Data path allocation using an extended binding model

Abstract

Existing approaches to data path allocation in high-level synthesis use a binding model in which values are assigned to the same register for their entire lifetimes. The authors describe an extended binding model in which segments of a value's lifetime may reside in different registers if there is a cost advantage in doing so. In addition, the model supports multiple copies of values and the use of functional units to pass through unmodified values to reduce interconnects. This model was exploited in an allocation tool that used iterative improvement to search for low-cost designs. Results showed that allocation costs can be substantially reduced by using this model.<>