Date: Sun, 1 Jun 1997 17:43:29 -0400 From: Joel Ray Holveck <joelh@gnu.ai.mit.edu> To: shag@concentric.net Cc: syssgm@dtir.qld.gov.au, freebsd-chat@FreeBSD.ORG Subject: Re: IDE or Ultra SCSI Message-ID: <199706012143.RAA32408@diazepam.gnu.ai.mit.edu> In-Reply-To: <338F2F83.55301D85@concentric.net> (message from Joshua Fielden on Fri, 30 May 1997 13:50:27 -0600)
next in thread | previous in thread | raw e-mail | index | archive | help
>SCSI can still only access two devices at any given time, one initiator ><usually, but not always the host adapter> and one target. SCSI II >allows for "disconnect" which means a drive can aquire a queue of >commands, remove itself from the bus, and reattatch when finished >processing those commands, but it obviously is no help for read/write >operations. :-) It does speed things up, but not quite to the extent I >believe you are thinking of. Although it is nice for long seek operations on devices such as CD-ROMs, tapes, scanners, etc, etc. And the bus mastering issue is wonderful for multimedia, but its effectiveness is lessened without disconnect. >> Or I could buy an 8x SCSI CD-ROM for twice the price of the same model >> drive with IDE interface. Sigh. >A CD-ROM <until 16x and higher hit> was always so slow that there would >be little improvement and I had these IDE busses "rusting" on my board, >so I have always gone ATAPI for CD-ROM, but with 16x, I'm not so sure >about that. I would say for 8x, don't bother with SCSI if you already >have the bus, and you get the added bonus of accessing IDE and SCSI in >parallel. i.e: read from the CD and write to the drive simultaneously. Really? I have a decent number of CD-ROM games that rely heavily on multimedia operations, and my old Texel 2x CD-ROM kept up with apps that normally wanted a 4x with no sweat. I had attributed this to disconnect and bus mastering, myself. >I have run three mixed IDE/SCSI machines, and used to work for FWB >Software, who makes IDE and SCSI device drivers for Mac and Windows, and >never have I seen a true benchmark that puts equivalent EIDE drives near >enough to SCSI performance to warrant consideration, except for the >newer 7200 RPM EIDE that have come out. Of course, I'm also the guy they >used to play "guess the drive speed" with at work, because after 1/2 >hr-45 minutes I could tell you which was the fastest/slowest drive on a >system with differences of only a couple of ms, so I'm a tad biased. :-) But for drives, benchmarks rarely are a good indicator of RL performance. -- http://www.wp.com/piquan --- Joel Ray Holveck --- joelh@gnu.ai.mit.edu All my opinions are my own, not the Free Software Foundation's. Second law of programming: Anything that can go wrong wi sendmail: segmentation violation -- core dumped
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199706012143.RAA32408>