Description: Application developers are facing new and more intricate performance tuning and optimization problems as parallel architectures become more complex. Hence, users of high-performance computing systems normally complain that it is difficult to achieve a good performance for their application and are constantly asking for more and better performance analysis tools. However, the availability of tools development resources across the industry is shrinking rapidly because development of performance tools is complex and generally not profitable. In summary, more performance tools are needed, but fewer tools can be created by the industry. In this tutorial we present an overview of the major issues, techniques, and re-sources in performance tools development. Our goals are twofold: first, we will provide enough information, such that users who need specialized tools could attempt to do in-house development, in order to fulfill their needs. Second, we will discuss open problems in the area for researchers and students interested in working in the field of performance tools. Areas covered will include instrumentation, performance measurement, performance data representation, analysis, and visualization techniques. For more information see http://www.fz-juelich.de/zam/tools101/
Object-Oriented Middleware and Components for the Grid:
Description: Object technology and middleware are increasingly used within parallel and distributed computing. At the same time, components are becoming a very effective tool for the composition and deployment of business applications. While the difficulties to actually program and deploy on the Grid have been, to a great extend, acknowledged, components technology is in the process of bringing many advances for such parallel programming. The aim of this course is to explain and detail this evolution.
Description: The distributed shared-memory programming paradigm has been lately getting rising attention. Recent developments have resulted in viable distributed shared memory languages that are gaining vendors' support, and several early compilers have been developed. This programming model has the potential of achieving a balance between ease-of-programming and performance. As in the shared-memory model, programmers need not to explicitly specify whether accesses are local or remote. Meanwhile, programmers can exploit data locality in distributed memory systems using an abstract model that can enable program portability.
Send questions and
comments to firstname.lastname@example.org
Designed and Maintained by MediaGirl
Copyright © 2003. All rights reserved.