Re: Postgres refusing to use >1 core

From: Scott Marlowe <scott(dot)marlowe(at)gmail(dot)com>
To: Aren Cambre <aren(at)arencambre(dot)com>
Cc: Kevin Grittner <Kevin(dot)Grittner(at)wicourts(dot)gov>, pgsql-performance(at)postgresql(dot)org
Subject: Re: Postgres refusing to use >1 core
Date: 2011-05-10 02:34:58
Message-ID: BANLkTim3CJ=-Uww3J7z1NA7PTWfWykXwuQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Mon, May 9, 2011 at 8:15 PM, Aren Cambre <aren(at)arencambre(dot)com> wrote:
>> Your OS won't *see* eight processors if you turn of HT.  :-)
>> I'm going to pursue this digression just a little further, because
>> it probably will be biting you sooner or later.  We make sure to
>> configure the BIOS on our database servers to turn off
>> hyperthreading.  It really can make a big difference in performance.
>
> OK, OK, I need to admit that this is a Core i7 720QM on an HP Envy 14
> laptop. :-) There is no BIOS option to disable HT.
> I am a doctoral student (but married with kids, about 5-10 years over
> traditional doctorate student age) and am trying to speed up some of my data
> analysis with parallelism. Right now the current operation,if run in series,
> takes 30 hours and only stresses one of the 8 (fake) cores. I'd rather see
> something that maximizes CPU use, provided that it doesn't overwhelm I/O.

The easiest way to use more cores is to just partition the data you
want to work on into 4 or more chunks and launch that many
multi-threaded processes at once.

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Merlin Moncure 2011-05-10 02:37:37 Re: Postgres refusing to use >1 core
Previous Message Aren Cambre 2011-05-10 02:15:27 Re: Postgres refusing to use >1 core