Yearly call for readings and a draft of the new README to the biblio
Eugene N. Miya
processing community what they think the 10 REQUIRED readings for a
1st or 2nd year grad student should be. Email me. I will change the
"standings" as needed and post the list in a couple of weeks.
Annotations appreciated.
Also attached is the new README to the biblio. You might read that.
--eugene
README:
Contents
========
Necessities
Optional Details, History, etc.
Necessities
===========
*************************************
This bibliography is not for redistribution.
Inquiries for copies, see the address listed below.
*************************************
Reference:
%A E. N. Miya
%T Multiprocessor/Distributed Processing Bibliography
%J Computer Architecture News
%I ACM SIGARCH
%V 13
%N 1
%D March 1985
%P 27-29
%K Annotated bibliography, computer system architecture, multicomputers,
multiprocessor software, networks, operating systems, parallel processing,
parallel algorithms, programming languages, supercomputers,
vector processing, cellular automata, fault-tolerant computers,
some digital optical computing, some neural networks, simulated annealing,
concurrent, communications, interconnection,
%X Notice of this work. Itself. Quality: no comment.
Also short note published in NASA Tech Briefs vol. 12, no. 2, Feb. 1988,
pp. 62. Also referenced in Hennessy & Patterson pages 589-590.
About an earlier unmaintained version. TM-86000 and ARC-11568.
Maintaining for ten years with constant updates (trying to be complete
but not succeeding). Limited verification against bibliographic systems
(this is better than DIALOG). Storing comments from colleagues
(DIALOG can't do this.) Rehash sections on a Sequent as a test of parallel
search (this work exhibits unitary speed-up). 8^).
The attempt is to collect respected comments as well as references.
Yearly net posting results hopefully updated "grequired" and "grecommended"
search fields. Attempted to be comprehensive up to 1989.
$Revision:$ $Date:$
Introduction
In 1982, NASA management asked for a survey of concurrent MIMD
computational research. Like many organizations, it was seeking yet another
Agency-cited paper. What became very clear to me was that there were
too many surveys of parallel computing in the literature. Most people
were not bothering to look them up much less read them. The field is
still fairly young. How could my survey be useful? I did a dozen drafts of
over 100 pages when I had an inspiration based on Goldberg and Kay's
concept of personal dynamic media. This survey would a sort of
hyper-document (long before hypertext became popular, but after it
was defined by Ted Nelson). After all where would personal dynamic
documents start but our finger tips?
Who & Where
Eugene N. Miya
MS 258-5
NASA Ames Research Center
Moffett Field, CA 94035-1000
(415)-604-4407
(FTS 464-4407)
ARPAnet: eug...@wilbur.nas.nasa.gov
uucp: {known gateways}!ames!eugene
If you are not on the Internet, you can obtain an older version
(with source files) from
COSMIC (NASA's software distribution center)
Univ. of Georgia
Athens, GA
A new added feature:
The folks who run Computer Literacy Bookshop have given me
access to their inventory system. This will allow me access to the latest
books on parallelism without US Government forms and 6-8 weeks
to buy them. In return, I will do short reviews for their Newsletter.
If you are unfamiliar with Computer Literacy,
they are THE computer bookshop in the "Silicon Valley," why they
even have hard bound UMI PhD thesis (that is hard core). Why I've seen
comp.parallel moderator Steve Stevenson spend $300 on computer books
just like a snap, wonder if his wife knows? I hope to get to the
store once a month and check out new books (if this store does not have it,
then maybe its not worth it). Also, organizations can establish corporate
accounts for email orders (do not email Visa numbers [bad, bad, very bad]).
To establish email orders:
ord...@clbooks.com
If you need a better UUCP path, ask me.
Several style considerations were made:
1) full author name when possible (and affiliation where possible %Z
field [dated])
2) complete Journal name when possible. Non-computer science
people read for things like parallel algorithms, and they may be
unfamiliar with the strange ways of computer people.
3) as much information will be presented as possible including
differing points of view in the annotations and so forth. This
includes old TR numbers once something makes it to press. You are
free to edit out.
