Static Rate-Optimal Scheduling of Iterative Data-Flow Programs Via Optimum Unfolding
IEEE Transactions on Computers
Computer
Computer
Hi-index | 0.00 |
Data-flow graphs (DFG) can be used to represent a signal-flow graph, a synchronous system or a digital circuitry. Retiming and unfolding are two useful techniques which have been effectively applied to optimize the realization of DFGs. Retiming technique reorganizes an iteration; unfolding technique schedules several iterations together. These two techniques can be combined to schedule DFGs optimally. We show a very efficient algorithm to combine these two techniques. We also present simple methods for examining an unfolded graph by simulating it on the original DFG.