Software development strategies for parallel computer architectures
At present, parallel computer architectures with distributed memories are not mature enough to be used by a wide range of application engineers. Efforts should be made to bring these very promising computers closer to the users. One major flaw of massively parallel machines is that the programmer has to take care of the data flow which is often different on different parallel computers. To overcome this problem, it is proposed that data structures be standardised. The database can then become an integrated part of the system and the data flow for a given algorithm can be easily prescribed. Fixing data structures forces the computer manufacturer to adapt his machine to the user's demands rather than forcing the user to adapt to the innovative computer science approach of the computer manufacturer. This paper presents data standards chosen for the ASTRID programming platform for research scientists and engineers, as well as a plasma physics application which was successfully ported on an INTEL iPSC/2 hypercube.
Bibliographic Reference: Article: Physics Reports (Review Section of Physics Letters), Vol. 207 (1991) Nos. 3-5, pp. 167-214
Record Number: 199210090 / Last updated on: 1994-12-02
Original language: en
Available languages: en