Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!watmath!clyde!burl!ulysses!allegra!princeton!caip!topaz!uwvax!uwmacc!dubois From: dubois@uwmacc.UUCP (Paul DuBois) Newsgroups: net.micro.mac Subject: Lightspeed and Ramdisk with 512K Mac Message-ID: <32@uwmacc.UUCP> Date: Thu, 26-Jun-86 13:58:39 EDT Article-I.D.: uwmacc.32 Posted: Thu Jun 26 13:58:39 1986 Date-Received: Sun, 29-Jun-86 05:16:06 EDT Distribution: net Organization: UW-Madison Primate Center Lines: 44 A short while back, there was a query about using LightspeedC with a Ram disk on a 512K Mac. The problem is that you can't really do anything interesting with LS *and* the system in a ramdisk - in fact, you can't even run it. You can run with one or the other in the ramdisk, but: with just LS in the rd you still can't really do much because it's *just* big enough to give itself memory problems on a compile. Things go ok with just the system file in the rd. I conjectured that perhaps one could hack out *some* of LS and put it in the system file. This would allow one to put the system in the rd and also *part* of LS, the idea being to push up closer to the critical point. Having said that, I'm not sure exactly what the critical point *is*, but I can say that its worthwhile to do the following. What I do: I used ResEdit to put CODE 10 and CODE 12 into the system that gets booted into the ramdisk. (I'll bet *that* violates the fascist licensing agreement!! :-) Then I deleted CODE 10 and 12 from the copy of LS that I launch (If you leave them in, then the resource manager will find the ones in LS itself, making the ones in the system file useless). I boot up, eject the boot disk, run the ramdisk's minifinder (to make sure the system being used is the one on the ramdisk, yes you need to do this!), put in the LS disk and launch LS. It works. I chose CODE 10 and 12 because they are of pretty good size, so a fairly hefty chunk of code is swapping from ram, not disk. I end up with a ramdisk of 120K. This is big enough to prevent compiling large files, but the project mechanism is sufficiently nice that you shouldn't be doing that anyway! (...he says, experiencing cognitive dissonance...) My question: does anybody know what kind of code is in each segment? I chose 10 and 12 on the basis of size only - it would be nice to load in seqments that correspond to whatever activities I do most often. I'd rather not have editor segments in ram if I'm compiling all the time, for instance. -- Paul DuBois UUCP: {allegra,ihnp4,seismo}!uwvax!uwmacc!dubois | ARPA: dubois@easter --+-- | "Well, we can't give an award to a _dead_ pig," said | the loud speaker. "It's never been done."