Advanced architecture as a commodity

Abstract

h HARDWARE ACCELERATION IN Computational Biology is the theme of this issue of IEEE Design & Test. In graduate school, a long time I worked on the acceleration of things that used to be considered compute-intensive, such as compilers. Then I was involved in the acceleration of fault simulation through special purpose hardware. None of these efforts was successful. Advances in general purpose computing technology outpaced that of acceleration, so our speed advantage shrank by the time these accelerators were ready. Since work today is leveraging general purpose hardware like GPUs and FPGAs, this shouldn’t be a problem. But a bigger problem was that our target applications were for a limited market, so the investment in developing acceleration wasn’t worth it. Maybe we can find some other things out there with a vast number of users that we could accelerate. Perhaps in the future, we’ll see ads like the following.