Date: Wed, 27 Sep 2000 19:01:40 -0500 From: David Kelly <dkelly@hiwaay.net> To: chat@FreeBSD.ORG Subject: Re: CD writers - recommendations Message-ID: <200009280001.e8S01ea28683@nospam.hiwaay.net> In-Reply-To: Message from John Galt <galt@inconnu.isu.edu> of "Wed, 27 Sep 2000 12:00:35 MDT." <Pine.LNX.4.21.0009271146100.28764-100000@inconnu.isu.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
John Galt writes: > > Adaptec 1542 has a Z-80 onboard to share in the disk processing tasks, I'm > betting that the better cards have beefier processors. That poor little Z-80 on my 1542CF couldn't do 5 million NOP's/second, much less keep up with the FAST SCSI 10M bytes/sec data rate. Seems like the Z80 was clocked at 5 MHz? Then was it between 3 to 8 clocks/ instruction? There are custom scripting engines in the 2940 and Symbios SCSI cards. Lesser SCSI cards, probably the Adaptec 2906, are little more than a data latch and line drivers. > I have never seen a GP processor on an IDE card. No place to put one. At it simplest, an IDE "interface" is only an extension of the CPU bus. The CPU uses this IDE bus to tickle registers in the IDE device which emulate the old Western Digital HD controller. Last time I saw a discrete IDE interface it was little more than one 74LS245 and one 74LS373. > [...]cheaper[...] For all the deleted verbage, that is the one word you used that makes sense. IDE devices are cheaper. They are cheaper because in their PC market "cheap" sells better than "quality". In the SCSI market if you lack quality then you are burnt toast. Lately I had the opportunity to select parts and assemble a new machine. "Reliable" and "value" was important. Raw performance was not. Purchased a 15G IBM UDMA100 HD. Selected IBM because I believe they are building the best HD's these days, and the $120 price was only $20 or $30 more than a cheap Maxtor or Western Digital. Connected to a UDMA33 MB (Asus P3B-F). Was awfully pleased when bonnie reported over 20MB/sec sustained in most tests using a file twice the size of core. Gee, that's roughly twice as fast as any of my UWSCSI systems. "systat -v" routinely shows 150 or more tps. Noticed the 64k per transfer limit must have been removed as often the blocksize field was ###'d. Got that machine connected to another via 100baseT full duplex thru a switch and decided to see what it could do. Got 8.5MB/sec sustained that way. Repeated using the same file and got 11.5 MB/sec playing the file out of cache. Same HD in an old junk AMD 5x86/133 (a 133 MHz 486) can't do DMA and sucks most of the CPU cycles to do over 1 MB/sec, while an Adaptec 2940 and IBM DCHS does a little better on 1/5 the CPU utilization. -- David Kelly N4HHE, dkelly@hiwaay.net ===================================================================== The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200009280001.e8S01ea28683>