From: sinster@scintilla.capitola.ca.us (Darren Senn) Subject: Re: Swapping (Was Re: X386 (Was Re: 0.96 out next week)) Date: 5 May 1992 20:21:31 GMT
Ok, so my subject line is getting too nested... so shoot me. *bang* *thump* :)
In article <1992May5.140604@hammer.Prime.COM>, cummings@hammer.Prime.COM (Kevin Cummings) writes:
> > As to memory: I'm using it in 8MB ram, and no swapping (with a couple of
> > xterms, xclock and xcalc - nothing major). If I want to recompile the
> > kernel in an xterm, I'll have to start up swapping (well, actually I've
> > done it without swapping, but it's tight). I assume it's still useable
> > in 4MB and a big swap-file, but I'm happy I haven't tested it.
>
> As the software gets bigger, swapping will become more and more the norm.
> Maybe we are still afraid of Linux's swap code? Other than that, the only
> wrong with swapping is that it slows things down. C'est la vie.
Just for the record:
I'm using a 486/25 w/ 4Mb of memory and 5Mb of swap with MCC-interim 0.95c+.
Often I run multiple compiles, and this usually makes me swap. I've only
had one problem with swapping, and that shows up a lot: if I've got a compile
that starts and stops under a vc that I'm not viewing, and any swapping
happens, then some of the child processes of the compile will get messed
up, and they will produce the following warning messages after terminating
normally:
Younger sibling x: invalid os link to PID y
Older sibly y: invalid ys link to PID x
(or somesuch).
Other than that, I haven't had any problems related to swapping. This
particular problem could be with the MCC-interim release, or GCC, or 0.95c+,
or virtual consoles, or swapping, or who knows. The point is, that's the
only problem I've encountered, and I swap a lot.