System Reliability-Aware Energy Optimization for Network-on-Chip with Voltage-Frequency Islands

In this paper, a novel energy optimization algorithm for Network-on-Chip (NoC) based on voltage-frequency islands (VFIs) is proposed under the system reliability constraints. The proposed algorithm mainly includes the three processes of VFI partitioning, assignment and task mapping to reduce the energy consumption. A system reliability-aware mapping algorithm is proposed which employs a hybrid algorithm based on quantum-behaved particle swarm optimization and multi-neighborhood simulated annealing strategy. The optimum particles upon QPSO are further optimized by annealing operation mentioned above. Besides, the catastrophic operation is executed to enlarge the search scope and avoid getting into the local optimum. It minimizes the communication energy consumption of NoC system on the condition of meeting system reliability constraints. Experimental results show that the proposed algorithm is quite effective in optimizing energy consumption of NoC with voltage-frequency islands under system reliability constraints.