4) special characters, a problem, we tend to use certain refer and
Unix [read troff] conventions for things like dashes (\*-)
subscripts (using eqn) and so forth. If this is a problem, consult an
eqn and troff manual. Want to try and minimize this
Can I print it out? # not recommended
What's the BEST way to use the biblio?
This depends on the tools you have available to use this bibliography.
Most Unix systems come with refer (a troff tool) and some other BSD tools:
addbib (additions), lookbib (searching), roffbib (printing), and so
forth. The real beauty: these tools are simple to use or replace.
The files containing the text are separated from search tables,
so it's possible for a person with an EBCDIC/IBM
system to use a text editor to search for citations (as is done).
After specific citations are found, I CACHE them into smaller
sub-bibliographies rather than search the whole bibliography.
This has performance advantages, but it uses a bit more space.
Send your personal annotations to me for inclusion. I will do so
anonymously unless otherwise requested.
To start reading, search for the keywords: grequired and grecommended.
If you are going to use this thing, I've been told it helps to cite it
and me. I am not much of a self-promoter. I'd prefer real assistance.
Optional
========
When
The intent is to maintain the bibliography over a ten-year period (starting
1985). The most significant update times are shortly after ICPP
(Parallel Processing conference) around September and at the start
of a new year.
Why
It's frequently said that this function is better done by librarians
and services like Lockheed DIALOG. The bibliogaphy by Louie is one such
bibliography.
Librarians are not computer architects or software designers.
They are not able to analyse important technical reports.
Library summaries are usually just abstracts and the annotations
herein are usually quick reflections by reviewers and not the authors
themselves. The people in the dynamic field must collect these comments.
In particular, it is dependent upon the readers to make critical
annotations (anonymous or otherwise) to save reading time.
Why refer?
Historical really. Unix wars were rampant until 1985. TeX was just
getting out of infancy. The Mac was new. Refer offered minimal overhead
to implement this database and was useable over on EBCDIC machines
(after dd'ing). It also offered reformatting within a document
based on incomplete citation.
Besides, how many computational fluid dynamicists know what the ACM is?
But then how many ACMers know about the AIAA? Few and few. Enough said.
Why copyrighted?
(The Jack Dongarra question)
I made the decision to start with a pre-existing biblio (with annotations)
for better or worse. I was trying to stand on the toes of those before me.
Don't worry, the US Government paid the copyrights, you get it for free,
I just have to keep track of who has it.
How
The compiler (me) wishes to maintain this bibliography. Your suggestions for
additions, deletions, comments to the keyword list, etc. should be
forwarded to me so I may consider and update this work on a regular
basis. I can be reached (preferably) by electronic mail on several
networks, telephone, and via snail mail.
In particular, short contributed, annotations would help immensly.
This version of the multiprocessor/supercomputer bibliography
is composed of entries from a number of sources.
The first source is the annotated bibliography from
Satyanarayanan's book [1980] which also appeared in Computer 1980
and is reproduced here with the permission of Prentice-Hall.
Starting here was my decision and it is the reason for the copyright notice.
The US Government paid the royality for this biblio.
1958.other 1985.other Rcccp Rpp77 add.bib
1960.other 1986.other Rconcert Rpp78 add.conf
1961.other 1987.other Rcsac Rpp79 check.lines
1962.other 19xx.other Rcyber Rpp80 check.tails
1963.other AIcodes Rdado Rpp81 cnbib
1964.other N.add Rdap Rpp82 connect
1965.other RCHiP Rdcs79 Rpp83 connect.bib
1966.other RCedar Rdcs81 Rpp84 copyright
1967.other README Rdcs82 Rpp85 dataflow
1968.other RPASM Rdcs84 Rpp86 docs
1969.other RUltra Rdcs85 Rpp87 email
1970.other Rarch74 Rdcs86 Rpurdue gregs
1971.other Rarch75 Rdf Rrisc klein.bib
1972.other Rarch76 Rdpsdis.bib Rsisal leff
1973.other Rarch77 Repri Rsosp mailing
1974.other Rarch78 Rhibb80 Rspdc82 makefile
1975.other Rarch79 Rhlca Rspdc84 objtest.r
1976.other Rarch80 Ricase Rspdc86 parallel1.bib
1977.other Rarch81 Ricase2 Rspliss83 problems
1978.other Rarch82 Rlou80 Rsrvy r2bib
1979.other Rarch83 Rmpp Rsuper85 r2bib.c
1980.other Rarch84 Rnaslib Rsuper87 search
1981.other Rarch85 Rpp73 Rsystolic.scr search2
1982.other Rarch86 Rpp74 Rtar80 tape.ann
1983.other Rarch87 Rpp75 Rwa test
1984.other Rasplos82 Rpp76 WARNING
Note: I changed organizations in 1988 so coverage after that time is
becoming more and more spotty. I am unable to pay full attention
and it becomes even more critical for readers to help out.
