Performance Optimization and Tool Support for High Performance Computing

Wolfgang E. Nagel
Center for High Performance Computing (ZHR)
Dresden University of Technology

With the growing demands in model-building on computer simulation, the programming conversion techniques for efficient algorithms and procedures must keep up with rapid progress on the instrumental platforms in computer architecture. Beyond moderately parallel vector computers, parallel computers in quite "massive parallel" form have been pushing into the market for several years. In many cases, the efficient use of such parallel computers is a challenge to parallel programming. Therefore, in all areas of research and development, strengthened - and in many cases, successful - efforts have been made to port sequential applications onto parallel computers.

Even if the scale results of the ported codes often are satisfactory and prove the message-passing concept - which seems to be an important method at least for the near future - worthwhile as a programming paradigm especially useful in obtaining portable and scalable code for new program development, it is just very hard to believe from an application point-of-view, that this programming model can serve as the central basis in proving massive-parallel computers as the technology trend for the future development. This is especially valid for large, commercial applications in which, until now, the explicit programming of data communication through message-passing has proven a lasting software technology barrier.

The talk will discuss aspects of programming and optimization HPC applications on parallel computers. Special emphasis will be placed on supporting software tools like VAMPIR.

Back to seminar homepage