mwolfe@ogicse.ogi.edu (Michael Wolfe) (03/20/91)
SIXTH DISTRIBUTED MEMORY COMPUTING CONFERENCE (DMCC6) Portland, Oregon April 28-May 1, 1991 CONFERENCE INFORMATION AND REGISTRATION INFORMATION This information is available electronically via anonymous ftp from % ftp cse.ogi.edu [or 129.95.40.2] Name: anonymous Password: [your name] ftp> cd pub ftp> cd dmcc6 ftp> get Register [this file] ftp> get Schedule [full conference schedule] ftp> quit MESSAGE FROM THE CONFERENCE CHAIRMAN The sixth annual Distributed Memory Computing Conference (DMCC6) in Portland, Oregon, April 28 - May 2, 1991 is nearly upon us. We expect that the growing use of parallel supercomputers for "real world" commercial and industrial computing will make this one of the most significant conferences of 1991. The theme of the 1991 conference - "Enabling Technologies for Parallel Computing" - combined with locating the conference in Portland, will make DMCC6 a focal point for examining key issues facing the advanced computing industry over the next decade. The preliminary program reflects the emphasis on program content and quality that was set as a primary goal for the 1991 conference. We have shortened the conference by one day. This means that, of several hundred papers and posters submitted, only the very best have been selected for presentation at DMCC6 - again, emphasizing quality, not quantity. Oregon is second to none as a destination site, and you will find the entire Pacific Northwest a special place. Special pre- and post-conference tours can be arranged to enable attendees to extend their enjoyment of the beauty of Portland and the Pacific Northwest. Register early for DMCC6 using the attached registration form. The sixth annual Distributed Memory Computing Conference (DMCC6) in Portland, Oregon, April 28 - May 2 will be the most exciting conference of 1991. Our famous "Northwest Hospitality" will be waiting for you. See you soon in Portland! Walter Rudd DMCC6 General Chairman Oregon Advanced Computing Institute ******************************************************************************* TUTORIALS & SUMMARY PROGRAM GUIDE ****************************************************************************** Sunday: April 28, 1991 - Registration - All Day Tutorials =================================Half Day Tutorials=========================== 8:30 - Noon Half-day Tutorial Session I T1: "Computational Fluid Dynamics on Parallel Processors: Algorithms and Applications" Ramesh K. Agarwal, McDonnell Douglas Research Labs. T2: "Teaching Parallel Computing" Marilynn Livingston, U. of Southern Illinois-Edwardsville and Quentin F. Stout, University of Michigan 1:15 - 5:00 Half-day Tutorial Session II T3: "Monte Carlo Methods for Partial Differential Equations: Highly Parallel Algorithms based on Statistical Sampling" Michael Mascagni, Supercomputing Research Center T4: "Compiling for Distributed-Memory Machines" Hans Zima, University of Vienna ===================================All Day Tutorial============================ 8:30 - 5:00 Full-day Tutorial Session T5: "The Future of Supercomputing: The Next Decade and Beyond" Stephen F. Lundstrom, PARSA 5:30-8:30 - Attendee Reception ******************************************************************************* TUTORIAL DESCRIPTIONS T1: "Computational Fluid Dynamics on Parallel Processors: Algorithms and Applications" Ramesh K. Agarwal, McDonnell Douglas Research Labs. This half-day tutorial is concerned with the implementation of algorithms to solve the Euler and Navier-Stokes equations on concurrent computers with grain sizes ranging from fine to coarse. Modern numerical algorithms will be discussed, and their implementation on machines such as the CRAY X-MP, Intel iPSC, and Connection Machines CM-2 will be described. Particular attention will be paid to differences in implementation on SIMD and MIMD architectures. In addition, user interface and software portability issues will be addressed. T2: "Teaching Parallel Computing" Marilynn Livingston, University of Southern Illinois-Edwardsville, and Quentin F. Stout, University of Michigan This tutorial will help persons institute introductory courses in parallel computing by providing specific information on course organization, hardware and software options, course materials, and funding opportunities. Emphasis will be on parallel computing alternatives suitable for a teaching environment rather than a research/production one. A spectrum of courses in parallel computing will be outlined, ranging from application-oriented parallel programming to more abstract parallel algorithms. Classroom materials such as syllabi, suggested reading lists, some sample homework problems and lab exercises will be made available. T3: "Monte Carlo Methods for Partial Differential Equations: Highly Parallel Algorithms based on Statistical Sampling" Michael Mascagni, Supercomputing Research Center This half-day tutorial begins with an introduction to the Monte Carlo method (MCM) via the example of numerical integration. This includes basic sampling theory and examples of variance reduction techniques. Next MCMs based on functional integration for elliptic PDEs are discussed. These are MCMs based Brownian motion and random walks. Methods for linear and nonlinear parabolic PDEs will then be presented via the random gradient method (RGM). Included will