Newsgroups: comp.parallel.pvm
From: john@elroy.jpl.nasa.gov (John Wright)
Subject: Re: pvm_exit - What's the point?
Organization: Image Analysis Systems Group, JPL
Date: 3 Feb 1995 15:57:46 GMT
Message-ID: <3gtjpq$p9v@elroy.jpl.nasa.gov>

In article <3gtgqs$nuo@elroy.jpl.nasa.gov>,
John Wright <john@elroy.jpl.nasa.gov> wrote:
>I have a question which I hope someone can answer.  I call
>pvm_exit at the end of my program.  Immediately before I do
>I call pvm_tasks which reports one task (this one) running.
>Then I call pvm_tasks again after calling pvm_exit and it
>still reports one task (this one) running.  What's the point?
>I even tried sleeping for a couple of seconds in case it
>needed some time to propagate.  Is this a bug?  Thanks for 
>any info you can provide.
>

There is an additional point here which is very irritating.
What I'm trying to do is check if pvm is idle when my task
finishes.  If so, and if my task launched pvm, it halts pvm
so the demons are not left running.  However, if I do a
pvm_exit, then a pvm_halt, it terminates my task which should
not be registered with pvm anymore.  Thus I can't do any of
the other cleanup things I wish to do.  Thanks for any help.

John
john@tone.jpl.nasa.gov



