ECOOP / Workshop 02

Parallel/High-Performance Object-Oriented Scientific Computing

Abstract

While object-oriented programming is being embraced in industry, particularly in the form of C++ and to an increasing extent Java, its acceptance by the parallel scientific programming community is still tentative. In this latter domain performance is invariably of paramount importance, where even the transition from FORTRAN 77 to C is incomplete, primarily because of real or perceived performance loss. On the other hand, three factors together practically dictate the use of language features that provide better paradigms for abstraction: increasingly complex numerical algorithms, application requirements, and hardware (e.g. deep memory hierarchies, numbers of processors, communication and I/O).

This workshop seeks to bring together practitioners and researchers in this developing field to 'compare notes' on their work. The emphasis is on identifying specific problems impeding greater acceptance and widespread use of object-oriented programming in scientific computing, and proposed and implemented solutions to these problems, though experience papers are welcome if they present a constructive lesson. Work-in-progress descriptions are welcome.

In brief we seek to advance, by use of modern OO languages and design and implementation concepts,the state of the art of parallel/high-performance scientific computing.

Main Topics

Specific areas of interest include, but are not limited to

tried or proposed programming language alternatives to C++;
performance issues and their realized or proposed resolution;
issues specific to handling or abstracting parallelism;
specific points of concern for progress and acceptance of object-oriented scientific computing;
existing, developing, or proposed software;
frameworks and tools for scientific object-oriented computing.

The workshop will consist of sequences (3-4) of short presentations followed a discussion session. The workshop will conclude with an overall discussion. We expect the majority of the participants to give presentations.

An overall summary of the workshop will be published in the ECOOP 2003 Workshop Reader. In addition, full papers will be published as an issue of the NIC (John von Neumann Institute for Computing) series to be distributed at the workshop, and also made available at the POOSC 2003 WWW site. Authors will retain copyright.

For authors of accepted presentations who require justification for travel the organizers can provide official letters of invitation.

Paper Submission

Prospective authors are invited to submit abstracts of their positions. Submissions will be reviewed by committee. Authors ofaccepted papers are responsible for submitting the final version by May 2 to ensure inclusion in the proceedings. Submission and email correspondence to the POOSC 2003 Committee

Type of Papers position / vision / experience

Formatting
Size Abstracts can be submitted in ASCII, PDF, or postscript.
Any common encoding (MIME or uuencode) or compression (zip, gzip, bzip2)is acceptable.
Authors ofaccepted papers are responsible for submitting the final version using the Springer LNCS LaTeX template

Deadlines Abstract Submission April 8^th
Notification April 16^th
Full papers due May 2^th

Details

Organizers	Kei Davis, Co-chair (Los Alamos National Laboratory) Jörg Striegnitz, Co-chair (Research Center Jülich) Andrew Lumsdaine (Indiana University) Bernd Mohr (Research Center Jülich) Jörg Nolte (FhG FIRST) Vladimir Getov (University of Westminster)
Date Location	Tuesday, 22^nd
Workshop Home Page	http://www.c3.lanl.gov/poosc03
Number of participants	10-15
Rules for Attending	Paper presenters preferred, non-presenter only as space allows.
Contact Person	Kei Davis