DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.co.uk;
	h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=qBYLcA/3aKrkvWbAqBDMuQ+9Ku+BezQ/NKr+MuKBnHVjiuNx6BWMxLmPO71wGEU9o547G9dRRleFY1wIvM/kVqX4oOvWlMTjogiugOOvvDo3fQky63NYjTbang5hWDbfcFuayMHSzPT1SqQDM7qNDaF/mimbYWru6tXzu1g/Atg=;
Message-ID: <860299.67368.qm@web26003.mail.ukl.yahoo.com>
Date: Mon, 11 Apr 2011 14:04:08 +0100 (BST)
From: Glyn Astill <glynastill@yahoo.co.uk>
Subject: Linux: more cores = less concurrency.
To: pgsql-performance@postgresql.org
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Hi Guys,=0A=0AI'm just doing some tests on a new server running one of our =
heavy select functions (the select part of a plpgsql function to allocate s=
eats) concurrently.=A0 We do use connection pooling and split out some sele=
cts to slony slaves, but the tests here are primeraly to test what an indiv=
idual server is capable of.=0A=0AThe new server uses 4 x 8 core Xeon X7550 =
CPUs at 2Ghz, our current servers are 2 x 4 core Xeon E5320 CPUs at 2Ghz.=
=0A=0AWhat I'm seeing is when the number of clients is greater than the num=
ber of cores, the new servers perform better on fewer cores.=0A=0AHas anyon=
e else seen this behaviour?  I'm guessing this is either a hardware limitat=
ion or something to do with linux process management / scheduling? Any idea=
 what to look into?=0A=0AMy benchmark utility is just using a little .net/n=
pgsql app that runs increacing numbers of clients concurrently, each client=
 runs a specified number of iterations of any sql I specify.=0A=0AI've post=
ed some results and the test program here:=0A=0Ahttp://www.8kb.co.uk/server=
_benchmarks/=0A