home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!yoyo.aarnet.edu.au!frodo.cc.flinders.edu.au!hobson.cc.flinders.edu.au!mmtl
- From: mmtl@hobson.cc.flinders.edu.au (T.Lampre)
- Newsgroups: comp.os.os2.misc
- Subject: Re: Speed OS/2 vrs. Windoze. sx slugs
- Message-ID: <1992Jul23.005509.51572@frodo.cc.flinders.edu.au>
- Date: 23 Jul 92 00:55:09 GMT
- References: <e04e9cd4@garfield.catt.ncsu.edu> <6515150d@p4.f302.n242.z2.fidonet.org> <1992Jul21.221700.585@mala.bc.ca>
- Sender: @frodo.cc.flinders.edu.au
- Organization: Flinders University
- Lines: 69
-
- In article <1992Jul21.221700.585@mala.bc.ca> bigras@mala.bc.ca writes:
- >In article <6515150d@p4.f302.n242.z2.fidonet.org>, Hauke_Hess@p4.f302.n242.z2.fidonet.org (Hauke Hess) writes:
- >> > OS/2 is also showing poor performance on my 386SX 16 with 6 megs.
- >> > Using the 386test burn-in testing program in vanilla DOS 5 and OS/2
- >> > with a single, full-screen DOS window yielded the following:
- >>
- >> > Calculations/sec (with turbo) | Calculations/sec (turbo off)
- >> > DOS 5 | 9980 | 6595
- >> > OS/2 2.0 | 1716 | 1469
- >>
-
- Under what DOS settings? Is it getting accurate info from the clock?
- These questions have been asked before but I still haven't seen any
- answer.
-
- [stuff deleted]
-
- >> Noone noticed until now, and if you think a little longer about
- >> your hardware configuration, you might notice that you used a 386SX,
- >> that means: a 16Bit CPU. Yes, 16Bit, try to imagine what happens if
- >> you do a context switch on a 32Bit architecture that is emulated by a
- >> 16Bit Bus. It really halves performance (if not more). And if OS/2
- >> works in the background on the fly on a real 32Bit Bus, it crawls on
- >> 16Bit. Try the same on a 386DX and post the benchmark once more.
- >> Perhaps you will be surprised (perhaps we will be surpries, but I
- >> don't believe in this)
-
- I've had OS/2 running on 386sx/16 machines through to 486/50's and the
- bus width makes little difference. I can't tell much difference
- between the 386sx/16 and the 386dx/16 that I've tested.
-
- >
- >I use a 386/40 with 8mb / trident SVGA at home, bought just
- >for 2.0
- >
-
- Good setup.
-
- >At work I use a 16MHz sx with 5mb and I run the HPFS. The
- ^^^^^^^^^^^^^^^^^^^^^^
-
- This is what makes the difference. 5mb with HPFS? What was the
- commandment? 'Thou shalt not run HPFS with less than 6mb of RAM'.
- Actually I wouldn't run it with less than 8. (I wouldn't run OS/2 at
- all with less than 8mb to be completely honest)
-
- >speed to load something is ugly, but I just leave it running
- >all the time {several weeks MTTReboot} and that helps
- >a lot as do the 720K multiple dos sessions I use for
- >Novell network administration { multi servers multi windows }
- >and the couple of connections to our VAX { through the net }
- >that come in handy for mail and the internet.
- >
- >This system , slow though it is , is far more productive than
- >win3.1 was on the system, and I don't think that
- >Windows_Next_Turkey will quite install on that system.
- >
- >So if you can't stand the speed of your sx slug just spend
- >$300 and get a new 386/40 motherboard, it will transform
- >your system.
- >
-
- Spend less on a few more meg of RAM. If you're still not happy then
- you can always get a new motherboard later and the money on RAM won't
- have been wasted.
-
- Trevor Lampre
- MIS Unit
- Flinders University of South Australia
-