X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f From: Hans-Bernhard Broeker Newsgroups: comp.os.msdos.djgpp Subject: Re: Ramdrive for developers (not only...) Date: 7 Mar 2004 23:03:34 GMT Organization: Aachen University of Technology (RWTH) Lines: 36 Message-ID: References: NNTP-Posting-Host: accip02.physik.rwth-aachen.de X-Trace: nets3.rz.RWTH-Aachen.DE 1078700614 4841 137.226.33.41 (7 Mar 2004 23:03:34 GMT) X-Complaints-To: abuse AT rwth-aachen DOT de NNTP-Posting-Date: 7 Mar 2004 23:03:34 GMT To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com Carlo wrote: > Hello everyone, > many of us know that a ram drive could improve a lot the compiler > speed. That "could" is now mostly an expression of past tense, not of a hypothesis. A sufficiently large cache with write-back mode enabled is almost exactly as good as a ram-drive, but a heck of a lot more flexible in usage. > In my opinion it would be better that ram drive use memory only when > it's required. But unfortunately that's not something the design MS-DOS can accomodate without major hassles. MS-DOS doesn't expect or even tolerate drives that change size during normal operation. > At the moment, all existing ramdrives allocate the memory and then his > usage is reserved. Wrong. Resizable ones have existed for a long time, but have always had various quirks and pitfalls, and it's quite sure none of them can ever work under a Windows system, which treats all DOS programs like second-class citizens. > But what about using only the exact memory required for storing files? > When files are deleted, that memory could be released, so our programs > can use it. That's quite exactly how a well-designed disk cache already works. Even the ones built into various versions of Windows appear to get that roughly right. -- Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de) Even if all the snow were burnt, ashes would remain.