home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!Sirius.dfn.de!chx400!forty2!amiga!cschneid
- From: cschneid@amiga.physik.unizh.ch (Christian Schneider)
- Newsgroups: comp.sys.amiga.applications
- Subject: Re: Virtual Memory, a first test.
- Keywords: VMEM
- Message-ID: <2315@forty2.physik.unizh.ch>
- Date: 30 Jul 92 12:30:29 GMT
- References: <1992Jul28.101448.858@unibi.uni-bielefeld.de>
- Sender: news@forty2.physik.unizh.ch
- Organization: ICU - University of Zurich, Dept. of Comp. Sci.
- Lines: 52
-
- In article <1992Jul28.101448.858@unibi.uni-bielefeld.de> markus@techfak.uni-bielefeld.de (Markus Illenseer) writes:
-
- >AG/Switzerland, thereby written by Christian A. Schneider, Friedeman
- ^ This is obsolete ;-)
-
- >I missed some hints on the use of the real Memory. Duing the test, VMEM
- >was somewhat (real) RAM-greedy, i explained that due to the fact that a
- >MMU-Table for large VMEM-Partitions can be large, too. This is not mentioned.
-
- Facts about memory usage: 5K per 10MB vmem, which leads to a 30K for 'common'
- 60MB virtual memory.
-
- >1 GB space, i thionk the authors didn;t had access to a larger drive,
- >VMEM runs fine with 1.9 GB :)
-
- Since programs might get into troubles if you'd use all that 1.9GB, up to
- 1GB was tested. Explanation: addresses would get in the 0x80000000 range,
- i.e. 'negative' addresses which confuses certain programs comparing pointers.
- This is an application problem, not really a GigaMem one.
-
- >Also i missed hints for possible dangers or problems, e.g. that VMEM doesn't
- >run with 'CPU FASTROM' (needs MMU table itself).
-
- It _does_ work on our test machine. It links an active MMU table to its own
- table, except for certain 'bogus' (very specialized) tables, e.g. Enforcer.
-
- >Under OS 2.0 the mainpart is installed in Sys:wbstartup.
-
- 'Mainpart' is somehow misleading. Only a project icon is copied to WBStartup.
-
- >provided, one may discuss this, i think the way VMEM does it is better.
-
- My opinion as well ;-) (You could have guessed that, couldn't you?)
-
- >As i am not a music-freak, i had no intention to test VMEM with a Soundtracker
- >or Digitiser.
-
- Audiomaster has been tested and works fine. Editing of 100MB samples?
- No prob ;-)
-
- >All in all VMEM was very stable and simple to deal with. The only problem
- >which did acutally occure was a problem with a 'MAsk'-entry in the Mountlist.
- >This was ou only crash in about 10 hours testing.
-
- This has been improved already.
-
- >Markus Illenseer EMail: markus@TechFak.Uni-Bielefeld.de
-
- --
- Chris Schneider - cschneid@amiga.physik.unizh.ch BIX: hschneider IRC: cschneid
- The goal of science is to build better mousetraps. The goal of nature is to
- build better mice. - Human beings were created by water to transport it uphill.
-