Processor/memory co-exploration on multiple abstraction levels

Recently, the evolution of embedded systems has shown a strong trend towards application-specific, single-chip solutions. As a result, application-specific instruction set processors (ASIP) are more and more replacing off-the-shelf processors in such systems-on-chip (SoC). Along with the processor cores, heterogeneous memory architectures play an important role as part of the system. According to last year's ITRS, in 2004 about 70 percent of the chip area will be made up of memories. As such architectures are highly optimized for a particular application domain, processor core and memory subsystem design cannot be apart, but have to merge into an efficient design process. In this paper, we present a unified approach for processor/memory co-exploration using an architecture description language. We show an efficient way, of considering instruction set and memory architecture during the entire exploration process. Finally, we illustrate the feasibility of our approach with a real-world case study.