The next sources come from the collected bibliographies on
multiprocessing (and array processing) by Tartar (1980 ACM),
Hibbard (1980 ACM), and Louie (1980 IEEE computer).
It was interesting to track down the typos and other problems in these
biblios: another reason to make this machine readable.
I am grateful that these people compiled earlier lists.
This is followed by the bibliography from my own multiprocessor
survey [Miya 1984]. I have also added a short EPRI survey, and
a survey on distributed processing from Stanford.
Duplicates are painstakingly eliminated, and other information
(keywords, abstracts, and annotations) are consolidated.
What
The original README file was written when the bibliography was
distributed on tape rather than ARPAnet or mail file transfer.
The backbone of those files of that distribution looked like this:
Rwa (Satya's)
|
Rtar80
|
Rhibb80
|
Rlouie80
|
Rsrvy
/ | \
/ | \
Other sources
The proceedings from the annual conference on Parallel Processing (in toto),
the conference on Distributed Processing, and Architecture
conference (selective) are also added. Lastly, this cut contains
references used in my own survey on multiprocessing.
The data flow bibliography comes from Cathy Schulbach who attended
Jack Dennis's data flow class.
The MPP and DAP bibliographies were added with the help of Jim Fischer
at NASA GSFC in January 1984.
Vineet Singh gave me his distributed processing bibliography in SCRIBE format.
The Ultracomputer project is very helpful.
The SISAL project is inside. The CEDAR/CSRD biblography is here.
The bibliography as it currently stands is organized by source, but I
wish to consider other ways of organizing it for maintenance.
Future versions of this bibliography will have additional references
to such unreferred journals such as Computer Architecture News,
SIGPLAN, SIAM publications, and so on. I would also like to further
check some of the sources of the papers I look at.
The topical surveys (Special issues and Computing Surveys) are hit first.
This bibliography has several obvious problems.
Some of the papers are too old, some have little to do with
multiprocessors or supercomputers (etc.), some papers have no
technical merit (position papers). Ideally, a good bibliography
should guide one through the morass. It is hoped to accomplish this
using keywords as machine readable ones such as this bibliography
is suitable for referencing tools like refer. Hopefully, the
search time will not become too unbearable, less a reorganization
and deletion of files.
Older technical reports have been included.
1) Requests for a bibliography range from those seeking introductory
survey papers to specialized technical documents.
2) Because this bibliography is in machine readable form, those
wishing to use it can rely on machine filtering to weed out
papers not needed in a reference list.
n) Lastly, mature fields of study can afford to have fixed mature
bibliographies. Someday the study of multiprocessors will be mature.
Acknowledgements
I am grateful to my management who had the patience to oversee my work over
the last year (1988), and to the Life-Science Division and NAS project people
who allowed me to use their machines in compiling this work.
Lastly, I could like to thank the following people whom typed
hard and long on this initial version (with me):
Wayne Anderson Joe Fasel, III
Annette Asano Joy Scherr
Suzanne B. Fuller Liz Brusky
Becky Getz Saundra Clay
Steve Stevenson
Dejan Milojicic <de...@uklirb.informatik.uni-kl.de>
Additional thanks goes to Peter Bain and Lawrence Leff for posting
various papers to "net.mag" and "ai-digest" on the USENET and saving
me work. A similar note of thanks go to other Usenet posters
including Josh Knight (IBM Yorktown) and Herbie Chong (also IBM YT).
Special thanks goes to Suzanne whom endured many hours typing
as well as waiting for me at home.
