Keywords: HPCN, CFD, shallow water models, weather prediction, distributed computing, PVM, workstation networks, load prediction
Start Date: 1 February 95 / Duration: 30 months
[ participants / contact]
The objective of this project is to improve distributed applications dispatch criteria using statistical information to predict workstations load. The results will be tested in two ongoing projects that run in parallel over a workstation network: the Rio de la Plata Parallel Tidal Model, developed at the University of Uruguay, and the Atmospheric General Circulation Model, developed at the university of California (UCLA).
Both models described above are now running in three different architectures acquired (two of them) through the ITDC project: ethernet (10 mbit) network, FDDI (100 mbit) network and an SMP (shared memory multiprocessor). Statistical information about network tools has been gathered during the last year, and algorithms that take advantage of IT are being developed.
At this moment, results about the Parallel Tidal Model have been presented in some international conferences on computational mechanics and fluid dynamics. More information about the project at http://www.fing.edu.uy/uruparallel.
Centro de Calalo-Engineering School
Universidad de la Republica
Julio Herrera y Reissig 565
CP 11300-Monreuldeo, UY
Universidad de la Republica, UY
Mr. Roberto Oliveira Mattos
Tel: +598 2 714 229
Fax: +598 2 715 446
URUPARALLEL - ITDC-194, May 1997
please address enquiries to the ESPRIT Information Desk
html version of synopsis by Nick Cook