[comp.benchmarks] Query about non-numerical applications

angela@Lessing.first.gmd.de (Angela Sodan) (03/07/91)

Thanks to all those who replied to my query about non-numerical applications.
There have been only a few answers referring to concrete applications.
Below follows a summary of them.

Angela

_______________________________________________________________________________

From: Anurag Acharya <unido!DRAVIDO.SOAR.CS.CMU.EDU!acha>

Try Common Lisp OPS5 implementation available from carnegie mellon. there
are a small number of benchamrks available too. let me know if you are
interested.

anurag

_______________________________________________________________________________

Cc: jap@maths.bath.ac.uk

A student of Paul Hilfinger at Berekeley was collecting parallel Lisp
applications a little while ago.  He also contributed one of his
own--some kind of expert system I think.  His name, as I recall, was
Kinson Ho, so you could try sending mail to ho@berkeley.edu.

As you probably know, our preferred dialect here is EuLisp, for which
we now have a reasonably portable implementation for multiprocessors,
with a base written in C and a Lisp to C compiler.  We only have
relatively small parallel programs at present in order to test out the
concurrency features.  However, we do have a large simulation system
called PSE (Persistent Simulation Environment ) which is a port of the
Lisp part (originally Common Lisp) to EuLisp of the Picasso+Postgres
system from Berkeley.  Within this system we have persistence coupled
with event, process and Petri net based simulation facilities.  The
next significant step for this system is to insert parallel constructs
(manually).

--Julian.

_______________________________________________________________________________


Cc: priol@irisa.fr (Thierry Priol)

We have developped a SVM, named KOAN, on an iPSC/2 hypercube in order to
parallelize a non-numerical application in image synthesis (ray-tracing).
The ray-tracing algorithm (VM_pRAY) is in public domain. We have also several
numerical algorithms running with KOAN. All these algorithms are written
in C language.

Thierry

_______________________________________________________________________________


*** please, use mail address provided below, do not apply return path ***
------------------------------------------------------------------------------
Angela Sodan                              angela@gmdtub.uucp
GMD Berlin                                ...!uunet!unido!gmdtub!angela
Hardenbergplatz 2 
D-1000 Berlin 12 
Germany
phone (49) (30) 25499 275 
fax   (49) (30) 25499 172
------------------------------------------------------------------------------