Routing in multithread environment

Given the enormous amount of detailed geometry information, the large number of local nets, and the ability to properly partition a design routing has been studied thoroughly to utilize parallelism intensively. In this paper we first discuss how to divide the routing space into regions that reduce run time and memory usage without scarifying the quality of the results. Then we cover scheduling among the routing regions; because scheduling determines the effectiveness of parallel routing. We consider the locking, quality of results, and scaling for scheduling in a multithread environment. Experiments show good routing quality with significant speed up in detailed routing.