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:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=4vKsy1xv7uKf9Ci9jPIW88c/ddhjV4s8nFRxOLLJWOQInJSXmY6Y9EOkv9M5gxjjdus+bbWTLZHBzwRC7SpSDjz+QbDDkbed0MRbY7Pcx4wJv4ra8X6xH12lTwdJB3crdfq+IAMcCiwarqfjVLqLUb5+rw4PGeKRb6NoMEzFn7o=;
Message-ID: <976931.36062.qm@web26004.mail.ukl.yahoo.com>
Date: Mon, 11 Apr 2011 21:17:04 +0100 (BST)
From: Glyn Astill <glynastill@yahoo.co.uk>
Subject: Re: Linux: more cores = less concurrency.
To: Steve Clark <sclark@netwolves.com>, david@lang.hm
Cc: Scott Marlowe <scott.marlowe@gmail.com>,
	"Joshua D. Drake" <jd@commandprompt.com>,
	Kevin Grittner <Kevin.Grittner@wicourts.gov>,
	pgsql-performance@postgresql.org
In-Reply-To: <alpine.DEB.2.00.1104111259350.32165@asgard.lang.hm>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

--- On Mon, 11/4/11, david@lang.hm <david@lang.hm> wrote:=0A=0A> From: davi=
d@lang.hm <david@lang.hm>=0A> Subject: Re: [PERFORM] Linux: more cores =3D =
less concurrency.=0A> To: "Steve Clark" <sclark@netwolves.com>=0A> Cc: "Sco=
tt Marlowe" <scott.marlowe@gmail.com>, "Joshua D. Drake" <jd@commandprompt.=
com>, "Kevin Grittner" <Kevin.Grittner@wicourts.gov>, pgsql-performance@pos=
tgresql.org, "Glyn Astill" <glynastill@yahoo.co.uk>=0A> Date: Monday, 11 Ap=
ril, 2011, 21:04=0A> On Mon, 11 Apr 2011, Steve Clark=0A> wrote:=0A> =0A> t=
he limit isn't 8 cores, it's that the hyperthreaded cores=0A> don't work we=
ll with the postgres access patterns.=0A> =0A=0AThis has nothing to do with=
 hyperthreading. I have a hyperthreaded benchmark purely for completion, bu=
t can we please forget about it.=0A=0AThe issue I'm seeing is that 8 real c=
ores outperform 16 real cores, which outperform 32 real cores under high co=
ncurrency.=0A=0A32 cores is much faster than 8 when I have relatively few c=
lients, but as the number of clients is scaled up 8 cores wins outright.=0A=
=0AI was hoping someone had seen this sort of behaviour before, and could o=
ffer some sort of explanation or advice.