Abstract
A well-known challenge during processor design is to obtain best possible results for a typical target applica- tion domain by combining flexibility and computational performance. ASIPs (Application Specific Instruction Set Processors) provide a tradeoff between generality of processor (flexibility) and its physical characteristics (computational performance and silicon area).This paper evaluates an ASIP design methodology based on the extension of an existing instruction set and architecture described with LISA 2.0 language. The objective is to accelerate the ASIPs design process by using partially predefined, configurable RISC-like em- bedded processor cores that can be quickly tuned to given applications by means of ISE (Instruction Set Extension) techniques. A case study demonstrates the methodological approach for the JPEG algorithm and motion estimation encoding algorithm of H.264 encod- ing standard.