From ns.itep.ru!Radio-MSU.net!Gamma.RU!srcc!satisfy.kiae.su!demos!solace!nntp.uio.no!nntp.uib.no!nntp-bergen.UNINETT.no!nntp-trd.UNINETT.no!due.unit.no!invalid.ed.unit.no!geirfrs Tue May 21 10:19:41 1996 Path: ns.itep.ru!Radio-MSU.net!Gamma.RU!srcc!satisfy.kiae.su!demos!solace!nntp.uio.no!nntp.uib.no!nntp-bergen.UNINETT.no!nntp-trd.UNINETT.no!due.unit.no!invalid.ed.unit.no!geirfrs From: geirfrs@invalid.ed.unit.no (Geir Frode Srrensen) Newsgroups: comp.os.os2.beta Subject: Re: Hard Drive size & HPFS Date: 15 May 1996 20:53:49 GMT Organization: Omega Verksted, NTH, Norway Lines: 60 Distribution: world Message-ID: <4ndg8t$g83@due.unit.no> References: <4nd7ia$2ql@hammer.msfc.nasa.gov> NNTP-Posting-Host: invalid.ed.unit.no X-Newsreader: TIN [version 1.2 PL2] Dr. Doug Rickman (doug@hotrocks.msfc.nasa.gov) wrote: : Does anyone know if WARP using HPFS will be happy with one of the : new Seagate ulta wide SCSI-3 drives? Specifically the ST19171W : which has a 9.1 GByte capacity (and a transfer rate of 40MBytes/sec. : - drool, slobber, drool) There was an discussion some time back about HPFS partition size limit. I'm not sure what the conclusion was, as it was large enough to be of only theoretical interest to me. 64 GB perhaps? But HPFS should do equally well on large partition as it does on small (well, medium.) But what triggered me was your estimate of transfer rate. 40 MByte/s is the _limiting_ rate on _data block_ transfer for the SCSI-3 SPI _narrow_ media (which is nothing but good ole SCSI-2 "16-bit wide" with narrowed in impedance tolerances) 8 bit narrow is likely to disappear. (Well, it's still there, but you will have to extract it from line 11-60 yourself. Or get an adapter.) SCSI-3 Wide has the same width as SCSI-2 Wide, namely 32 bit, but in 16 bit chunks over P and Q cables, instead of the SCSI-2 8 bit A cable and 24 bit B cable. (X3T9.2 never defined SCSI-2 _16 bit_ wide as an _physical_ option. If you do not believe me, feel free to read the spec. www.symbios.com/pub/... ) Besides, the SCSI-3 draft has no mention of Ultra. It is formally called Fast-20; as in 20 MHz. Besides, SPI is just one of the five alternative transport medias, rangeing from Fibre Channel (GB/s+++ ;), via IEEE 1394 to SPI and in fact also any suiteable networks like ethernet... Boy, are we going to have some notation confusion :-) But what I meant to say, is that your harddrive is not going to be any faster just because the interface becomes faster. (Disregarding buffer transfer.) It is still going to be ~3-6MBytes/s because that is the maximum speed it can pick up it's data from the discs! On the other hand you could get nine 1GBytes HDDs instead, and RAID them to one mega ultra etc. fast ~7 GByte drive. What you really gain from SCSI-3 is lower _bus utilization_ per. device! Just so you won't be disappointed... : Doug Rickman : -- : _____________________________________________________________ : Doug Rickman | internet doug@198.116.59.82 : ES41 | doug@hotrocks.msfc.nasa.gov : MSFC | : Huntsville, AL 35812 | tel: 205-922-5889 fax: -5723 : National Aeronautics and Space Administration : Marshall Space Flight Center : _____________________________________________________________ ****************************************************** Never ever underestimate the power of human stupidity. -Robert Anson Heinlein GeirFRS@invalid.ed.unit.no ******************************************************