Computer architecture challenges in the next ten years

摘要

Over the last 50 years, there has been rapid progress in some areas of computer technology - like the golden years of VLSI in the 80's or the emergence of the web in the 90's. Other areas, which at some time seemed promising, such as neural networks, have stalled. The talk will extrapolate from these historical trends and attempt to forecast which branches of computer systems architecture and technology will blossom during the first decade of the 21th century. It will conclude with the description of a research system 'Icecube' being developed in IBM Research which attempts to capture some of these expected trends.