The objective of TROPICS was to ensure the exploitation of the next generation of data-processing systems, by reducing:
- execution times for a wide range of applications, through the use of parallelism
- the cost of computation, through the exploitation of VLSI technology
- the cost of producing computer software, by the development of methods and tools.
The objective of the project was to ensure the exploitation of the next generation of data processing systems, by reducing:
execution times for a wide range of applications, through the use of parallelism;
the cost of computation, through the exploitation of very large scale integration (VLSI) technology;
the cost of producing computer software, by the development of methods and tools.
The results fall into 4 categories: system architecture, language, user interfaces and the design methods and applications. On the system side, the POOMA prototype was further evaluated. The layout of the communication processor chip was produced and the node was also redesigned to allow a greater integration of the machine. The first components of the envisaged communication server was produced and the boards to interconnect to integrated services digital network (ISDN) and fibre distributed data interface (FDDI) were tested.
The object oriented language POOL was further developed and implemented on the POOMA machine.
The work on the database system progressed toward the development of a multimedia distributed database server. An extended structured query language (SQL) and the implementation of abstract data types and complex objects were worked on, and first implementations were demonstrated. To develop applications on the parallel system, design tools were constructed that allow a developer to conceive the application in an object oriented manner and experiment with the granularity of the parallelism to get the best performance. Generic tools to construct user interfaces to applications were also developed and implemented.
On the application side, the major achievement has been the geographic application. It aimed to provide a parallel geographic data server allowing the user to ask complex queries on maps and early prototypes have been successfully demonstrated.
TROPICS aimed to research and develop a symbolic, parallel computer system technology to meet the high-performance demands of symbolic applications that constitute the next step forward in office automation, and in other fields. This will be expressed in the delivery of an industrial prototype of a parallel processing system, using the object-oriented approach for exploiting parallelism.
Work-packages were planned for the following tasks:
- defining the architecture of the processing node and its constituent components, followed by VLSI designs and implementations for the components
- the design and implementation of the parallel processing units and their integration into a complete TROPICS systems
- designing and implementing the TROPICS operating system
- designing and implementing support software to provide a platform on which applications can be developed
- making available the object-oriented programming language POOL as the TROPICS main programming language, other languages to offer a migration path, and a set of software development tools
- realising application software targeted for office environments on the TROPICS system
- facilitating the transfer of results and the investigation of the application domain by a transfer desk
- project management.
It was planned to develop demonstrator and industrial prototypes of a workstation to serve as a user interface to the system, and multimedia, cartographic, multilingual and knowledge support applications were to be made in the office area.
55049 Viareggio Lucca
6525 ED Nijmegen
1009 AB/10 Amsterdam
7500 AE Enschede