Received: from localhost (unknown [200.46.204.191]) by postgresql.org (Postfix) with ESMTP id D4D7C9F9D8F for ; Tue, 11 Sep 2007 20:00:30 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.204.191]) (amavisd-maia, port 10024) with ESMTP id 99578-08 for ; Tue, 11 Sep 2007 20:00:25 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.5 Received: from noel.decibel.org (noel.decibel.org [67.100.216.10]) by postgresql.org (Postfix) with ESMTP id C12A29F9C6B for ; Tue, 11 Sep 2007 20:00:25 -0300 (ADT) Received: by noel.decibel.org (Postfix, from userid 1001) id 44F15564B2; Tue, 11 Sep 2007 23:00:24 +0000 (UTC) Received: (hashcash-sendmail, from uid 1001); Tue, 11 Sep 2007 18:00:22 -0500 Date: Tue, 11 Sep 2007 18:00:22 -0500 From: Decibel! To: Mindaugas Cc: pgsql-performance@postgresql.org Subject: Re: Barcelona vs Tigerton Message-ID: <20070911230022.GH38801@decibel.org> References: MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="MFP99WsVFCW8Z+Yt" Content-Disposition: inline In-Reply-To: X-Operating-System: FreeBSD 6.1-RELEASE-p10 amd64 X-Distributed: Join the Effort! http://www.distributed.net User-Agent: Mutt/1.5.13 (2006-08-11) X-Hashcash: 1:20:070911:ml@kilimas.com::4OQSbUMSw0TkjjNg:0001l5D X-Hashcash: 1:20:070911:pgsql-performance@postgresql.org::uLU9MVsA3/1YWxqz:00000 0000000000000000000000000XUJ X-Virus-Scanned: Maia Mailguard 1.0.1 X-Archive-Number: 200709/206 X-Sequence-Number: 26751 --MFP99WsVFCW8Z+Yt Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Sep 11, 2007 at 10:57:43AM +0300, Mindaugas wrote: > Hello, >=20 > Now that both 4x4 out it's time for us to decide which one should be be= tter for our PostgreSQL and Oracle. And especially for Oracle we really nee= d such server to squeeze everything from Oracle licenses. Both of the datab= ases handle OLTP type of the load. You might take a look at replacing Oracle with EnterpriseDB... but I'm a bit biased. ;) > Since we plan to buy 4U HP DL580 or 585 and only very few of them so po= wer ratings are not very critical in this our case. >=20 > First benchmarks (http://www.anandtech.com/IT/showdoc.aspx?i=3D3091) sh= ow that Intel still has more raw CPU power but Barcelona scales much better= and also has better memory bandwidth which I believe is quite critical wit= h 16 cores and DB usage pattern. > On the other hand Intel's X7350 (2.93GHz) has almost 50% advantage in C= PU frequency against 2GHz Barcelona. Databases are all about bandwidth and latency. Compute horsepower almost never matters. The only reason I'd look at the clock rate is if it substantially affects memory IO capability; but from what I've seen, memory seems to be fairly independent of CPU frequency now-a-days, so I don't think there's a huge difference there. --=20 Decibel!, aka Jim Nasby decibel@decibel.org EnterpriseDB http://enterprisedb.com 512.569.9461 (cell) --MFP99WsVFCW8Z+Yt Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.3 (FreeBSD) iD8DBQFG5x4GdO30qud8SkgRAoU9AJ9lcT3lCig9GIWpHFagpbzYHi0bSgCbB8Vl WRwRA+ONhva3X/NvTewRwQ0= =4aPO -----END PGP SIGNATURE----- --MFP99WsVFCW8Z+Yt--