PARDIS: An Architecture for Application-Level PARallel DIStributed
			     Computation

			     Kate Keahey
	Indiana University and Los Alamos National Laboratory

			      Abstract

High-speed networks make possible the development of applications capable
of combining the computational power of several remote supercomputers,
scientific instruments and sophisticated display devices within one system.
While the physical infrastructure necessary to support such applications is
becoming widespread, distributed supercomputing poses new challenges for
software developers. This talk will present PARDIS, a general-purpose
programming environment for developing applications comprising parallel
components implemented using different software systems and interacting in
a distributed domain.

PARDIS addresses interoperability at application-level which allows the
programmer to build metaapplications from independently developed and
tested components; this strategy promotes component reuse and encourages
modular programming. The design of PARDIS is based on the ideas underlying
the Common Object Request Broker Architecture (CORBA), in particular the
idea of interaction through object interfaces. PARDIS extends the CORBA
object model by the notion of an SPMD object representing data-parallel
computations, and distributed arguments which can be used in their
interactions. Further, the programming abstractions implemented in PARDIS
also contain support for concurrency. The talk will present the performance
evaluation of SPMD objects and present a few examples demonstrating how the
abstractions implemented in PARDIS allow the programmer to develop
distributed supercomputing applications. We will discuss issues of parallel
interaction, computational steering and give examples of interoperability
with two data-parallel libraries: POOMA and HPC++ PSTL.

bio:

Kate Keahey has recently received a Ph.D. in Computer Science from Indiana
University. Her research interests focus on distributed and parallel
computing and the issues arising on the intersection of these two areas.
She is planning to continue her work in the field of distributed
supercomputing as a member of the Advanced Computing Laboratory in Los
Alamos National Laboratories in New Mexico.


urls:

for PARDIS: http://www.cs.indiana.edu/hyplan/kksiazek/pardis.html
for me: http://www.cs.indiana.edu/hyplan/kksiazek.html