Newsgroups: comp.parallel From: sigss97@tcsa1.dl.ac.uk (Kevin Maguire) Subject: ERCOFTAC Summer School Organization: Daresbury Lab, Warrington, U.K. Date: 22 Apr 1997 11:58:27 GMT Message-ID: <5ji953$h60@server1.ctc.com> /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ ======== Please Distribute this to Colleagues ========== ======== Who May be Interested in this Event ========== \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ ERCOFTAC Summer School High Performance Parallel Computing for Fluid Dynamic Applications Daresbury Laboratory, May 12-16, 1997 Summer School Background ------------------------ The first summer school in this application area was held in the J. M. Burgers Centre for Fluid Dynamics, June 24-28, 1996 in the Netherlands at TU Delft. The summer school was organised by Prof. P. Wesseling under the title "High Performance Computing in Fluid Dynamics". This was a highly successful course and it was decided to continue the summer school at other locations in Europe. The second summer school will be held at Daresbury Laboratory, England, and the title has been modified to reflect to increasing interest in utilising and optimising applications for parallel computers. Summer School Objectives ------------------------ This is an intensive 5 day course intended to provide practical tuition and experience and is addressed to CFD and computational mechanics specialists at a graduate level who are interested in learning more about high performance computing and/or want to port their code to a parallel platform. A copy of the book "High Performance Computing in Fluid Dynamics", which was edited by Prof. Wesseling (T. U. Delft) and published by Kluwer, will be provided to all participants attending the course. In addition to this, new lectures are being added to reflect recent progress in this rapidly changing field. Additional course notes will be provided by the new lecturers. Course Contents --------------- The course is both theoretical and practical and will include both lectures and exercises. The lectures will deal to a large extent with the algorithmic, programming and implementation issues, as well as experiences gained so far on parallel computers. Attention will also be given to mathematical aspects, notably domain decomposition and scalable algorithms. Computer science aspects will not be emphasised. A range of exercises will be provided by the lecturers to demonstrate practical aspects of parallel programming using PVM, MPI and native message passing approaches. High level languages, such as HPF, will also be covered. The optimisation of CFD codes, for a range of architectures, will also be covered to enable students to get the best performance from their application. Some of the topics that will be covered are: basic concepts of parallel computers; parallelisation strategies; programming aspects; parallel algorithms; sequential and parallel grid partitioning. Who Should Attend ----------------- Whilst focusing on fluid dynamics as an application area, the course is suitable for all engineering application areas and particularly appropriate for people involved in finite difference, finite volume and finite element discretisation methods. No prior knowledge is assumed and the course will be very beneficial for engineers and graduate level students interested in learning more about high performance computing and/or want to port their application to a parallel platform. Summer School Fees (inclusive of VAT) ------------------------------------- The course fee is: 450 Pounds Sterling - Industrial participants 350 Pounds Sterling - Academic participants ERCOFTAC members are entitled to a 50 Pounds Sterling reduction in course fee. ERCOFTAC Scholarships --------------------- A limited number of ERCOFTAC scholarships may be available to assist Ph.D. students or researchers from East European countries with their course fees. Please contact the course organiser for further details. Participants of the Summer School --------------------------------- All people attending the ERCOFTAC Summer School will each receive: * A copy of the book "High Performance Computing in Fluid Dynamics"; * Lunch each day of the course; * Morning and afternoon breaks; * An invitation to an evening dinner; * Additional notes for any new course material * Access to parallel systems for practical experience. Computing Facilities -------------------- Daresbury Laboratory has a wide range of computing facilities on site, including an IBM SP2, Intel iPSC/860. It also has several workstation clusters available. These facilities will be available for the course students to gain a substantial practical experience of parallel computing. Registration for the Summer School ---------------------------------- To register for the summer school, please contact the course organiser: Dr. D. R. Emerson Head of Computational Engineering Department for Computation and Information Daresbury Laboratory Keckwick Lane Daresbury Warrington WA4 4AD England Email: sigss97@dl.ac.uk Fax: +44 1925 603634 Summer School Venue ------------------- Daresbury Laboratory is located in the Cheshire countryside alongside the Bridgewater Canel, the first working canal in England. The author Lewis Carroll was born in the village of Daresbury where his father was Rector of the local church. The stained glass windows of the church contain characters from his books. Provisional Programme --------------------- Monday 12th May 09:15 - 10:00 Registration and coffee 10:00 - 11:00 Introduction to Parallel Computers: Architecture and Algorithms Dr. D. R. Emerson (CLRC Daresbury Laboratory) 11:00 - 11:30 Break 11:30 - 12:45 Introduction to Parallel Computers: Architecture and Algorithms Dr. D. R. Emerson (CLRC Daresbury Laboratory) 12:45 - 14:00 Lunch 14:00 - 15:00 Introduction to Parallel Computers: Architecture and Algorithms Dr. D. R. Emerson (CLRC Daresbury Laboratory) 15:00 - 17:00 Practical Session I - Performance Tuning/Cache Optimisation Tuesday 13th May 09:00 - 10:30 Porting Industrial Codes on High Performance Computers Dr. M. J. Dayde (CERFACS) 10:30 - 11:15 Break 11:15 - 12:45 Porting Industrial Codes on High Performance Computers Dr. M. J. Dayde (CERFACS) 12:45 - 14:00 Lunch 14:00 - 17:00 Practical Session II - PVM and MPI in practice Wednesday 14th May 09:00 - 10:30 Load Balancing for Unstructured Mesh Applications Dr. Y. F. Hu (CLRC Daresbury Laboratory) 10:30 - 11:15 Break 11:15 - 12:45 Load Balancing for Unstructured Mesh Applications Dr. Y. F. Hu (CLRC Daresbury Laboratory) 12:45 - 14:00 Lunch 14:00 - 17:00 Practical Session III - Grid partitioning examples Thursday 15th May 09:00 - 10:30 Data Parallelism for Engineering Application: HPF/HPF-2 Dr. A. Geiger (University of Stuttgart) 10:30 - 11:15 Break 11:15 - 12:45 Data Parallelismfor Engineering Application: HPF/HPF-2 Dr. A. Geiger (University of Stuttgart) 12:45 - 14:00 Lunch 14:00 - 17:00 Practical Session IV - Practical Applications of HPF Friday 16th May 09:00 - 10:30 High Performance Computing of Turbulent Flows Dr. F. S. Lien (UMIST) 10:30 - 11:15 Break 11:15 - 11:45 High Performance Computing of Turbulent Flows Dr. F. S. Lien (UMIST) 11:45 - 12:45 Reducing Communication Costs for Parallel CFD Applications Prof. A. Ecer (Purdue University) 12:45 - 14:00 Lunch 14:00 - 15:15 Reducing Communication Costs for Parallel CFD Applications Prof. A. Ecer (Purdue University) 15:15 - 15:30 Concluding Remarks 15:30 - 16:00 Break and Close -- Articles to parallel@ctc.com (Administrative: bigrigg@ctc.com) Archive: http://www.hensa.ac.uk/parallel/internet/usenet/comp.parallel