Hierarchical Scheduling in High Level Synthesis Using Resource Sharing Across Nested Loops

This paper presents a resource-constrained scheduling algorithm for hierarchical behavioral speciications containing nested loops. The algorithm attempts to share resources across levels, to schedule operations that belong to diierent levels of the nested loop structures in the speciications as well as operations that belong to the same level. We compare the results of scheduling using our algorithm with those obtained using traditional list scheduling with no sharing of resources among diierent levels of the speciication. These results show an average improvement of 23.47% in terms of number of control steps.