Dynamically reconfigurable processors with multi-context facility have been used for various applications. The relationship between context size and performance of such processors is analyzed based on real designs. The Parallelism Diagram which shows the required PEs in each step of the algorithm is introduced as the basis of the analysis, and models for performance and cost are shown. Evaluation results show that the performance is degraded about 23% when the size of a context becomes 1/2. The performance per cost is improved 7-14 times than that of the case without time-multiplexed execution.
Citation:
Hideharu Amano, Shohei Abe, Yohei Hasegawa, Katsuaki Deguchi, Masayasu Suzuki, "Performance and Cost Analysis of Time-Multiplexed Execution on the Dynamically Reconfigurable Processor," fccm, pp.315-316, 13th Annual IEEE Symposium on Field-Programmable Custom Computing Machines (FCCM'05), 2005