Contents ======== Necessities Optional Details, History, etc. Necessities =========== ************************************* This bibliography is not for redistribution. Inquiries for copies, see the address listed below. ************************************* This biblio comes with a file named "copyright." It is a violation of an agreement with Prentice-Hall publishers to not have that file with the main file. The ACM and IEEE would also be a bit happier as well. 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: eugene@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: orders@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 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 (lls@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 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. ******** Sept. 1988 A big year for supercomputers and parallel processing. Lots of conferences a hot topic. Falling way behind. Addition of humor. More papers than I can keep track. David Bailey helps out with the program and titles of ICPP. David Kotz is also a help. Bart Miller provided the preliminary parallel/distributed debugging workshop papers and Sam Cole at Utah provided the complete set of FPS user group proceedings. The people with the German (IBM/Cray site) parallel processing database using INSPEC and network discussion about "Why not use commercial services?" inspired me to verify what other bibliographies have. There is a thesis that computer science is a clobber's poor child when it comes to coverage of computer references. I would confirm that it is uneven at best. I selected 4 references arbitrarily based on the expression: something old, something new, something borrowed, something blue... The old was the oldest independent reference I can find (S. Gill, 1958), the newest reasonable reference was the last paper in ICPP 87. Mao and Yeh 1980 journal article was selected at random, just a year and a number of screens forward. The blue was what I felt was an important CMU TR with TRs perhaps not being well covered enough. Well the CMU report was well covered (found first) in NASA RECON, and DIALOG NTIS. Lacking in INSPEC and COMPENDIX. Gorin (from ICPP'87) was lacking in RECON and NTIS but was COMPENDIX and INSPEC. Mao and Yeh was also lack in RECON and NTIS, and was in COMPENDIX/INSPEC. Gill was in nothing. None of these paper was in the DIALOG Computer DB (but they could have been as Jones did have another article). Searching for 4 papers on the 5 DBs took close to an hour and cost about $50. Some keywords were found which I did not have (merged them for future reference with my stuff (careful to avoid copyright). Coverage of this was uneven. Next time I will give the library the "grequired" list of 10 papers. ******** 1989 Integrated most of Peter Highnam's biblio. ****** 1990 Paula Sue Utter's parallel debugging biblio (in bibtex) is available devvax.tn.cornell.edu. moved eagle.cnsf.cornell.edu cd pub get debugger.bib Refer version shortly. ******** 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. Added Andrew Burke's parallel graphics bibliography, various Usenet TR lists. Ken Jopp Gave Cray Channels stuff. Jean Shuler (LLNL, CUG and SIGGRAPH) provided more CUG stuff.