Re: Memory management by Share[]?
- To: mathgroup at yoda.physics.unc.edu
- Subject: Re: Memory management by Share[]?
- From: Boss Man <BOSS at noaapmel.gov>
- Date: Fri, 9 Oct 1992 17:51 PDT
>Warren Wiscombe writes: > Memory is a severe constraint when running Mma on a MacIIfx, as I do. >Mostly I just watch the memory bar creep toward maximum, trying to >understand how my commands cause its motion, but with little understanding. > So I was interested to see, in sec. 2.11.5 of The Bible, that there are a >couple of methods to exert user control. I have questions about Share[]. > Share[] seems like a command you would want to execute in your init.m >file, since it promises to significantly reduce memory requirements; in >fact you would expect it to be built into the kernel and not under user >control. Thus it must be dangerous; indeed there is a remark about "the >sharing of subexpressions is often destroyed as soon as you use an operator >like /.". Does anyone have any experience with Share[]? Why doesn't >everyone use it? I am running on a Quadra 700 with 20MB at work and an SE/30 with 8MB at home. I have overcome a certain amount of memory limitation by using the virtual memory feature of System 7. To turn it on, go into the "memory" control panel and check the radio button to turn virtual memory on. Then click in the amount of swap file space to use. As a rule of thumb give yourself half again as much as your real memory figure. I have 8MB of real memory so I configure the swap file for 12MB. (Of course this requires that you have 12MB or whatever of free disk space.) Then reboot. Then give Mathematica more room using get info in the file menu. I think I am giving it 6MB. You can give it as much room as you need as long as you leave about 2.5 MB for system software. This will carry a performance penalty, but at least you'll be able to do calculations you couldn't do before. Good luck, Ed Boss NOAA/PMEL