Newsgroups: comp.parallel.pvm
From: ismisa@wi.uni-muenster.de (Mika Saastamoinen)
Subject: Re: Invisible processes?!?!? (pvm_spawn and stuff)
Organization: Wirtschaftsinformatik Uni Muenster
Date: Wed, 11 Jan 1995 14:16:59 UNDEFINED
Message-ID: <ismisa.12.000B544C@wi.uni-muenster.de>

In article <ismisa.11.001E7911@wi.uni-muenster.de> I wrote:

>Hello.

Yes, hello again.

>...[snip]... They're created (or are 
>they?) and they can be told to do things (I think), but they don't show up 
>anywhere (not that I could find out) and they never send anything back to the 
>master (that's for sure)... [snap]...

... All that because those slaves were "guruing" (to use an Amiga term, 
"dumped core", "crashed", "kicked the bucket" to use other terms... :) due to 
a stupid stupid stupid stupid bug in the slave code. Given certain parameters 
the slave would start writing (or trying to do so anyway) outside the 
allocated table space... Hmmpph... :) So it wasn't a memory problem as 
was suggested to me (thanks), but a "Stupid Programmer(tm)" problem... Well, 
I've fixed it now and the master program and its slaves works beautifully 
now (nice Mandelbrots! :). 

-MiS-
--
Mr. Mika Saastamoinen (ismisa@wi.uni-muenster.de || mikasaas@utu.fi)
        Comes from Finland, the Funland, but is currently an exchange 
        student in WWU Mnster in Germany. For more information, see
        http://www.utu.fi/~mikasaas