***** A Late Notes: *****
During the preparation for the first release distribution, a word
processor floppy containing certain files was damaged. Several
files pertaining to the parallel processing and computer architecture
symposium were damaged. Please report any problems to me as soon as
you find them. We have done as much possible to try and locate damage
in the little time available. Most records were in the 1983-1984 time frame.
***** March 1985 note *****
To repeat, if you are trying to maintain a current copy of the bibliography
via ARPAnet ftp, check our directory every few months. Please avoid adding
annotations, corrections and changes to your copy, forward them to me so
that the concurrent processing community can benefit.
Use care when using refer and lookbib. We have had some suspect searches.
We may be reaching some refer size limitations. (WARNING file)
***** July 1985 note *****
We have added a file of distributed system references from the
University of Lancaster with the assistance of Vince Aragon and
Dr. C. D. Paice. Their annotations will be added to those references
where collisions with the existing data base occur.
.....Upon later inspection, an added note: many of the spellings and
referencing styles in this file are based on English rather
than American conventions. I have altered some, and corrected
obvious typographic errors, others I have integrated with existing
American entries. I am reticent at this time to Americanize all
entries for "international" reasons, but if I get complaints,
and problems, I will modify them further [as a last resort].
****** May 1986: Note I've incorporated a Connectionism bibliography
posted to PARSYM@SUMEX. Connectionism might seem a bit strange to the
numeric types, but I'll keep many of the references until some
significant complaining takes place.
SPECIAL NOTE: I can detect text which has come from V. Singh's and
other bibliographies which probably came thru me. First, I have to
assume this text came from legal copies of my bibliography. Second,
it would help save time if people who use references from my
bibliography to say so, thus saving me time from checking to make
certain that some references don't have to be reentered. This is the
case with some network and connectionism bibliographies I have received.
I have also detected great inconsistencies in spelling in titles and
authors' names in various papers. Because search programs like lookbib
and even more sophisticated programs, cannot detect these differences,
I have taken some liberty to "correct" typos where the sources themselves
are wrong. Sigh!
A similar problem is the American versus English spelling of certain words.
I have made two efforts to help this problem. Where ever possible, I kept
spelling consistent with the original source. The further problem comes
with abbreviations. While `bib` can handle abbreviations in a consistent
way, refer and most people cannot. I keep "obvious" ones like IBM.
Others, I will make an effort to expand.
******* Sept. 1986
The source file Rmp was merged into dated files. This does not remove
some problems like the lack of TR numbers on the TRAC papers referenced,
but it makes for one less file. The source file old.bib was also removed
after contents (Rtar80) were verified as duplicated.
H.T. Kung has gratiously donated his systolic array bibliography.
It's in Scribe format, but I don't plan to change it. (Rsystolic.scr)
Lauren Smith (l...@mimsy.umd.edu) donated a functional progamming bibliography
from Andy Cheese, but it is not in refer or Scribe, just text. (funct)
We are currently trying to convert it (into refer format, DONE).
But it's, separate cover. Ask if you want it.
******** Dec. 1986
Peter Sember of Monash University sent me Anthony Maeder's parallel
computing bibliography which will be merged. Maeder cites it
as an introductory bibliography, not complete, but hopefully, a useful
guide to the literature.
******** March 1987
Rod Oldehoeft <r...@lll-crg.ARPA> added the r2bib program for refer to
bibtex format (to help those interested in TeX formatting). I should
get similar help from Brian Reid with Scribe.
/*
r2bib - convert refer input files to bibtex .bib files
Author - Rusty Wright, Center for Music Experiment, UCSD
Modified by - Rod Oldehoeft, LLNL & Colorado State University.
*/
******** Sept. 1987
Changed "required" and "recommended" to "grequired" and "grecommended"
to avoid collisions in titles and other fields.
******** 1989
Integrated most of Peter Highnam's biblio.
******** 1990
Also added Paula Utter's debugging biblio. Separate cover, ask if you
want it.
******** 1991
Cray Research, Inc., Cray Computer Corp., the Cray User Group,
the FPS User Group, and the Convex User Group:
all deserve acknowledgement for providing materials which will be integrated
as soon as I find the time.