[comp.parallel] EPCC Invitation to exchange information on research in parallelism

bjnw@castle.ed.ac.uk (B Wylie) (11/27/90)

The EPCC would like to contact other institutions and organisations
involved in research into parallel computation.  A summary of our
activities is listed below; similarly detailed responses from interested
groups which would like to be added to our mailing list are invited.


                Edinburgh Parallel Computing Centre
                -----------------------------------

The University of Edinburgh has been investigating massively-parallel
SIMD and coarse-grained MIMD parallelism since 1982, recently
culminating in the initiation of the new Centre.   Current staffing
totals over 30, including 6 full-time research staff, about 10 staff on
software development and several more responsible for industrial
liaison and servicing a user community of around 300.  This year's
Project Directory lists some 150 active projects.

Current multiprocessor hardware:
 * Meiko Computing Surface: 430+ Inmos T800s (1.7 Gbytes memory)
 * Meiko Computing Surface:  64  Intel i860s (1.0 Gbytes memory)
 * AMT Distributed Array Processor: DAP-608/64 (4096 PEs)
 * AMT Distributed Array Processor: DAP-510/8  (1024 PEs)
 * Parsytec Multicluster II: 16 T800 nodes
 * Transtech: 16 T800 nodes
 * Silicon Graphics 4D-220 GTXB dual processor system

Existing software environments and utilities:
 * Tiny - performance-optimised message-router for transputer-based machines
 * Rian - Meiko OPS configuration and debugging environment
 * CAPE & DAP-CellSim - cellular automaton programming environments
 * Rhwydwaith - distributed neural networks simulator
 * D3 - distributed polygon-renderer
 * Also: OPS, MeikOS, gfx, CStools, Linda, TDS, Helios, X, GKS, DGL, ...

Primary research interests include:
 * numerical algorithms & optimisation
 * parallelisation tools from algorithmic templates
 * mapping & load-balancing in relation to data decomposition strategies
 * tools for interactive parallelisation on multicomputers
 * object-orientated programming systems for parallel computers
 * performance measurement & visualisation of message-passing traffic
 * application development support from tools & environments

Selected recent or on-going applications include:
 * quantum chromodynamics
 * cellular automata for computational fluid dynamics
 * generative communications systems (Linda)
 * protein and gene sequence comparison algorithms
 * surface detection & display within three-dimensional datasets

For more information, or to reply, please contact:

Brian Wylie

Edinburgh Parallel Computing Centre           E-mail: EPCC-contacts@uk.ac.ed
James Clark Maxwell Building
University of Edinburgh
Mayfield Road                                    tel: [+44] (31) 667-1011 x5030
Edinburgh  EH9 3JZ                               fax: [+44] (31) 667-4712
Scotland