Lattice-basedmemory allocation for data-intensive signal processing applications

Many signal processing systems, particularly in the multimedia and telecom domains, are synthesized to execute data-intensive applications: their cost related aspects - namely power consumption, performance, and chip area - are heavily influenced, if not dominated, by the data access and storage aspects. This paper presents a memory allocation methodology for multidimensional signal processing applications, focusing on the problem of mapping the signals from the algorithmic specification into the physical memory such that the overall amount of data storage be reduced. This mapping approach is formally modeled, which is advantageous from the point of view of computational efficiency. This algorithm is currently embedded in an energy-aware memory allocation framework, which also performs the assignment of the multidimensional signals to the memory layers, the ultimate goal of the entire framework being the reduction of the energy consumption in the memory subsystem.