Design Space Exploration of Memory Architecture for Reconfigurable Computing

Memory is a critical component in reconfigurable computing (RC). With investigating the application performance of application algorithms (AAs) for RC, this paper presents algorithms for estimating the high bound for local data memory (LDM) and configuration context memory (CCM). Base on estimation results, the design space exploration methodology considering area constrains for memory architecture (MA) is proposed. The experimental results show that it is efficient to allocate LDM and CCM and improve the performance of application algorithms