Google gives its $1m Turing prize to, er, top Google bods: RISC men Hennessy, Patterson

Well deserved

I've never met John Hennessy, but I did work on one of Dave Patterson's projects at Berkeley in the 80s. You couldn't imagine a nicer, more inspiring guy.

Silicon valley view of the world

Hennessy and Patterson got all of the credit for RISC but there were others who preceded them. The IBM 801 project, which lead to PowerPC, was built before either the MIPS or SPARC chips, I remember reading the 801 papers in late 1982 and thinking that they were doing the correct thing. However before any of them there was Seymour Cray who was the original proponent of simple architectures. Cray machines weren't Reduced Instruction Set Computers because he had never complicated them in the first place. The CDC 6600 and the Cray 1 were examples of minimal instruction set machines, the CDC 6600 was contemporaneous with the IBM 360, the machine that could be considered the first CISC machine. There were other simple instruction set machines in the 60s and early 70s, the DEC PDP8 and the Data General Nova, both design by Ed DeCastro, however those machines were simple by necessity, cost was the driving factor which at the time meant simple. Cray's machines were simple as a matter of philosophy, they were designed to be the fastest computers in the world and the way Cray achieved that was by using very simple instruction sets and running them at very high clock rates (for the time).

Lots of RISC in schools.

Primary school was BBC then Acorn Archimedes, big school was, after a couple of years of CISC Mac LC IIIs, PowerPC Macs.

I don't think we saw a normal Newton (ARM) in the flesh, but the Design and Technology teacher was given an Apple eMate (a clamshell Newton with a keyboard only sold in education market, the translucent curved design foreshadowing Sir Jony's first famous product, the iMac) to play with. He didn't bother.

Obviously being a schoolboy, it was the contrast in how these things ran video games that interested me. The Acorns ran games very well compared to the PC I had at home. Lots of sprites and colours and sound, loaded from the native WIMP GUI.

In the summary of their accomplishments, the earlier IBM 801 work by others is explicitly referenced. What is unique about the work of Hennessy and Patterson is their systematic method for evaluating instructions and their process for eliminating less used instructions to reduce the instruction set to a more efficient and smaller set. The headline is misleading. The award is made by the Association for Computing Machinery. Google supplies the funds but has no role in the selection process. I am a former President of ACM and a Turing Award winner and have chaired the Turing Award committee in the past. I am also a Google employee, and I am pleased that the ACM has chosen to recognize the pioneering work of these two computer scientists.

Vint Cerf

Excellent authors and teachers -- well deserved

All their tomes are wonderful (regardless of the authors' name precedence). And nice to see Dave with hair.