Artificial intelligence

Abstract: Issues of control of failure safety of data-flow multiprocessor systems with dynamic task distribution between processors are considered. Possibility of the automatic recovery of errors appeared due to processor failure is shown. There is proposed a hardware and software system reconfiguration method, which enables to reduce the list of instructions necessary for system recovery.

Keywords: fail-safe system, reconfigurable computer systems, dynamic task distribution, data-flow computation.

References:

1. Silva J., Wood J.V. Design of processing subsystems for Manchester data flow computer // IEEE Proc.N.Y. – 1981. – Vol. 128, N 5. – P. 218 – 224.
2. Watson R., Guard J. A practical data flow computer // Computer. – 1982. – Vol. 15, N 2. – P. 51-57.
3. Kovalenko A.E., Gula V.V. Fail-safe microprocessor systems. – Kyiv: Technika, 1986. – 150 p. (inRussian)
4. Zhabin V.I. Method of process parallelizing in computer systems // NTUU “KPI” Reporter. Computerscience, control and computer engineering. – 2000. – N 34. – P. 136-142. (in Russian)
5. Zhabin V.I. Parallel process realization in computer systems // Artificial intelligence. – 2002. – N 3. – P.235-241. (in Russian)
6. Zhabin V.I. Realization of calculations controlled by the data descriptor flow in multiprocessor systems //Electronic simulation. – 2003. – V. 25, N 1. – P. 35-47. (in Russian)
7. Zhabin, V.I. The architecture of the real-time computer systems. – Kyiv: VEK+, 2003. – 176 p.(in Russian).