be new results on the RGM in two and three dimensions Finally, the parallel implementation of these PDE MCMs is presented with examples for both MIMD and SIMD machines. T4: "Compiling for Distributed-Memory Machines" Hans Zima, University of Vienna This half-day tutorial describes a new compilation technique for mapping sequential programs to distributed memory systems (DMS). It is based on the SPMD model of computation in which each processor performs the same computation on different sets of data. The task of the compiler is to create a distributed program and insert all required communication. Currently the main task of the user in the compilation approach is the specification of array distribution. We will outline current work on automatic data decomposition, and will include a demonstration of the SUPERB system, which curently transforms sequential FORTRAN programs into parallel programs for SUPRENUM, the Intel iPSC/860, and Transputer arrays. T5: "The Future of Supercomputing: The Next Decade and Beyond" Stephen F. Lundstrom, PARSA Important multidisciplinary applications are driving the development of new scientific supercomputing capabilities toward the TeraFLOP range. This full-day tutorial will look at the facts of where available technologies are now, how they are advancing, and how they might be used to implement the fantasy - the dreams - of future fully-scaled TeraFLOP computing systems. Topics will include significant aspects of hardware, systems software, application devolopment support, and visualization, discussing what capabilities are feasible over the next decade and specific challenges which must be met in the near term. **************************************************************************** ===========================SUMMARY PROGRAM GUIDE============================== Monday: April 29, 1991 *** Registration - 7:00 to 8:15 am 8:15 General Session 8:30 - Invited Speaker: David Kahaner: High Performance Computing in Japan TRACK 1 TRACK 2 TRACK 3 TRACK 4 SST1: Software CA1: Computation PA1: Performance V1: Vendor Technology and and Algorithms: Analysis: Tutorials: Tools: Experiences Performance Parallel Software Simulation & Modeling Tools Task mapping SST2: Software CA2: Computation PA2: Performance V1: Vendor Technology and and Algorithms: Analysis: Tutotials(Cont): Tools:Operating Algorithms:Finite Benchmarks Parallel Software Systems Element Methods Tools (Cont) SST3: Software SS1: Special PA3: Performance V2: Vendor Technology and Session: Analysis: Tutorials: Tools: Intel/DARPA Performance Distributed Ada Compilers Touchstone Visualization Exp. & Aspirations Tuesday: April 30, 1991 *** Registration - 7:00 to 8:15 am 8:15 General Session 8:30 - Invited Speaker: H.T. Kung: Fine Grain vs. Coarse Grain Parallelism TRACK 1 TRACK 2 TRACK 3 TRACK 4 A1: Architecture: CA3: Computation and SS2: Special: V3: Vendor Tutorials: Interconnection Algorithms: Session: The Programming the Networks Optimization iWarp Project iPSC/860 Methods SST4: Software CA4: Computation and PA4: Performance V4: Vendor Tutorials: Technology and Algorithms: Analysis: Portable Parallel and Tools:Task Methodology Portable Programming Scheduling Parallelism SST5: Software CA5: Computation and MS1: Minisymposium: V4(Cont): Technology and Algorithms: Aerospace Appl. Vendor Tutorials: and Tools: System Solvers and Dist. Memory Portable Parallel Programming Computing Programming(Cont) Language Models Banquet Speaker: Terry Cole: Planetary Data Visualization at JPL Wednesday: May 1, 1991 8:15 General Session 8:30 - Invited Speaker: Ken Kennedy: Compiler Support for Portable Parallel Programming TRACK 1 TRACK 2 TRACK 3 TRACK 4 A2: Architecture: SS3: Special Session: SST6: Software Tech. V5: Vendor Interconnection Performance Analysis and Tools: Load Tutorials: Networks Tools Balancing Express & Automatic Parallelizing Tools A3: Architecture: CA6: Computation SST7: Software Tech. V5: (Cont) Special Purpose and Algorithms: and Tools: Load VendorTutorials:Designs Communication Balancing Express & Automatic Parallelizing Tools (Cont) 12:40 CONFERENCE ENDS ***************************************************************************** REGISTRATION INFORMATION **************************************************************************** HOUSING RESERVATIONS Blocks of rooms have been reserved at the hotels listed on the Housing Request Form. To guarantee a reservation at the special convention rate shown, make your reservation request early by completing and mailing the Housing Request Form. DO NOT CALL THE HOTEL DIRECTLY. You may not be offered the convention rate and DMCC6 may not receive credit for your reservation against our room block. DMCC6 must receive your Housing Request Form by April 1, 1991. After this date, the hotel's standard rates will apply. Space will be assigned on a first-come, first served basis. All DMCC6 hotels are located downtown and within walking distance of each other. AIR TRAVEL SAVINGS FROM UNITED AIRLINES AND DELTA - OFFICIAL CO-CARRIERS FOR DMCC6 United Airlines will offer a 45% discount off unrestricted coach fares for all DMCC6 attendees travelling to/from Portland between 4/23/91 and 5/9/91. The special fares are available through United's Meeting Desk (1-800-521-4041) [8am to 11pm EST]. Reservations must reference the DMCC6 Meeting ID # 439IT to receive the special