Skip to Main Content
The selection of the optimal architecture for a system is done through a process called Design Space Exploration (DSE). This paper presents a novel hybrid exploration process for multi objective modular computing architectures to provide even faster means of architecture selection. The introduced exploration approach radically reduces the number of architectural variants to be analyzed during the exploration process. The advantages from the two current DSE approaches have been derived and combined together with an objective of finding even faster solution to the architecture selection mechanism. This faster exploration helps in reducing the time to market for the multi objective designs. The proposed hybrid approach when applied on various DSP benchmarks showed respectable speedup in the architecture exploration process compared to the current existing DSE approaches with multi parametric objective.