gupta@prl.philips.co.uk (Ashok Gupta) (10/04/90)
HIGH PERFORMANCE
AND
PARALLEL COMPUTING
IN LISP
Nov 12-13 1990
Twickenham, U.K.
A EUROPAL Workshop
Sponsored by
The British Computer Society Parallel Processing Specialist Group
-*-*-*-
Production quality Lisp systems are fast approaching the performance levels of
applications developed in the more mainstream languages. Implementations of
Lisp span the spectrum of available hardware platforms - from p.c.'s to
workstations & dedicated Lisp machines through to multi-processor, parallel
computers.
This conference brings together
researchers working on developments in compiler technology,
hardware design, systems and algorithms
users of high performance Lisp systems in research & industry
and vendors of high performance Lisp systems and applications.
The programme has a truly international flavour with submissions from USA,
Japan, U.K., Germany, France, Israel & South East Asia, and consists of keynote
and invited speakers, contributed papers and a panel session.
Hear about parallel Lisps for a range of SIMD & MIMD architectures including
environments and applications on the Connection Machine, Transputer arrays,
the new 256 processor EDS machine being developed under ESPRIT-II, and more.
Topics covered also include models for parallel computation (Linda and Timewarp)
and their implementation, concurrent OOP, automatic partitioning of Lisp code
for concurrent execution and parallel debugging.
-*-*-*-
Keynote Address
Ten Ideas
Richard Gabriel (Lucid Inc. & Stanford Univ., USA)
-*-*-*-
Session 1 - Distributed Systems
Parallel Lisp for a Distributed Memory Machine
C. Hammer and T. Henties (Siemens AG, W. Germany)
Sharing Data Structures in a Distributed Lisp
J. Piquer (Ecole Polytechnique - INRIA, France)
Experiences with Lisp and Distributed Systems
David DeRoure (Univ. of Southampton, UK)
-*-*-*-
Session 2 - Shared Memory Systems
SYMPAL and its Optimising Compiler
Yariv Aridor & Dr. Shimon Cohen (Tel-Aviv University, Israel)
Scheme-Linda for the Meiko & Symmetry
Ulf Dahlen (Univ. of Linkoping, Sweden)
&
Neil MacDonald (Edinburgh Parallel Computing Centre, UK)
- Title to follow -
Kelly Murray (Top Level Inc., USA)
-*-*-*-
Session 3 - Implementations & Delivery
Generation Scavenging GC on Distributed-Memory Parallel computers
Hanpei Koike and Hidehiko Tanaka (University of Tokyo, Japan)
The Future of Futures (or, How Futures Can Be Implemented Efficiently
on Stock Hardware)
M. Radlhammer (Siemens AG, W. Germany)
Linda Meets Lisp
D. Hutchinson (U. Bath, UK)
Delphi: A Toolkit for Debugging of Parallel Lisp Programs
H. Ilmberger and Dr. S. Thuermel (Siemens AG, W. Germany)
-*-*-*-
Session 4
PolyScheme - Semantics for a Concurrent Scheme
C. Queinnec (Ecole Polytechnique - INRIA, France)
Automatic Partioning for Concurrent Execution
S. Kalogeropulos (U. Bath, UK)
A Data Driven, Direct Execution Architecture for a Parallel Lisp Dialect
W.F. Wong et. al (National University of Singapore)
-*-*-*-
Panel Discussion
How Can Parallel Lisp Succeed in the Market-Place ?
-*-*-*-
Session 6 - Models for Parallel Computing
Implementing Timewarp Using Lisp and Linda
D. Hutchinson & J. Fitch (U. Bath, UK)
Concurrent Object Oriented Programming in Lisp
J. Padget, R. Bradford & J. Fitch (U. Bath, UK)
-*-*-*-
Session 7 - Applications
PARKA : An A.I. Application on the Connection Machine
Matthew Evett & James Hendler (UMCP, USA)
Parallelising an Expert System Application
Jim Aragones & Benjamin Zorn (Boulder, Colorado)
-*-*-*-
Closing Address
What Price, Parallel Lisp ?
George K. Jacob (Franz Inc., USA)
-*-*-*-
Programme Committee :-
Professor John Fitch University of Bath, U.K.
Jeff. Dalton A.I. Applications Institute, University of Edinburgh, UK
Ashok Gupta Philips Research Laboratories, Surrey, U.K.
The delegate fee for this two-day event is :-
250 (pounds sterling) + 15% V.A.T.
[less 150 (pounds sterling) for all registered academics]
A copy of the Proceedings will be issued to all contributors and delegates.
