Article View: comp.sys.ibm.pc.hardware.storage
Article #1189Re: Rocket versus 32-bit file access
From: Peter Herweijer
Date: Fri, 11 Nov 1994 13:20
Date: Fri, 11 Nov 1994 13:20
43 lines
2845 bytes
2845 bytes
jsteele@insyte.com (John Steele) writes: >BJSpace@ix.netcom.com (B. E. Johnson) writes: > >>In <674529809wnr@rosebery.demon.co.uk> Brian@rosebery.demon.co.uk (Brian >>Sheehan) writes: > >>>I'm currently trying to decide to establish my windows 3.11 >>>setup with either 32 bit file and disk acccess enabled or >>>to use the utility Rocket. You can use 32 bit file access with Drive Rocket just fine. >>> Under DOS I'm certainly getting >>>better performance but under Windows the performance seems >>>almost identical (I'm using the Western Digital driver for >>>my two hard disks) [...] This is not surprising, as the WD driver does the same thing for Windows that Rocket does for DOS, that is, enabling block transfers. It is very unfortunate that the WD driver won't initialize with Rocket present. >>My jury is still out on DR but the 'feel' you get that the machine is >>faster with 32-bit enabled is real and has little to do with the >>transfer rate. What is making the big difference is that with standard >>Windows the Windows system drops repeatedly into real mode to do disk >>transfers and then back up into 386 enhanced after each file is read in. >[snip] This is certainly true. Normally, Rocket compensates this effect by making your disk faster in the first place, but if you compare it to Windows with WD driver, this advantage is negated because the latter does the same magic Rocket does. >The other factor is that as I recall, vcache is a file cache as opposed to >a track cache -- as a result all, or a significant portion, of a file a read >in to the cache so its available more readily. the track cache approach reads >one or more tracks in, which may or maynot contain the data needed next. But there is no reason to wave VCACHE bye-bye when you use Rocket. - Peter Herweijer pieterh@sci.kun.nl
Technical Details
Message-ID:
<Cz3uEz.EBB@sci.kun.nl>Path:
arc1-ca.pugleaf.net!nntp.gmd.de!xlink.net!news.dfn.de!swiss.ans.net!howland.reston.ans.net!news.sprintlink.net!EU.net!sun4nl!sci.kun.nl!pieterhReferences:
<674529809wnr@rosebery.demon.co.uk> <39sbrt$ogp@ixnews1.ix.netcom.com> <jsteele.31.046286F4@insyte.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 1036bc,41be0babe0d39d00 X-Google-Attributes: gid1036bc,public X-Google-Thread: f4d7c,41be0babe0d39d00 X-Google-Attributes: gidf4d7c,public X-Google-ArrivalTime: 1994-11-12 14:07:10 PST Newsgroups: comp.sys.ibm.pc.hardware.storage,demon.tech.pc Path: arc1-ca.pugleaf.net!nntp.gmd.de!xlink.net!news.dfn.de!swiss.ans.net!howland.reston.ans.net!news.sprintlink.net!EU.net!sun4nl!sci.kun.nl!pieterh From: pieterh@sci.kun.nl (Peter Herweijer) Subject: Re: Rocket versus 32-bit file access Message-ID: <Cz3uEz.EBB@sci.kun.nl> Sender: news@sci.kun.nl (News owner) Nntp-Posting-Host: wn3.sci.kun.nl Organization: University of Nijmegen, The Netherlands X-Newsreader: NN version 6.4.19 References: <674529809wnr@rosebery.demon.co.uk> <39sbrt$ogp@ixnews1.ix.netcom.com> <jsteele.31.046286F4@insyte.com> Date: Fri, 11 Nov 1994 13:20:58 GMT Lines: 43