Newsgroups: comp.sys.transputer
From: k3mi@jupiter.unb.ca (Brian d'Auriol)
Subject: Re: transputer (t8) process activation
Organization: University of New Brunswick, Fredericton, NB, Canada
Date: Thu, 10 Mar 1994 13:59:26 GMT
Message-ID: <1994Mar10.135926.8905@jupiter.sun.csd.unb.ca>

In article <CMEw4r.3MG@wizzy.com> andyr@wizzy.com writes:
>In article <1994Mar9.144239.9230@jupiter.sun.csd.unb.ca>,
>Brian d'Auriol <k3mi@jupiter.sun.csd.unb.ca> wrote:
>>hello all,
>>i am trying to figure out how the process activation occurs on the
>>transputer (t8).
>After this you define some PROCs, but do not say how they are
>instanced.
>
>As any transputer introduction will tell you, any output MUST
>have a parallel input, otherwise the system will deadlock.

*smile*  i have, for the sake of brevity, left out what i consider
the obvious occam syntax.  I do, in fact, have a number of programs
(working correctly) based on the above scheme.  yes, the procs are
instantiated as follows:  each proc is encased as a linked unit 
where each linked unit is specifically placed onto a specified transputer
in the CONFIG section of the .pgm file within a parallel construct.  
Therefore, all *four* of my procedures (in my earlier example) are
executing in *parallel* on the *same* transputer.  From my
studies of the architecture, the PAR outputs should spawn a new process
for each, therefore i should expect 6 parallel processes.  Now the
parent of the output process will block pending completion of the 
outputs, my question is: how does the transputer (via the queuing
system) regulate which of the potential 6 par processes execute.

Brian d'Auriol
ZZ