For opportunities of sponsorship and other details please contact the Conference
Organiser :-
David Lloyd
Applied Workstations Ltd.,
The White Cottage, South Holmwood, Dorking, Surrey, U.K., RH5 4NA
Phone:- (+44) (306) 889485 Fax:- (+44) (306) 741293
Or email your postal address to gupta@prl.philips.co.uk, if you would like a
copy of the Conference Brochure to be sent to you.yeh@cs.purdue.EDU (Wei Jen Yeh) (03/05/91)
Hello,
Does anyone know of a parallel lisp? Common lisp flavor is preferred.
And, can anyone provide me info. as to how to call C functions in AKCL?
Thanks in advance for any help.
Wei Jen Yeh yeh@cs.purdue.edu
Department of Computer Science
Purdue University
West Lafayette, Indiana
--
Wei Jen Yeh yeh@cs.purdue.edu
Department of Computer Science
Purdue University
West Lafayette, Indianaelee63@castle.ed.ac.uk (R Hamilton) (03/05/91)
In article <13658@medusa.cs.purdue.edu> yeh@cs.purdue.EDU (Wei Jen Yeh) writes: > Does anyone know of a parallel lisp? Common lisp flavor is preferred.... > I've been looking for a good //el implementation for while. As far as I know there isn't one (I say that in the hope that someone will prove me wrong!) The nearest I've seen (heard of) was a version of Betz's xlisp running on Helios on a transputer array. Robert Hamilton Dept of Electrical Eng. University of Edinburgh. Email: rjha@uk.ac.ed.ee
dourish@EuroPARC.Xerox.COM (Paul Dourish) (03/06/91)
In article <8861@castle.ed.ac.uk>, elee63@castle.ed.ac.uk (R Hamilton) writes: > In article <13658@medusa.cs.purdue.edu> yeh@cs.purdue.EDU (Wei Jen Yeh) writes: > > Does anyone know of a parallel lisp? Common lisp flavor is preferred.... > > > I've been looking for a good //el implementation for while. > As far as I know there isn't one (I say that in the hope that someone > will prove me wrong!) > The nearest I've seen (heard of) was a version of Betz's xlisp > running on Helios on a transputer array. Parallel Lisp is a research area (albeit a fairly active one). By that I mean that, while it's an area of considerable interest and work, it's by no means certain that you're going to find something that will suit your purposes. However, we can do better than transputer-specific systems. Most people working on parallel Lisp address specifically parallel machines, but there are often implementations which time-share on serial machines. There are various possible mechanisms for parallelising Lisp, though, and they differ greatly in approach and applicability. I'm presuming that the original poster wanted to run a parallel Lisp on a serial machine. Thinking Machines have a simulator for *Lisp (as used on the Connection Machine -- see Danny Hillis' book) which is available for anonymous ftp. Other interesting contenders are QLisp (Stanford) and Multilisp (MIT), although last time I looked neither of those were widely available for serial machines. It would be hard to imagine three systems which took more different approaches to parallelising Lisp! An early implementation of EuLisp, a Lisp dialect whose definition is currently being worked on by a CEC-sponsored group, incorporates a thread mechanism, and is freely available (to get a copy, try mailing eudist@maths.bath.ac.uk). It's not entirely Common Lisp-like; whether this is a good or bad thing is a matter of debate. For Rob Hamilton, my own contribution is in the building next door (a Linda-based parallel Lisp running on the Edinburgh Concurrent Supercomputer)! -- Paul. -- Paul Dourish, Rank Xerox EuroPARC, Cambridge, UK <dourish@europarc.xerox.com> "Ain't they got no barbers where you come from, boy?"
dandrews@bilver.uucp (Dave Andrews) (03/06/91)
> Does anyone know of a parallel Lisp?
