M.A.Stoker@newcastle.ac.uk (Michael Stoker) (08/13/90)
Practical Parallel Programming on Unix Shared Memory Multiprocessors September 10-14, 1990 Centre For Multiprocessors Computing Laboratory University of Newcastle upon Tyne Whilst some parallel programming languages are revolutionary in their approach, to programming, many tangible benefits can be gained by pursuing a more evolutionary approach, extending the present-day sequential mode of programming in languages like C and Fortran to encompass parallelism. This evolutionary approach is the prime focus of this 5-day course, which will concentrate on practical parallel programming on Unix-based shared memory multiprocessors. Such computer systems are expected to form a dominant part of the parallel computer marketplace in the coming decade, and hence programming and understanding of such systems is likely to become especially important. The course will introduce the basic concepts of parallelism, parallel programming paradigms and the design of parallelised programs. It will also cover the architecture of parallel computers, parallel programming on Unix systems, discuss the experiences of research staff at Newcastle in producing parallel programs, and cover some of the more revolutionary approaches and software tools being researched for the longer term. The course will contain a significant practical element; attendees will be expected to design and implement parallel programs to learn at first hand about the advantages and pitfalls of parallel programming. Parallel programs will be developed and run on a large, advanced shared memory multiprocessor (an Encore Multimax 520) containing 14 CPUs (120 MIPS) and running Unix. The course is intended for programmers, software engineers and software managers, and will assume familiarity in a sequential programming language such as C, Pascal or Fortran. Familiarity with the Unix system would be advantageous, but not vital. The course will be given by members of the Centre for Multiprocessors (CMP), part of the Computing Laboratory at the University of Newcastle upon Tyne. All of the course lecturers all have significant practical experience in parallel programming. CMP specialises in parallel processing on shared memory multiprocessors, and is sponsored by a grant from the DTI's Enterprise Initiative. Further details can be obtained from Ann Pretswell at, Centre for Multiprocessors, Computing Laboratory, The University, Newcastle upon Tyne, NE1 7RU, UK. Telephone: 091 222 8049. E-mail: Centre-for-Multis@newcastle.ac.uk