To address the growing need for domain-specific acceleration technologies, this paper describes a design methodology (i) to automatically generate a domain-specific coarse-grained array from a set of ...