Synapse tightly coupled multiprocessors: a new approach to solve old problems

ABSTRACT

The theoretical merits of a tightly coupled multiple-processor/shared-memory architecture have long been recognized. Two major problems in designing such an architecture are the performance limitations imposed by shared-memory bus contention in cached processors and multiple-processor data coherency. In the Synapse system, memory contention was significantly reduced by designing a processor cache employing a non-write-through algorithm, which minimized bandwidth between cache and shared memory. The multicache coherency problem was solved by a new bussing scheme, the Synapse Expansion Bus, which includes an ownership level protocol between processor caches. Using a non-write-through cache and the Synapse Expansion Bus, Synapse has designed a symmetric, tightly coupled multiprocessor system, capable of being expanded on line and under power from two through twenty-eight processors with a linear improvement in system performance.

REFERENCES

• 1: Inselberg, A. "Multiprocessor Architecture Ensures Fault-Tolerant Transaction Processing." Mini-Micro Systems, 16 (1983), pp. 165--172.
• 2: Jones, S. "High Application Availability." In Proceedings of Compcon83. New York: IEEE, pp. 12--17.
• 3: Alan Jay Smith, Cache Memories, ACM Computing Surveys (CSUR), v.14 n.3, p.473-530, Sept. 1982 [doi>10.1145/356887.356892]
• 4: James R. Goodman, Using cache memory to reduce processor-memory traffic, ACM SIGARCH Computer Architecture News, v.11 n.3, p.124-131, June 1983
• 5: Pohm, A. V., and O. P. Agrawal. "A Cache Technique for Bus Oriented Multiprocessor Systems." In Proceedings of Compcon82. New York: IEEE, pp. 62--66.
• 6: Ravishankar, C. V. and J. R. Goodman. "Cache Implementation for Multiple Microprocessors." In Proceedings of Comcon83. New York: IEEE, pp. 346--350.
• 7: James E. Smith , James R. Goodman, A study of instruction cache organizations and replacement policies, ACM SIGARCH Computer Architecture News, v.11 n.3, p.132-137, June 1983
• 8: WILLIAM D. STRECKER, Cache memories for PDP-11 family computers, Proceedings of the 3rd annual symposium on Computer architecture, p.155-158, January 19-21, 1976
• 9: F. A. Briggs , M. Dubois, Effectiveness of Private Caches in Multiprocessor Systems with Parallel-Pipelined Memories, IEEE Transactions on Computers, v.32 n.1, p.48-59, January 1983 [doi>10.1109/TC.1983.1676123]
• 10: M. Dubois , F. A. Briggs, Effects of Cache Coherency in Multiprocessors, IEEE Transactions on Computers, v.31 n.11, p.1083-1099, November 1982 [doi>10.1109/TC.1982.1675925]