Mixed-mode multicomputers with load adaptability
Dynamically reconfigurable mixed-mode computers are those systems that can be configured at run-time into Multiple SIMD and MIMD autonomous submachines. Therefore, they are also known as MSIMD/MIMD computers. In the existing MSIMD/MIMD architectures, the role of a processor to run as either a processing element “PE” or a control unit “CU”, is fixed at design time. It will be shown that such static assignment adversely affects the system's flexibility and performance. This paper introduces an MSIMD/MIMD architecture with dynamic processors assignments. Unlike the existing machines, this architecture dynamically configures/reconfigures each processor to run as either a PE or a CU as needed. A cost-efficient dynamically reconfigurable hardware is provided in order to establish a broadcasting link between any CU and its PEs in an SIMD submachine. It will be shown that the proposed architecture offers significantly better performance/cost ratio than the existing comparables.
Unable to display preview. Download preview PDF.
- H.J. Siegel et al., “PASM: A partitionable SIMD/MIMD system for image processing and pattern recognition, “IEEE Trans. on Computers, vol. c-30, Dec. 1981, pp. 934–947.Google Scholar
- M. Sharma et al., “NETRA: An architecture for a large multi-processor vision system, “Parallel Computer Vision, L. Uhr, editor, Academic Press Inc., Florida, 1987.Google Scholar
- Y.W.Ma & R. Krishnamurti, “The Architecture of REPLICAA Special Purpose Computer System for Active Multi-sensory Perception for 3-dimensional Objects, “Proc. of Int'l Conf. on Parallel Processing, Aug 1984, pp.30–37.Google Scholar
- Nutt G.J., “Microprocessor Implementation of a Parallel Processor,” 4th Annual Symposium on Computer Architecture, Mar. 1977, pp.147–152.Google Scholar
- F.A. Briggs et al., “PM4-A reconfigurable multiprocessor system for pattern recognition and image processing, “Proc. of National Computer Conf. 1979, pp.255–265.Google Scholar
- H.J.Siegel et al., “A survey of interconnection methods for reconfigurable parallel processing systems, “Proc. National Computing Conf., 1979, pp.529–541.Google Scholar
- W.D. Hillis, The Connection Machine, Cambridge, MA: The MIT Press, 1989.Google Scholar
- Shahid H. Bokhari, “On the Mapping Problem, ”IEEE Trans. on Computers, vol c-30, March 1981, pp. 207–214.Google Scholar
- L.Snyder, “Introduction to the Configurable Highly Parallel Computer, “Computer, Jan. 1982, pp.47–56.Google Scholar
- 11]W.Lin & C. L. Wu, “Reconfiguration procedures for a Polymorphic and Partitionable Multiprocessor, ” IEEE Trans. on Computers, vol c-35, Oct. 1986, pp. 910–916.Google Scholar
- F.Weil et al., “An analysis of fixed-assignment hypercube partitioning,” Int'l Conf.on Parallel Processing, vol. 1, Aug. 1990, pp. 222–225.Google Scholar
- Tse-yun Feng, “A survey of interconnection networks, “Computer, Dec. 1981, pp.12–27.Google Scholar
- M.S.Chen & K.G. Shin, “Processor Allocation in an N-cube Multiprocessor using Gray Codes,” IEEE Trans. on Computing, vol 36, Dec. 1987, pp. 1396–1407.Google Scholar