POV-Ray : Newsgroups : povray.off-topic : RIP Gary Gygax : Re: RIP Gary Gygax Server Time
11 Oct 2024 05:22:19 EDT (-0400)
  Re: RIP Gary Gygax  
From: Jim Henderson
Date: 11 Mar 2008 17:45:19
Message: <47d70b7f$1@news.povray.org>
On Tue, 11 Mar 2008 12:45:47 -0700, Darren New wrote:

> Warp wrote:
>>   The main point is whether you were able to run X with it or not.
>> To simply use the console you could use a microwave oven. ;)
> 
> I worked on a diskless solaris machine with 4M of RAM at one point,
> swapping over the ethernet. It ran.  It took about 10 minutes to switch
> focus on the windows, but it ran. ;-)

At the university I went to, we had some of these Sun SLC workstations, a 
few 1+s, a 2, and a big rack-mounted system (I forget the model number).  
The rack-mounted system hosted the disk for the diskless stations.

We went through a phase of playing around with the DikuMUD source code, 
and ran a MUD on the Sparc 2 (which was also running Ingres, which is 
relevant because we coded a character in the MUD called 'Ingres' that was 
impervious to attacks.  You could kill it by typing "Look at ingres", and 
that was it - because our experience was so bad with the product, we had 
to take it out on someone.  Nothing like being a god character, creating 
an Ingres character instance, and then looking at it just to watch it 
die....<g>).

We only had about two dozen machines in total, and they were used for 
teaching Ada, MODSIM II, and I think FORTRAN.  We had so many students 
who loved playing the MUD that often times we couldn't free machines up 
for students who had actual homework to do.

I wrote the following program to deal with the problem:

--- snip ---

#include <stdio.h>
#include <stdlib.h>

main()
{
	int bignum = 100000;
	while (malloc(bignum));
}

--- snip ---

(As I recall, the actual code included logic to divide bignum by two when 
it fails and try again - but this will suffice)

We could always tell when it had been run on one of the diskless 
stations, because the rack-mounted system's disks would start thrashing 
like mad.

The really neat thing about malloc'ing all the memory under SunOS 4.2 
(and probably other UNIX OSes) is that if the OS can't allocate memory to 
even open a file handle, the system just sits there - you can't find out 
what happened.  The only way out with those machines was a Stop-A restart.

Ah, the good old days....<g>

Jim


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.