Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!hubcap!roland
From: roland@sics.se (Roland Karlsson)
Newsgroups: comp.parallel
Subject: Re: Is message passing more efficient than shared mem ?
Message-ID: <12328@hubcap.clemson.edu>
Date: 18 Dec 90 13:52:49 GMT
Sender: fpst@hubcap.clemson.edu
Organization: Swedish Institute of Computer Science, Kista
Lines: 32
Approved: parallel@hubcap.clemson.edu
In-Reply-To: gdheinze@medusainformatik.uni-erlangen.de's message of 17 Dec 90 13:55:51 GMT

Hi Gerhard,

> In the Proceedings of the ICPP 1990 I found a very interesting paper
> suggesting that programs based on message passing are more efficient
> than those based on shared memory - even on shared memory
> architectures.

I hope that neither you nor the authors in the article get offended,
but this is nonsense.  I assume you have simplified the results from
the article.  You can not compare two different ways of implementation
without knowing what to implement.  I have seen several examples of
programs that have been improved when moved from a loosely coupled
machine to a shared memory machine.  The usual trick is to replace
message passing with direct access of other processes memory.  There
are lots of references in the area of parallel simulation.  Here is
one: "Time Warp on a Shared Memory Multiprocessor - Richard M.
Fujimoto - Technical report - CSD - University of Utah".  One more
example: I am working on a very efficient implementation of
Or-parallel Prolog.  There is no way of doing this, with the same
efficiency, with message passing.  There are no messages to transmit.
The Or-parallel branches are independent but reuse already created
(read: lots of) data.  Message passing is a very appropriate method
for many applications.  The main advantage is the possibility to
create issolated objects.  This simplifies the design task.  But
efficiency on a shared memory multiprocessor is not the main issue.



--
Roland Karlsson
SICS, PO Box 1263, S-164 28 KISTA, SWEDEN	Internet: roland@sics.se
Tel: +46 8 752 15 40	Ttx: 812 61 54 SICS S	Fax: +46 8 751 72 30