SCOUG-HELP Mailing List Archives
Return to [ 17 | 
September | 
2003 ]
<< Previous Message << 
 
 
 
=====================================================  
If you are responding to someone asking for help who  
may not be a member of this list, be sure to use the  
REPLY TO ALL feature of your email program.  
=====================================================  
 
 
Saved from a Usenet post, now six years old ....  
I've not tried it.  You're on your own here....  
 
================================================================================  
From: mdemooij@xs4all.nl (M.M.L. de Mooij)  
Newsgroups: comp.os.os2.setup.storage  
Subject: Re: HPFS386  
Date: Tue, 11 Feb 1997 00:34:13 GMT  
Message-ID: <3304be7f.43092735@news1.xs4all.nl>  
References: <5cbkdb$l85@netnews.ntu.edu.tw>  
Reply-To: mdemooij@xs4all.nl  
 
On 25 Jan 1997 00:33:47 GMT, jplin@tvlsi (Jeng-Ping Lin) wrote:  
>I've got an I/O intensive program to run. The current HPFS shipped  
>with Warp 4 still has the limit 2MB to cache data. I remember  
>someone has posted before that HPFS386 in Lan Server could be used  
>in Warp and has the capability to cache data more than 2MB.   
>Unfortunately, I lost the data I collected. Could someone be kind  
>enough to instruct me where to find the info?  
>  
>--  
 
HPFS386 howto  
 
It took some tinkering, but here goes:  
 
1. get either LS30HPFS.ZIP or LS40HPFS.ZIP. The LAN Server 3.0 version  
is easier to install and does not require an HPFS386.INI. Speed  
differences are marginal anyway (law of diminishing returns in action,  
keep that law in mind when you set your cache size). Both versions  
seem (I repeat SEEM) to work well in Warp 4.0, both in a stanalone  
situation as well as a networked (peer) situation. As is always  
suggested and never done, backup before getting into this kind of  
tweaking. I did warn you in time.  
 
I continue on the assumption you got LS30HPFS, which is on Hobbes, I  
think. I have the 12/12/95 version.  
 
2. After unzipping and LOADDSKF-ing on a Super Virtual Disk, no time  
wasted on floppy disks, you get four files that count: HPFS386.IFS,  
CACHE386.EXE and two .MSG files.  
 
3. Create C:\IBM386FS directory and put in the .IFS and .MSG files.  
Put CACHE386.EXE into C:\IBMLAN, on the assumption that's your LANroot  
directory. Copy the file HPFS200.386 from there into the IBM386FS dir.  
 
4. From the .MSG and the .INF files, you can easily deduce the  
parameters you need or want.  
 
5. Use lazy write on the cache.  
 
6. Add these two lines to your CONFIG.SYS BEFORE (I repeat BEFORE) the  
existing HPFS.IFS line.  
 
IFS=C:\IBM386FS\HPFS.IFS C:\IBM386FS\HPFS200.386 [+ other options]  
RUN=C:\IBMLAN\CACHE386.EXE /lazy:on  
 
7. Do not REM the regular HPFS line yet, as it functions as a safety  
catch: should HPFS386 fail for some reason, the normal HPFS will  
start, providing acces to your harddisk (which is definitely useful in  
such a situation).  
 
8. Reboot, check and enjoy. Now you can REM.  
 
On the following notes, please correct me if I'm wrong, as these  
observations are a mixture of fact and assumption.  
 
Afterthoughts  
-  if you are not going to use more than 2 Mb of cache, it may not be  
worth the hassle or that insecure feeling when tampering with such  
central issues. Keep those boot diskettes at hand.  
- if you can afford something like 8Mb or even 12 for the cache, you  
should really notice the change.  
- I don't know what HPFS200.386 is good for, please enlighten me.  
- Cache eats away RAM for applications. If you run disk-intensive  
programs, you'll love the change. If you run RAM hoggers, that's  
another matter. If you have RAM in quantity, it won't hurt.  
 
Interesting:  
- as the HPFS386 supposedly is a ring 0 driver, versus ring 3 for  
HPFS, the 386 version is better in principle.  
- even better: as a ring 0 driver with built-in SMB-networking, there  
are definite advantages when used in an SMB network environment (i.e.  
Win95 and Warp peer networking). The network disk access passes by the  
ring 3 stuff, which would otherwise control disk i/o on the Warp peer.  
So you gain on two accounts: less CPU work in your OS/2 peer and  
faster disk access from the Win95 peer.  
- Speaking of which, check your NIC's for full duplexing. Make sure  
you use it if your NIC's support it.  
- I installed HPFS386 in conjunction with the Intel BusMastering  
PIIXide (triton) IDE driver (v.2.06) which itself proved a speed gain,  
as it was easier to configure for speed than the standard IBM1s506.ADD  
driver.  
 
It works for me. MAYBE it works for you, too.  
 
As disk i/o is not Warp's best feature, any speed gains are welcome.  
You may not like this, but even with HPFS386 installed, Win95 still  
feels more responsive. Ah well, what can you do.  
 
Cheers/2 you all.  
 
M.M.L. de Mooij  
mdemooij@xs4all.nl  
 
 
 
 
 
=====================================================  
 
To unsubscribe from this list, send an email message  
to "steward@scoug.com". In the body of the message,  
put the command "unsubscribe scoug-help".  
 
For problems, contact the list owner at  
"rollin@scoug.com".  
 
=====================================================  
 
  
<< Previous Message << 
Return to [ 17 | 
September | 
2003 ] 
  
  
The Southern California OS/2 User Group
 P.O. Box 26904
 Santa Ana, CA  92799-6904, USA
Copyright 2001 the Southern California OS/2 User Group.  ALL RIGHTS 
RESERVED. 
 
SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group.
OS/2, Workplace Shell, and IBM are registered trademarks of International 
Business Machines Corporation.
All other trademarks remain the property of their respective owners.
 
 |