From tautges@nucst9.neep.wisc.edu Wed Mar 21 14:14:15 1990
Return-Path: <tautges@nucst9.neep.wisc.edu>
Received: from vms.macc.wisc.edu by cs.utk.edu with SMTP (5.61++/2.3-UTK)
	id AA24454; Wed, 21 Mar 90 14:13:44 -0500
Received: from nucst9.neep.wisc.edu by vms.macc.wisc.edu;
          Wed, 21 Mar 90 13:11 CST
Received: by nucst9.neep.wisc.edu (4.1/25) id AA04791;
          Wed, 21 Mar 90 13:09:12 CST
Message-Id: <9003211909.AA04791@nucst9.neep.wisc.edu>
Date: Wed, 21 Mar 90 13:09:12 CST
From: tautges@nucst9.neep.wisc.edu
Subject: SCHEDULE for Sparcstation
To: dongarra@dasher.cs.utk.edu
Status: RO

Jack,
  The next message will be a share file containing SCHEDULE for the
Sparcstation 1.  It uses the brute force method to pass fortran arguments
into c structures.  The original version didn't work on the Sparcstation, due
to a change in the c-fortran interface.
  I made a few other changes in the code:
1) The ability to allow more arguments to be passed through SCHED is now
   implemented using #ifdefs with the c preprocessor; this is simpler and
   it doesn't require the user to know anything about c.
2) A single (sorted) ready q can be optionally used, again with the use
   of cpp definitions.  I implemented this for my work, which requires a
   little more control over the task scheduling.  The default is to use
   a Longest Processing Time first algorithm to sort the ready queue, although
   this can be easily changed (see the README file and the function addrq
   in ftsubs.F).

I included only the things that changed in the shar file.  The things that
are not in there are:
makefile
cputm.c
all the test problems and data files

I'm hoping that those already exist in a shar file there and can be patched
in easily.

Regards,

Tim Tautges
Univ. of Wisconsin - Madison





.