NSC logotype

NSC Seminars on High Performance Computing

Aim

This seminar series is arranged by the National Supercomputer Centre (NSC) at Linköping University. Seminars presented in the series focus on various aspects of high performance computing in scientific and technological research, such as results from actual computations, numerical algorithms, as well as hardware and software solutions in high performance computer systems. The seminars are intended for a scientific cross-disciplinary audience, i. e., scientists, engineers and students with an interest in computational modelling as a research tool.

Information

If you have questions, suggestions or would like further information, please contact Jan Fagerström (fager@nsc.liu.se), NSC.

Programme Autumn 2001

Time Place Speaker Title (click to see abstract)
Fri., Oct 19, 2001, at 13.15-14.00 Room Jordan-Fermi Dr. Charles Grassl, IBM Research Parallel processing including MPI, OpenMP and parallel tools

Please click here to view previous seminars in the series.


Detailed Programme


Title:           Parallel processing including MPI, OpenMP and parallel tools

Speaker:    Dr. Charles Grassl, IBM Research

Time:          Friday, October 19, 2001, at 13.15-14.00

Place:          Room Jordan-Fermi, F-building (Fysikhuset), Campus Valla

Abstract

Parallel processing issues will be covered. Central to this issue is the concept of multiple shared memory nodes. Within a single node, the user can exploit parallel processing with either shared memory (OpenMP) or with a distributed memory model (MPI). To exploit parallelism throughout the entire system, the user can implement either the distributed memory model (MPI) or a multilevel model involving both OpenMP and MPI.
 

About Charles Grassl, Ph.D.

Charles Grassl is an Information Technology specialist in the Advance Computing Technology Center (ACTC) at the IBM's T.J. Watson Research center. Charles works on high performance computing issues involving performance and programming models. These program models include shared memory parallelism involving OpenMP and distributed memory parallelism involving MPI. He on the "multilevel" programming model which combines MPI and OenMP.

Charles' duties include giving workshops of programming and conversion, program conversion and optimization. He also works with high performance computing centers on issues regarding conversion, optimization and benchmarking.

Prior to joining IBM in 1998, he has worked in the benchmarking and in the strategic marketing groups at Cray Research. In the benchmarking department he worked on program optimization using both vectorization and parallelization. In the strategic marketing department he worked on competitive analysis and issues dealing with computer performance.


NSC logotype button
info@nsc.liu.se     Last modified: October 8, 2001       fager@nsc.liu.se