IPDPS
2000 TUTORIAL1
Monday, May 1st
8 AM – 12 Noon
Multithreaded Programming for Windows NT/2000: A Practical Guide to Writing Programs for Multiprocessor PCs
John Thornley
Department of Computer Science, University of Virginia
WHO SHOULD ATTEND
This tutorial is intended for experienced C and C++ programmers who are interested in using multithreading to speed up their applications on shared-memory multiprocessors. No previous experience with multiprocessing is required. The tutorial will be of particular interest to those who are considering purchasing multiprocessor PCs, and need up to date information on relevant trends in hardware, operating systems, and programming tools. Although the material will be presented in the context of Windows NT/2000, it will also be relevant to Linux and other Unix variants.
COURSE DESCRIPTION
Powerful multiprocessor PCs are now within the budget of individuals and small businesses, and commodity operating systems provide standardized support for lightweight threads. However, most material on multithreaded programming deals with either systems programming or large-scale scientific supercomputing. This tutorial will provide practical instruction on how to use multithreading to speed up general-purpose applications on
multiprocessor PCs running Windows NT/2000. The emphasis will be on structured programming techniques and tools that make it easier to manage the complexities of concurrent threads. Performance measurements will be given for applications running on 4-way and 8-way Pentium III Xeon systems. By arrangement with Microsoft Corporation, all attendees will receive a free copy of Windows 2000 Server (value over $1,000).
LECTURER
John Thornley is an assistant professor in Computer Science at the University of Virginia, where he heads the Multithreaded Software Engineering Laboratory and teaches classes on operating systems and multithreaded programming. He received B.Sc. and M.Sc. degrees from the University of Auckland, and M.S. and Ph.D. degrees from Caltech. He has been actively involved in research on parallel computing for ten years and is currently writing a book on multithreaded programming.
|