Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!ispd-newsserver!rpi!uupsi!tate!cy From: cy@dbase.a-t.com (Cy Shuster) Newsgroups: comp.sys.ibm.pc.misc Subject: Re: Sample CONFIG.SYS file for DOS 5.0, QEMM 5.1x and 4DOS Message-ID: <1991Jun19.183640.2031@dbase.a-t.com> Date: 19 Jun 91 18:36:40 GMT References: <1991Jun14.070037.17625@nntp-server.caltech.edu> Organization: Ashton-Tate, Inc. Lines: 25 Originator: cy@dbase Nntp-Posting-Host: dbase In article halcyon!ralphs@seattleu.edu (Ralph Sims) writes: >nyet@nntp-server.caltech.edu (n liu) writes: > >> >BUFFERS=4 > >> I assume that by setting the buffers,fcbs, and stack vars so small you can >> save a couple of K. How low is "safe" for most uses (windows, mathematica, >> etc. etc.). Do i just set 'em all real tiny and wait for an app to bitch abou >> not having enough? or is that inviting problems? > >We had set up low DOS buffers to allow maximum efficiency in a cache. >Most of our applications use dBASE III+, which DEMANDS (it appears) >at least 8 DOS buffers. For some strange reason, even allocating more >buffers via LOADHI didn't work. We were plagued with scrambled databases, >which required major surgery to resuscitate. Since setting BUFFERS=10 >and sacrificing a tad in cache performance, I don't have to spend a lot >of time playing with the tape backup. III+ was written about a million years ago. dBASE IV 1.1 includes its own cache now, and only requires a 450K footprint (compared to 516K for dBASE IV 1.0). --Cy-- cy@dbase.a-t.com