home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.misc
- Path: news99.sunet.se!liuida!news
- From: c92manen@und.ida.liu.se (Mans Engman)
- Subject: Re: Doubling Memory
- X-Nntp-Posting-Host: astmatix.ida.liu.se
- Message-ID: <1017.6583T1226T2233@und.ida.liu.se>
- Sender: news@ida.liu.se
- Organization: CIS Dept, Linkoping University, Sweden
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP) *UNREGISTERED*
- References: <rdingem.445u@grafix.xs4all.nl> <1356.6580T1113T2606@mbox3.swipnet.se>
- Date: Wed, 10 Jan 1996 20:36:24 GMT
-
-
- [ram compression stuff deleted]
-
- >It simply can't be done on the Amiga! The reason it works on Mac is first
- >of all the bad memory handlings. You have to specify how much memory each
- >program can get on a Mac. And only one program have the CPU, although
- >several can be in the memory. This differs a lot from the Amiga.
-
- Well, pre-emptive multitasking makes things a little harder, but with
- an MMU I don't see any big problems. Paged virtual memory works fine,
- so analogously a page can be *compressed* to a pool instead of swapped
- out. Compression would in many cases be much faster than disk paging
- because of the slow disk access. Perhaps a good approach would be to combine
- the two techniques: When the system is low on memory, "old" (not recently
- used) pages are compressed, and even older, compressed, pages are spooled to
- disk.
-
- >It won't
- >work so well on Win3.11 either and not at all in win95 because they have more
- >complicated memory allocations than mac.
-
- Hmm, this is out of my scope...no comment! :)
-
-