The one that comes to my mind was the one described in Danny Hillis'
dissertation, "The Connection Machine". They have a list variant
called a "xector", in which each element is processed simultaneously
by a processor on the CM. I don't remember if it is very CL compliant.jpff@maths.bath.ac.uk (John ffitch) (03/06/91)
It is very hard to by-pass this oportunity to talk about EuLISP. EuLISP was designed to provide parallel processing, and there are primitive for just that. The FEEL implementation runs on our shared memory Stardent, and also runs pseudo-parallel on SUN-OS. We believe that porting FEEL is not very hard, so moving to other shared memory machines should be OK. Yes this is a research program and so things change, but the current state etc can be determined by sending e-mail to eudist@uk.ac.bath.maths Code is free (FEEL = Free and Eventually Eu Lisp) ==John ff
af@spice.cs.cmu.edu (Alessandro Forin) (03/07/91)
I have ported Multilisp to a variety of machines running Mach, multiprocessors and not. The best parallel lisp currently is, as far as I can tell, Mul-T from Dave Kranz for the Encore. Both are Scheme-based. Qlisp is/was an extension of Common Lisp for the Alliant-FX, I do not know if it is/was ever made available to people. sandro-
welch@sacral.cis.ohio-state.edu (Arun Welch) (03/07/91)
BBN had/has a parallel CL for their Butterfly, based on Multi-Scheme. A massively wierd beasty, but it works. The reason I say has/had is because they only sold 6 copies of it, and it's now officially unsupported. ...arun ---------------------------------------------------------------------------- Arun Welch Lisp Systems Programmer, Lab for AI Research, Ohio State University welch@cis.ohio-state.edu
ahlenius@motcid.UUCP (Mark Ahlenius) (03/08/91)
Top Level out in Cambridge? MA, has a parallel CL that they sell for a number of multiprocessor platforms. I have only talked with these folks and have not used the product. Contact them @ (413) 256-6405 for more information. The were demo'ing the 2 processor version on the Data General Avion Platform at the AAAI in Boston last summer. They were also working on a few other ports too. 'mark -- =============== regards 'mark ============================================= Mark Ahlenius voice:(708)-632-5346 email: uunet!motcid!ahleniusm Motorola Inc. fax: (708)-632-2413 Arlington, Hts. IL, USA 60004
teskridg@nmsu.edu (Tom Eskridge) (03/09/91)
In article <1991Mar5.215242.12812@bilver.uucp> dandrews@bilver.uucp (Dave Andrews) writes: > Does anyone know of a parallel Lisp? Lucid has a parallel lisp for the sequent called CLiP (Common Lisp in Parallel) which we have here at NMSU. Somewhat buggy, but does have some nice features too. -- tom eskridge computing research laboratory, new mexico state university teskridg@nmsu.edu (505) 646-6247
barmar@think.com (Barry Margolin) (03/09/91)
In article <1991Mar5.215242.12812@bilver.uucp> dandrews@bilver.uucp (Dave Andrews) writes: >The one that comes to my mind was the one described in Danny Hillis' >dissertation, "The Connection Machine". They have a list variant >called a "xector", in which each element is processed simultaneously >by a processor on the CM. I don't remember if it is very CL compliant. You are describing CM Lisp. It is completely CL compliant, because it is simply an orthogonal extension to CL. Unfortunately, it was never implemented fully, as far as I know. CM Lisp was a very ambitious design, because it abstracted a good deal away from the Connection Machine hardware, and expected many Common Lisp primitives to deal with xectors directly. Instead, what we implemented was *Lisp, which is much closer to the CM architecture, and requires programmers to use a separate set of constructs for dealing with "pvars" (parallel variables) and parallel operations. Functions that accept pvars are generally named <name>!! and map the CL function <name> over the elements of the pvars (e.g. +!!); global and control constructs are generally named *<name>, where <name> is the corresponding CL operator (e.g. *IF, *SUM). Gary Sabot developed Paralation Lisp as part of his dissertation. I believe his book "The Paralation Model" includes a disk with a simulator on it. -- Barry Margolin, Thinking Machines Corp. barmar@think.com {uunet,harvard}!think!barmar
DanN@juliet.cs.umass.edu (Dan Neiman) (03/09/91)
Just to correct some information in a previous message...
Top Level, Inc. is located in Amherst, MA. Their phone number is
(413) 549-4455. Their parallel common lisp runs on a number of
multiprocessors including the Sequent Symmetry and the Encore.
The parallel Lisp supports futures and lightweight processes called
threads. The version I'm using (version 1.0) has a debugging
interface which runs on a TI Explorer and allows the programmer to debug
the individual parallel processes. As I understand it, version 2.0
of Top Level Common Lisp will include a X-window based debugger with
debugging and metering software.
I'm using the parallel lisp in my research on parallel production
systems and know the people at Top Level, but am not otherwise
associated with them.
Dan N.ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe) (03/11/91)
In article <1991Mar8.172014.9920@Think.COM>, barmar@think.com (Barry Margolin) writes: > Gary Sabot developed Paralation Lisp as part of his dissertation. I > believe his book "The Paralation Model" includes a disk with a simulator on > it. I have that book and there was no disc in it. -- The purpose of advertising is to destroy the freedom of the market.