Article View: comp.sys.ibm.pc.misc
Article #1022Re: Optimizing 386SX-16's for Windows
From: Randy Spangler
Date: Wed, 12 Aug 1992 21:51
Date: Wed, 12 Aug 1992 21:51
43 lines
2541 bytes
2541 bytes
In article <1992Aug12.121839.12941@uni-paderborn.de> hosp@uni-paderborn.de (Holger Spangardt) writes: %Though Windows is an graphic-oriented-shell the speed of the %graphiccard is not important, a Herculescard might be faster than a %vga-card, but a colorscreen makes more fun working with Windows.(of %course expensive) The speed of the graphics card is very important, in certain situations. 1) CASE WHERE IT ISN'T: On my 386DX-16, going from a Paradise Pro (mid-speed non-coprocessed) to a Video-7 VRAMII (high-speed non-coprocessed) only got me a 10% speed increase (even though the card could theoretically accept data 700% faster than the Pro). The problem here is that the 386 simply couldn't render the graphics fast enough. 2) CASE WHERE IT IS: Using an accelerated (around $200) or coprocessed ($500+) video card takes some of the load of calculating graphics away from the 386, allowing it to do other things with its time. This WILL provide a substantial speed increase. %The best way to tune the system with Windows is to buy as much RAM as %possible. A Cachecontroller makes the system much faster,too, but %this is not very cheap. Holger is correct that Windows runs much faster with more memory. I consider 4MB the bare minimum. And at $30 or so per MB, it's a pretty cheap solution. If you CAN add a RAM cache to your 386, it will probably give you a speed boost of 40-50 percent. But then again, for the same $200 or $300 you could get a 386DX-25 motherboard, which would give you an even greater speed gain. -------------------------------------------------------------------------- | Randy Spangler | Get your mind out of the gutter | | rspangle@tazdevil.llnl.gov | you're blocking my periscope | --------------------------------------------------------------------------
Technical Details
Message-ID:
<133151@lll-winken.LLNL.GOV>Path:
arc1-ca.pugleaf.net!sparky!uunet!icd.ab.com!usenet.ins.cwru.edu!magnus.acs.ohio-state.edu!cis.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!pacbell.com!iggy.GW.Vitalink.COM!lll-winken!tazdevil!rspangleReferences:
<Bsu89r.3K6@cs.dal.ca> <1992Aug12.121839.12941@uni-paderborn.de>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!icd.ab.com!usenet.ins.cwru.edu!magnus.acs.ohio-state.edu!cis.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!pacbell.com!iggy.GW.Vitalink.COM!lll-winken!tazdevil!rspangle From: rspangle@tazdevil.llnl.gov (Randy Spangler) Newsgroups: comp.sys.ibm.pc.misc Subject: Re: Optimizing 386SX-16's for Windows Message-ID: <133151@lll-winken.LLNL.GOV> Date: 12 Aug 92 21:51:11 GMT References: <Bsu89r.3K6@cs.dal.ca> <1992Aug12.121839.12941@uni-paderborn.de> Sender: usenet@lll-winken.LLNL.GOV Distribution: na Organization: Lawrence Livermore National Laboratory Lines: 43 Nntp-Posting-Host: tazdevil.llnl.gov