From owner-freebsd-hackers@FreeBSD.ORG Thu Mar 1 02:22:32 2007 Return-Path: X-Original-To: freebsd-hackers@freebsd.org Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 7E01016A400 for ; Thu, 1 Mar 2007 02:22:32 +0000 (UTC) (envelope-from doconnor@gsoft.com.au) Received: from cain.gsoft.com.au (cain.gsoft.com.au [203.31.81.10]) by mx1.freebsd.org (Postfix) with ESMTP id D81D813C47E for ; Thu, 1 Mar 2007 02:22:31 +0000 (UTC) (envelope-from doconnor@gsoft.com.au) Received: from inchoate.gsoft.com.au (inchoate.gsoft.com.au [203.31.81.25]) (authenticated bits=0) by cain.gsoft.com.au (8.13.8/8.13.8) with ESMTP id l211h2Fw062238 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Thu, 1 Mar 2007 12:13:02 +1030 (CST) (envelope-from doconnor@gsoft.com.au) From: "Daniel O'Connor" To: freebsd-hackers@freebsd.org Date: Thu, 1 Mar 2007 12:12:58 +1030 User-Agent: KMail/1.9.5 MIME-Version: 1.0 Content-Type: multipart/signed; boundary="nextPart2424579.2sQUaeOuOM"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit Message-Id: <200703011212.59136.doconnor@gsoft.com.au> X-Spam-Score: -3.165 () ALL_TRUSTED,AWL,BAYES_00 X-Scanned-By: MIMEDefang 2.58 on 203.31.81.10 Subject: PCI slave performance (OT?) X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 01 Mar 2007 02:22:32 -0000 --nextPart2424579.2sQUaeOuOM Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Hi, We make a custom PCI card used for transferring data from our radar system= =20 acquisition chassis to a PC. It is rather old and we are looking at updatin= g=20 it but I would like to try and squeeze out as much performance as possible = to=20 satisfy our existing customers. It is a PCI slave only device using a PLX-9051 to read data from a FIFO. Th= is=20 means that we can't use prefetch or bursting because the PLX chip can throw= =20 data away in these circumstances :( Anecdotal evidence suggests that the maximum sustainable speed has dropped= =20 when we switched from 4.x to 6.x, but we also changed from i386 to amd64 an= d=20 used different motherboard chipsets so it's difficult to know where the=20 problem really lies. I plan on doing some tests to try and determine what is causing the slow do= wn=20 but I would appreciate any suggestions people may have :) One idea I did have was to get a dual core CPU and then one core would spen= d=20 all its time reading from the PCI bus while the other did data processing,= =20 etc.. Any suggestions gratefully received! =2D-=20 Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C --nextPart2424579.2sQUaeOuOM Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQBF5i+j5ZPcIHs/zowRAgrQAJ90VVA8WmgIdDkPGca5VuCH4ozLSQCdHMIy po2kNp8jtfXB06c3TT/aY4s= =s+T9 -----END PGP SIGNATURE----- --nextPart2424579.2sQUaeOuOM--