COMRANCE: A Rapid Method for Network-on-Chip Design Space Exploration

As the communication sub-system that connecting various on-chip components, Network-on-Chip (NoC) has a great influence on the performance of multi-/many-core processors. Because of NoC model contains a large number of parameters, the design space exploration (DSE) for NoC is a critical problem for the architects. Similar to the core design, existing DSE process mainly depends on iteratively time-consuming simulations. To lower the time budget, many previous studies focus on reducing the simulations. However, most of the proposed works based on regression or machine learning techniques, whose accuracy will be significantly affected by the scale of training set. It still needs a lot of simulations to build the training set. Previous work of network analysis has shown that, the upper limit of network latency can be estimated at the network diameter and congestion. Based on this perspective, we propose a novel approach to DSE of NoC, called COMRANCE, which provides rapid searching for the configuration with the best performance. We experimental show that, the method can provide exploration for a variety of parameters for detail designing while allowing the architect to balance the exploration granularity and time budget. The method also supports the comparison of area and power consumption among various configurations, and it is convenient to incorporate with expert knowledge in the method. We implement the framework of COMRANCE and evaluate it with different configurations and benchmarks. The experiment shows that, on average, COMRANCE achieves the reliability at 90.91% with benchmark driven and 92.32% with synthetic traffic driven. Besides this, we analyze the factors which affects the reliability of COMRANCE, such as injection rate and NoC scale. We also find that the major factor that affects the reliability of COMRANCE varies as the injection rate changes. Finally, we show that the DSE time cost for COMRANCE linearly increased as the scale of NoC grows.