fares and a special Hertz rental rate. Special fares are also available through Delta Airlines. (Call 1-800-221-1212 and ask for the Special Meeting Network.) Reservations must reference the DMCC6 Meeting ID # R23069 to receive the special fares. Because these special fares can amount to a savings of as much as $500 on a round trip coach fare from the East Coast, attendees are urged to take advantage of this special offer. Both airlines will offer special Canadian discounts. ***************************************************************************** CONFERENCE AND TUTORIALS REGISTRATION For paid tutorials, the tutorial fee includes admission to one tutorial and notes. Please note that ADVANCE REGISTRATION IS REQUIRED for all paid tutorials. Tutorials with too few participants will be cancelled. Registration fees for the tutorials, with conference registration, are as follows: REGULAR STUDENT One-half Day Tutorials (T1, T2, T3, T4) - Postmarked Before April 1, 1991 $175 $ 90 Postmarked After April 1, 1991 $200 $110 Full-day Tutorial (T5) - Postmarked Before April 1, 1991 $275 $135 Postmarked After April 1, 1991 $300 $175 Registration fees for tutorials only are: One-half day tutorials $250 $175 Full Day Tutorial $350 $200 Tutorials Sub-Total =________ The regular registration fee includes admission to contributed sessions, poster sessions, all exhibits, the Conference Banquet, and a copy of the Conference Proceedings. The student fee includes all of the above except the Conference Banquet, for which tickets may be separately purchased at the conference. REGULAR STUDENT Postmarked on or before April 1, 1991 $245 $90 Postmarked after April 1, 1991 $275 $110 At the Conference $300 $120 Technical Program Sub-Total =________ Indicate quantity of each required: Conference Banquet ___ x $45 = _____ Additional Conference Proceedings* ___ x $95 = _____ Additional Items Sub-Total =________ GRAND TOTAL = ________ ****************************************************************************** CONFERENCE REGISTRATION FORM NAME:_______________________________________E- MAIL:__________________________ TITLE:________________________________________________________________________ COMPANY/UNIVERSITY:_________________________________________________________ ADDRESS:___________________________________________________________________ CITY:________________________________STATE:___________________ZIP:___________ COUNTRY:___________________________PHONE NUMBER: (____)_____________________ Do you want the above information included in the register of attendees? ____Yes ___No ______MasterCard Card No.__________________ Exp. Date________ _____VISA Signature_____________________ [FAX to OACIS (503) 690-1210) **************************************************************************** REGISTRATION Fill out all of the preseeding registration forms and mail or FAX to: DMCC6 Registration Knudsen Associates, Inc. 2701 Vaughn St, Suite 700 Portland, OR 97201, USA FAX: (503) 222-4520 If paying by credit card, you must FAX both pages of the SIGNED registration form to (503) 222-4520. Please note the following: (1) Students must include a copy of a current student ID with their payment to claim discounts. (2) Full payment must accompany the form, otherwise it will not be processed. (3) Checks or money orders for conference registration must be in U.S.funds, and must be made payable to DMCC6. Attendees transferring funds by wire must ensure that a copy of the bankerUs draft is sent with their registration form. (4) Applications for advance registration postmarked after April 1, 1991 will be processed at the higher rate of $275. Registration at the conference will be at the $300 rate. CANCELLATIONS To cancel your pre-registration and claim a refund, written notice of cancellation must be received by DMCC6 before April 1, 1991. Requests for refunds should be sent to DMCC6 Registration at the address given above. STUDENT VOLUNTEERS Students willing to trade a half-day's work at the conference for a free student registration should complete the Student Volunteer Form. (The Tutorials, Conference Banquet, and Proceedings ARE NOT included in this free registration, but may be purchased separately). Volunteers are needed Sunday through Wednesday, April 28 through May 1. The number of volunteer positions available is limited and work times will be assigned on a first-come, first-served basis. Volunteers will be accepted prior to April 1,1991. Student volunteers should complete the form below and mail or FAX to: DMCC6 Registration Knudsen Associates, Inc. 2701 Vaughn St, Suite 700 Portland, OR 97201, USA FAX: (503) 222-4520 ****************************************************************************** STUDENT VOLUNTEER REGISTRATION FORM Name______________________________________ College/University________________________ Address___________________________________ City______________________________________ State______________________Zip___________ Phone (_____)_______________________________ WORKDAY PREFERENCE (Circle Day/Time) 1st Choice: S M T W AM PM 2nd Choice: S M T W AM PM I will work at DMCC6 for 1/2 day in exchange for a free student conference registration. I certify that I am a full-time student at the institution indicated above. Signature_____________________________________ Student ID #__________________________________