Hot chips and soggy software

The author discusses the bottlenecks that impair performance of a computer system and discusses the success of the RISC (reduced-instruction-set computer) approach. He attributes it, at least in part, to the fact that all the seminal work on the RISC chips was carried out in close conjunction with a strong compiler team. He discusses issues that designers of computer systems must consider and examines trends that will affect the optimum design points for future systems. The author then addresses what he refers to as 'soggy software', i.e. the slow pace of progress in software development as compared to hardware, identifying standardization and reuse as necessary components of any solution to the problem.<
>