home *** CD-ROM | disk | FTP | other *** search
- Path: peer-news.britain.eu.net!uknet!str-ccsun!not-for-mail
- From: nbc@vulture.dmem.strath.ac.uk (Neil Brendan Clark)
- Newsgroups: comp.unix.amiga,comp.os.linux.m68k
- Subject: Re: PROBLEMS: IDE HD's with Linux
- Date: 9 Feb 1996 16:42:11 -0000
- Organization: University of Strathclyde
- Message-ID: <4ffth3$3g3@vulture.dmem.strath.ac.uk>
- References: <1771.6606T1007T296@stack.urc.tue.nl>
- NNTP-Posting-Host: vulture.dmem.strath.ac.uk
-
- E. Durmin <eka@stack.urc.tue.nl> wrote:
- >I just have 4 Mb fast and 2 Mb chip... this means that I can't use X-Windows...
- >Why can't I use X when I have virtual memory? Is it because not evrything
- >can be swapped to VM or is it just because it is tooo sloow?
-
- I used to use X on my 4MB fast 1200 before I acquired a 16 megabyte SIMM. Does
- this tell you something? The problem is twofold really. As you state, certain
- parts of the system cannot be swapped out. These include the kernel, and
- pages of RAM that are "wired" for several possible reasons, such as DMA
- transfers. If I remember correctly, approximately 2M of RAM was lcoked on
- a 4M fast machine - i.e. you have about 2M left to play with. Not good.
-
- The reason X is so bad is that there is something called the "working set"
- of pages for a process(es). This is akin to the number of pages that need
- to be present at a given moment of time in real RAM to avoid swapping to
- disk. If you exceed this, you will be constantly swapping to disk. This
- in fact happens - with two rxvts (small xterms) using twm, moving the mouse
- between the two would cause the system to swap.
-
- Other issues arise too; you cannot (seem) to use the disk cache to any great
- effect either with small amounts of RAM, so bang goes more performance.
-
- AFAIK, with BSD systems anyway, 4M is fine without X, 8M is OKish with X,
- 16M *flys* with X. Although I hear that Linux is a bit less demanding on
- RAM than NetBSD.
-
- --
- "I have trouble imagining death at that income level" - White Noise, D.Delillo
-
- Neil Clark, Transparent Telepresence Group
- http://telepresence.dmem.strath.ac.uk
-