Sunil Joshi
From Bit-slices to OpenSPARC: Innovation thru Simplification
I was lucky that my first job out of college was to work on Bit-slice processors. Loved it! Had learned all about it in school, done class projects with bit-slices. And here I was working with the creators of these simple, 4-bit, cascadable, scalable processors. In one year, I had designed/redesigned 12 chips (yes, no kidding!) and published a ton of articles. Plus designed a really cool coffee mug.
Then RISC happened. When the trend was complex instructions, here came the reduced instruction set computer. Simpler, but achieved more. This time I was there next to the guys who designed the first SPARC chip at Sun. Simpler, faster, better. I did not design that first chip, but did design a coffee mug and a T-shirt for it! It was a great time for everyone. Sun had reinvented itself for the first time in transitioning from the Motorola 68K to SPARC.
Now many years later, I suppose it was by chance that I got associated with Niagara from day one of its life @Sun. Right place at the right time. It started with the Afara acquisition, the integration/assimilation of the team and then all the hard teamwork to support it. There it was - gutsy, radical, innovative, and simple. Normally in most designs, complexity creeps up on you. The art is to simplify complexity. To take a step back to take three forward. Niagara did. It looks obvious now when you have the proof points and benchmarks ("hindsight is 20-20")! Kudos to those that had the foresight. Text books will talk about this. College students will do their projects with CoolThreads technology.
As for me, I am trying to think of a cool design for another mug.
On second thoughts, let me make it a T-shirt design... goes better with CoolThreads.
Sunil Joshi
p.s. shall do an intro in a future blog...
Posted at 05:17PM Dec 08, 2005 by sunil in Sun | Comments[0]