From: Javier H. Seen
Subject: A Problem with *LISP on the Connection Machine
Date: 
Message-ID: <kr04koINN519@paducah.cs.utexas.edu>
Hi.  For those who are parallel programming gurus on CM:

After including the *LISP package to be used at the top level,
I tried to initialize and clean the memory area by invoking the
primitive function *COLD-BOOT, first without indicating the number
of processes desired (I wanted to start with the default values).
An error message pops out saying that "No sequencers are available".
Then I tried to attach the front end to the physical size, say

   (cm:attach :16k)

and still the same error message appears.  One of the debugging options
was to "retry allocating 16384 processors until success".  I try that
one out, and it sits there, like going to sleep.  Furthermore, I can't
even quit the interactive interpreter, coz I think the loading wasn't
successful (it doesn't even understand the QUIT function).

My program runs fine in the *LISP simulator (command STARLISP-SIMULATOR),
but when I port it to the real stuff, I can't even try things out.

Please e-mail if you either work (or have worked) with *LISP on the 
Connection Machines before and/or knows about this problem common among
neophytes (like myself)...  Thanks in advance.

  --Javier H. Seen
    ······@cs.utexas.edu