Article View: comp.sys.ibm.pc.hardware.storage
Article #1016Re: Is it possible to actually get a 16 MB/sec data transfer rate w/ an EIDE drive?
From: James Bannon
Date: Sun, 14 Dec 1997 00:00
Date: Sun, 14 Dec 1997 00:00
85 lines
4135 bytes
4135 bytes
In article <671irb$3jk@bgtnsc01.worldnet.att.net>, Ron Reaugh <Ron- Reaugh@worldnet.att.net> writes > > > >No, UDMA EIDE drives provide sufficient bus bandwidth for video >applications including direct to HDs. SCSI is the best choice for servers. > > >> >> [The bit about sustained transfer rates is interesting. Consider SCSI-II: >The >> maximum theoretical transfer rate of such a device is 10 MB/sec (burst), >or >> roughly 4 MB/sec sustained, at least according to the official >specification. >> There is no way that SCSI-II could give 20 MB/sec sustained.] > > >Well, not quite. Fast narrow SCSI2 has a peak burst of 10MB/sec. and can >sustain about 8.5MB/sec. Fast wide peak burst of 20MB/sec. and sustained >about 17-18MB/sec. > >> Now the Seagate Barracuda! This drive is designed to give sustained >throughput >> of 4 MB/sec; i.e., the fastest a SCSI-II bus can run. > > >No, depending on which model Barracuda it'll sustain in the range of >5.5-11MB/sec. > > >> At that speed, >> interleaving reads/writes from applications with differing requirements >is going >> to be slower than EIDE, even when using a decent controller. > > >The Barracuda 4XL is as fast as any UDMA drive except possibly the Fireball >SE. > > >> The way to get >> around this is to install the drive on a separate controller and reserve >the >> Barracuda for applications it's designed for, specifically Video transfer >direct >> to HD. So you see, the Barracuda is not a good device for a >general-purpose >> workstation! > > >A Barracuda is a great drive for a workstation when connected to a >busmastering PCI controller; it's just that it's very expensive compared >to a UDMA drive. > >Regards, > >Ron Reaugh -- James Bannon Hi Ron, I stand corrected! It's just as well somebody actually reads the newest device specs correctly! I was using the Sun one, about 4 years old now! When I'm asked to calculate this sort of thing for somebody I usually divide the maximum burst rate by about two or thereabouts to get an "average" figure. Of course, so much depends on the application and spread of read/write requests and then you have to take into account the fact that the OS queueing system gets in the way! Not easy to give precise figures unless you run loads of timing tests with an application spread the user is expecting! Information r.e. the Ultra-EIDE is appreciated. I think I'll let RM (my machine supplier) know! By the way, do you where I can get descriptions of the specs? I can't seem to get one out of RM and we sell their kit! As far as the Barracuda is concerned, I think we're talking about two different models. My manual quotes 4 MB/sec as the sustained transfer rate. Anyway, I'll sign off this debate now having demonstrated my ignorance! Cheers.
Technical Details
Message-ID:
<n$PBRHAAyFl0EwX7@james-bannon.demon.co.uk>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailReferences:
<348f685a.259490677@news.ebtech.net> <881848354snz@michaels.demon.co.uk> <34913d16.2409107@news.wam.umd.edu> <66thpq$e4d@mtinsc04.worldnet.att.net> <66u25b$30i@mtinsc02.worldnet.att.net> <66ulbh$22l@bgtnsc03.worldnet.att.net> <iraAEKAV0Bl0EwWA@james-bannon.demon.co.uk> <671irb$3jk@bgtnsc01.worldnet.att.net>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 1036bc,aad589958d8c66fd X-Google-Attributes: gid1036bc,public X-Google-Thread: 1045b9,aad589958d8c66fd X-Google-Attributes: gid1045b9,public X-Google-Thread: 109e38,aad589958d8c66fd X-Google-Attributes: gid109e38,public X-Google-Thread: f4d7c,aad589958d8c66fd X-Google-Attributes: gidf4d7c,public X-Google-Thread: f5f89,aad589958d8c66fd X-Google-Attributes: gidf5f89,public From: James Bannon <jab@james-bannon.demon.co.uk> Subject: Re: Is it possible to actually get a 16 MB/sec data transfer rate w/ an EIDE drive? Date: 1997/12/14 Message-ID: <n$PBRHAAyFl0EwX7@james-bannon.demon.co.uk> X-Deja-AN: 298187299 Distribution: world X-NNTP-Posting-Host: james-bannon.demon.co.uk [193.237.58.159] References: <348f685a.259490677@news.ebtech.net> <881848354snz@michaels.demon.co.uk> <34913d16.2409107@news.wam.umd.edu> <66thpq$e4d@mtinsc04.worldnet.att.net> <66u25b$30i@mtinsc02.worldnet.att.net> <66ulbh$22l@bgtnsc03.worldnet.att.net> <iraAEKAV0Bl0EwWA@james-bannon.demon.co.uk> <671irb$3jk@bgtnsc01.worldnet.att.net> Organization: Home Machine MIME-Version: 1.0 Newsgroups: comp.sys.ibm.pc.hardware.storage,comp.sys.ibm.pc.hardware.misc,comp.sys.ibm.pc.hardware.chips,demon.tech.pc,alt.comp.hardware