From owner-freebsd-scsi Thu Nov 22 4: 5:57 2001 Delivered-To: freebsd-scsi@freebsd.org Received: from zang.com (zang.com [216.34.130.70]) by hub.freebsd.org (Postfix) with ESMTP id E739337B405; Thu, 22 Nov 2001 04:05:53 -0800 (PST) Received: from localhost (johne@localhost) by zang.com (8.9.1a/8.9.1) with ESMTP id EAA29353; Thu, 22 Nov 2001 04:05:53 -0800 (PST) Date: Thu, 22 Nov 2001 04:05:53 -0800 (PST) From: John Engelhart To: freebsd-questions@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Bizzare Adaptec 3210S results Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org I'm having the oddest problems with an Adaptec 3210S controller in FreeBSD. Eight drives, four to a channel, initially set up as a RAID 0/5. After spending a day in this configuration and getting horrible performance (the lone IDE drive in the system was beating it out, but only slightly), I decided it was time to start from scratch. Rebuilt as a pure RAID 0 array performance went up considerably. But look at this: Version 1.01d ------Sequential Output------ --Sequential Input- --Random- -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP localhost 8G 69499 87 112164 60 18697 12 51132 98 56762 27 316.6 5 Writes twice the speed of reads!? Did hell freeze over? I'm curious to hear anyones opinion on why write performance is besting read performance. For reference, the RAID 0/5 performance was around 27-33Mb/s across the board. Even more appreciated is pointers to where the problem lies. I may end up going through the hassel of installing another OS and using it to benchmark the RAID to eliminate FreeBSD from the list of problems. Write-back is enabled. Soft-updates are enabled. The drives are fast, 15K seagate ones that individually start at a minimum of 47Mb/sec. This thing should be close to pushing the 64/33 PCI bus it's on. Anyone else have this card? Offer any pointers? Even a "Yea, you should be doing better than that" would be helpful. As a curiosity item, both the boot rom utility and the dptmgr util report the PCI bus a 528Mb/s, or 64/66. The AMD 760 chipset only does 64/33. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message