Task data Optimization allocation with Data Migration

Chip Multi-Processors (CMPs) is applied in the embedded system. An embedded system with more cores means that there will be token more size and consumed more power. Scathed-Pad Memory (SPM) and Non-Volatile Memory (NVM) are new memory technologies. The embedded system can reduce its size and power consumption, which employ the SPM and the NVM. In this paper, we propose a Task data Optimization allocation with Data Migration algorithm (TODMA). Data migration and dynamic programming are combined to allocating task data, which has the dependency between them, in the TOMDMA algorithm. The Experimental results show that the TODMA algorithm can reduce time costs and the number of write activities on NVM by 36.25% and 24.58% respectively, and lower 34.41% systems energy consumption comparing to greedy algorithm And comparing to IDOA, the corresponding reduction are 33.82% , 10.00% and 24.27%.