in thread
From: Kjell Tore Fossbakk @ 2006-09-25 08:30 UTC (permalink / raw)
To: pgsql-performance
Hello!
I got two AMD Opteron 885 processors (2.6ghz) and 8 gig of memory.
Harddrives are 4 scsi disks in 10 raid.
I'm running gentoo, and the kernel finds and uses all of my 2 (4) cpu's.
How can i actually verify that my PostgreSQL (or that my OS) actually gives
each new query a fresh idle CPU) all of my CPU's?
Kjell Tore.
--
Social Engineering Specialist
- Because there's no patch for Human Stupidity
^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: Multi-processor question
2006-09-25 08:30 Multi-processor question Kjell Tore Fossbakk <kjelltore@gmail.com>
@ 2006-09-25 08:47 ` Markus Schaber <schabi@logix-tt.com>
0 siblings, 0 replies; 2+ messages in thread
From: Markus Schaber @ 2006-09-25 08:47 UTC (permalink / raw)
To: pgsql-performance
Hi, Kjell Tore,
Kjell Tore Fossbakk wrote:
> I got two AMD Opteron 885 processors (2.6ghz) and 8 gig of memory.
> Harddrives are 4 scsi disks in 10 raid.
>
> I'm running gentoo, and the kernel finds and uses all of my 2 (4) cpu's.
>
> How can i actually verify that my PostgreSQL (or that my OS) actually
> gives each new query a fresh idle CPU) all of my CPU's?
On unixoid systems, use top to display individual CPU loads, they should
be balanced, if you issue the same type of queries in parallel.[1] On
Windows, the Task Manager should be able to display individual CPU load
graphs.
Note, however, that if you issue different kinds of queries in parallel,
it is well possible that some CPUs have 100% load (on CPU-intensive
queries), and the other CPUs have low load (processing the other, I/O
intensive queries.
Btw, if your queries need a long time, but CPU load is low, than it is
very likely that you're I/O bound, either at the disks in the Server, or
at the network connections to the clients.
HTH,
Markus
[1] You might need some command line options or keys, e. G. on some
debian boxes over here, one has to press "1" to switch a running top to
multi-cpu mode, pressing "1" again switches back to accumulation.
--
Markus Schaber | Logical Tracking&Tracing International AG
Dipl. Inf. | Software Development GIS
Fight against software patents in Europe! www.ffii.org
www.nosoftwarepatents.org
^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2006-09-25 08:47 UTC | newest]
Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2006-09-25 08:30 Multi-processor question Kjell Tore Fossbakk <kjelltore@gmail.com>
2006-09-25 08:47 ` Markus Schaber <schabi@logix-tt.com>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox