simulation@uflorida.cis.ufl.edu (Moderator: Paul Fishwick) (11/20/89)
Volume: 12, Issue: 7, Mon Nov 20 09:39:42 EST 1989
+----------------+
| TODAY'S TOPICS |
+----------------+
(1) Correction about Tech. Reports
(2) Dynamic Parallelism
(3) Survey on Research Center
(4) Ternary Simulation
(5) NewsGroup Discussion: Virtual-Worlds
* Moderator: Paul Fishwick, Univ. of Florida
* Send topical mail to: simulation@bikini.cis.ufl.edu OR
post to comp.simulation via USENET
* Archives available via FTP to bikini.cis.ufl.edu, login as
'ftp', use your last name as the password, change
directory to pub/simdigest.
* Simulation Tools available by doing above and changing the
directory to pub/simdigest/tools.
-----------------------------------------------------------------------------
Date: Fri, 10 Nov 89 08:31:33 -0800
From: liny@cs.washington.edu (Yi-Bing Lin)
Return-Path: <liny@cs.washington.edu>
To: fishwick@bikini.cis.ufl.edu
Subject: Correction
Dear Prof. Fishwick,
Please post this message in the SIMULATION DIGEST. Thanks,
Jason (Yi-Bing Lin)
-------------------------------------------------------------------
I posted the abstracts of two papers in Simulation Digest
(V12, N16). These papers are JUST technical reports submitted for
publication. They are NOT accepted by 1990 Distributed Simulation.
Sorry for the mistake.
Jason Lin
------------------------------
Date: Sat, 11 Nov 89 14:30:29 PST
From: chkim%pollux.usc.edu@usc.edu (Chinhyun Kim)
To: fpst@hubcap.clemson.edu, simulation@bikini.cis.ufl.edu
Subject: Dynamic parallelism
Hi, the following is an article I want to post to your newsgroup.
---------------------------Article starts here---------------------------
In numerical analysis, arrays being indexed by other arrays as shown below
are supposed to be very common in programs involving sparse matrices,
fluid flow simulations, etc. If one wanted to parallelize this loop during
compilation, he/she would have to know the values of B(i) and C(i). If they
are not known at compile time, it will be impossible to parallelize the
operations in the loop and thus must be executed in sequential manner.
There are, however, ways to parallelize the operations in the loop by checking
data dependency during run time and letting operations which are free of data
dependency to execute in parallel. This kind of parallelism which can be
determined only during run time is called, "dynamic parallelism".
A(i) = ..... ; A(i)'s are initialized.
B(i) = .... ; B(i)'s are determined.
C(i) = .... ; C(i)'s are determined.
DO i = 1..N
A(B(i)) = ..... ; Array A is updated.
T(i) = A(C(i)) + .. ; Array A is consumed for some operation.
ENDDO
I am currently working on ways to exploit dynamic parallelism as opposed to
static parallelism which the compiler assesses parallelism during compilation.
The model of computation is data-flow and the specific type of data-flow
computer model is tagged token data-flow computer. I am using a mechanism
known as "token relabeling" to do this.
The results of my simulation so far is good. However, I have some fundamental
questions to ask.
Q1. In exactly what type of numerical simulations do such loops appear and why
are they necessary?
Q2. Are there other ways to do it without having the subscripts unknown at
compile time, i.e. is exploitation of dynamic parallelism avoidable while
still achieving parallel execution?
Q3. Is it possible for me to get the fragment of the simulation program which
has such loop? i.e. assuming answer to the first part of question 1 is
true.
Answers to these questions will be appreciated very much.
Chinhyun Kim
chkim@pollux.usc.edu
------------------------------
From: Robert Cole <cole@cica.indiana.edu>
Date: 15 Nov 89 12:01:39 GMT
To: comp-simulation@rutgers.edu
Subject: Submission for comp-simulation
Responding-System: cica.cica.indiana.edu
Path: cica!cole
From: cole@cica.cica.indiana.edu (Robert Cole)
Newsgroups: comp.graphics,comp.simulation,comp.software-eng
Subject: repost survey - PLEASE help
Date: 15 Nov 89 12:01:38 GMT
Reply-To: cole@cica.indiana.edu (Robert Cole)
Distribution: na
Organization: Indiana University, Bloomington
This is a repost of a message which only went out on comp.graphics
So far I've only received one response (which was excellent!) and
two messages from SIGGRAPH committee members expressing interest
in the survey results. If you are associated with a Center such
as below or have contacts at such centers, please see that this
gets into appropriate hands so that we can get a good selection
of answers. Thanks!
****************************************************
CICA is the Center for Innovative Computer Applications at Indiana
University. As a special projects organization for the Office of
University Computing, CICA has begun a two year commitment to usage
of high-end graphics in the various sciences (such as Chemistry,
Geology, Astronomy, Computer Science, Mathematics) and other
disciplines such as Fine Arts and Theatre at Indiana.
We are in the process of evaluating our existence in relation
to other special university computing centers and trying to
determine what it is that is needed for success as a nationally
recognized center.
Please help us by e-mailing a reply to the following questions.
Your response can be sent to cole@cica.cica.indiana.edu Thank
you for your help...
1. What is a 4-5 sentence mission statement for your organization?
2. What are your main goals?
3. What are your main accomplishments?
4. When was the organization founded?
5. Please fill in the answers to the following based upon the first year
of existence:
a. Number of staff
b. Number of projects undertaken
c. Approximate budget
d. Percentage of budget from institution
e. Percentage of budget from external sources
6. Please fill in the answers to the following based on current levels:
a. Number of staff
b. Number of projects undertaken
c. Approximate budget
d. Percentage of budget from institution
e. Percentage of budget from external sources
7. Any words of wisdom for a fledgling center?
------------------------------
To: comp-simulation@cwi.nl
Path: tuewsd!tuewsk!wstomv
From: wstomv@lso.win.tue.nl (Tom Verhoeff)
Newsgroups: tue.general,nlnet.general,comp.theory,comp.lsi,comp.simulation,sci.electronics
Subject: 2nd NFI seminar by J. A. Brzozowski at EUT
Keywords: Asynchrononous Circuit Theory, (Ternary) Simulation
Date: 16 Nov 89 09:03:49 GMT
Sender: news@lso.win.tue.nl
Distribution: nlnet
EINDHOVEN UNIVERSITY OF TECHNOLOGY
DEPARTMENT OF MATHEMATICS AND COMPUTING SCIENCE
The second seminar in the NFI seminar series
ASYNCHRONOUS CIRCUIT THEORY and TESTING OF SEQUENTIAL CIRCUITS
by Professor J. A. Brzozowski will be held on Friday December 1st 1989
from 10.45 to 12.30 in lecture room Hoofdgebouw 8.78. The topic is
TERNARY SIMULATION
Abstract:
In 1964 Eichelberger proposed the use of ternary simulation for the
analysis of races and hazards in asynchronous circuits. His method was
applied to the "feedback-delay" model of sequential circuit in which
the circuit state is determined by the state of a set of wires having
the property that cutting all the wires in the set results in a
loop-free circuit. Eichelberger's treatment was rather informal and
the results of his simulation were not completely understood.
Nevertheless, the method has been used rather widely; for example, it
is available in commercial logic simulators. In 1976 Eichelberger's
ternary algorithm was applied by Brzozowski and Yoeli to the
"gate-delay" model in which each gate is assumed to have a delay. The
results of this ternary simulation were then compared with the results
of the binary race analysis of the same circuit; in the binary analysis
the "General Multiple-Winner" model of races was used. A partial
characterization was found, but it was conjectured that the
correspondence would be complete if wire delays would be included, i.e.
if the "gate-and-wire-delay" model were used in the binary analysis.
Eichelberger's method consists of two parts, which he called Procedure
A and Procedure B. In 1983 Bryant found a characterization of
Procedure A using the "wire-delay" model. He has also incorporated
ternary simulation techniques in his switch-level simulator, MOSSIM,
designed to model MOS circuits. The method was introduced without
proper justification. A mathematical foundation for the use of ternary
simulation for MOS circuits was found by Lengauer and Naher. However,
these results cannot be directly carried over to gate circuits. The
1976 conjecture was finally proved correct in 1985 by Brzozowski and
Seger. A sketch of the proof of this result will be given in this
talk, and some important consequences of this result will be
discussed.
The main reference for this lecture is: J.A. Brzozowski and C.-J.
Seger, "A Characterization of Ternary Simulation of Gate Networks,"
IEEE Trans. Computers, Vol. C-36, No. 11, Nov. 1987, pp. 1318-1327.
For more information call:
Mrs. F. van Neerven 040 - 47 39 02, or
Prof. J. A. Brzozowski 040 - 47 27 44.
------------------------------
To: uunet.uu.net!comp-simulation@lll-crg.llnl.gov
Path: well!bluefire
From: well!bluefire@lll-crg.llnl.gov (Bob Jacobson)
Newsgroups: comp.simulation
Subject: CALL FOR DISCUSSION: SCI.VIRTUAL-WORLDS
Summary: Initiation of discussion.
Keywords: virtual worlds, virtual reality, artificial reality
Date: 20 Nov 89 06:30:51 GMT
Reply-To: well!bluefire@lll-crg.llnl.gov (Bob Jacobson)
Followup-To: news.groups
Organization: Human Interface Technology Lab, Univ. of Washington
The following message has been posted to news.announce.newgroups. A
discussion has begun in news.groups. We hope this conference's members
will join in.
CALL FOR DISCUSSION OF A PROPOSED NEW NEWSGROUP:
SCI.VIRTUAL-WORLDS
Why this new newsgroup?
-----------------------
This is a proposal to create a new newsgroup on Usenet,
sci.virtual-worlds. This proposal is being made by the Human
Interface Technology Laboratory at the University of Washington,
which proposes to host this newsgroup for the study of "virtual-
world" phenomena.
According to the procedures agreed upon by Usenet adminis-
trators, a new newsgroup must be offered to the Usenet community
for from 14 to 30 days' worth of discussion prior to the taking
of a formal vote on the newsgroup's desirability. Please accept
this as our offer to discuss the new newsgroup we propose,
sci.virtual-worlds. We have until December 15th.
We are making this proposal because we believe that the coming
proliferation of virtual-world phenomena, made possible by powerful
virtual-interface technology, requires the scientific community
served by Usenet to begin debating how this technology will be
employed. Further, with additional research on virtual-world
phenomena taking place at more and more research sites, and in a
growing number of fields -- aerospace, medicine, entertainment,
education, and science -- it is imperative that there be a forum
where the outcomes of this research can be shared most widely.
Initial directions for this newsgroup have been suggested by
various participants in the alt.cyberpunk newsgroup and its
derivative newsgroups. However, to ensure the highest level of
discussion regarding virtual worlds, we believe that this subject
should be elevated to the level of a sci. newsgroup, with the
formal endorsement of the Usenet community.
What is a virtual world?
------------------------
A "virtual world" is a unique, intangible but highly designed
information environment generated by a computer and transmitted
by "virtual-interface" technology to a user who "enters" the
virtual world via appropriate sensory-exchange devices. Best
known of these devices are the Super Cockpit, a cabled helmet-
and-uniform ensemble worn by Air Force pilots to "bond" them
to their planes; and a video facemask and data glove developed
by NASA and refined by VPL and Autodesk.
The virtual-world environment can be as complex as a three-
dimensional "sense surround" comprising seamless visual, aural,
and tactile cues; or as simple as a computer conferencing system.
Virtual worlds are designed to increase the bandwidth of communi-
cation between the computer and the human being, to facilitate
their interaction, and ultimately to improve the human being's
understanding and performance. In short, a virtual world is
created to exploit the perceptual and cognitive skills of human
beings: to design computer systems that communicate with
human beings on human terms, rather than the other way around.
Virtual worlds are also known by the popular term "cyber-
space," coined by the author William Gibson (NEUROMANCER) to
describe the environment within future computer networks as
experienced by future "cowboys" who "ride" the global Net via
headsets and laptop "decks." Other terms used to describe
virtual-world phenomena are "virtual reality" and "artificial
reality."
The substance of the proposed newsgroup
---------------------------------------
The subject of this newsgroup will be virtual worlds, in all
their aspects: the theory of virtuality, the technology that is
being developed and employed to create virtual-world environments,
the people and places working on virtual worlds, and the philoso-
phical questions and social consequences attendant upon the
emergence of this new medium of communication.
Sci.virtual-worlds will be a scientific, economic, philoso-
phical, and spiritual examination of the particular phenomena that
are associated with the creation and experience of virtual worlds.
Moderating the conference will be Howard Rheingold, WELL host and
author (TOOLS FOR THOUGHT) who is currently researching virtual-
world developments for a book in preparation. Assisting Rheingold
will be Drs. Thomas Furness and Robert Jacobson, director and
associate director, respectively, of the Human Interface Techno-
logy Laboratory at the University of Washington.
Who could join this newgroup?
-----------------------------
Anyone with a serious interest in virtual-world phenomena
and/or technology would be invited to join this newsgroup.
Researchers are currently developing sophisticated virtual
interface systems at the University of Washington, MIT, the
University of North Carolina, the University of Utah, Brigham
Young University, Autodesk, VPL, and NASA. In the next year,
several books on virtual worlds will be published. There are a
number of articles about virtual-world theory and practice
already available in professional and popular journals.
Additionally, practitioners in many professional fields and
practices are beginning to investigate the use of virtual-world
technology to solve problems in the conduct of their work. Among
these are chemists, aerospace engineers, medical doctors, educa-
tors, and motion-picture special-effects producers.
Sci.virtual-worlds would be a place where this diverse and
active community of researchers and practitioners can join
together to share knowledge and reflect upon its meaning.
Future plans
------------
The Laboratory intends to make available via Usenet a database
referencing the items in its considerable library regarding
virtual-worlds phenomena and research. The database is in prepara-
tion. An announcement will be made when this archive is publicly
available.
Please join in the discussion!
------------------------------
This is your opportunity to voice an opinion regarding the new
newsgroup proposed in this article. If you agree with our general
premises, your can also help shape the future of sci.virtual-worlds.
Please take advantage and let us know what you think. The dialogue
will be conducted in newsgroup news.groups, where you should
direct your comments.
For more information, contact Robert Jacobson at the follow-
ing telephone or email numbers:
(206) 543-5075 (Laboratory) bluefire@well.uucp
(206) 726-1537 (Home) MCI Mai: Robert Jacobson
You may also contact Howard Rheingold at hlr@well.uucp.
------------------------------
END OF SIMULATION DIGEST
************************