4 ms·
You've been designing microprocessors for 20 years? By yourself, or do you mean that you're contributing to your team's microprocessor design? Were you sendin
by kragen 4y ago
You've been designing microprocessors for 20 years? By yourself, or do you mean that you're contributing to your team's microprocessor design? Were you sending out your own processor designs to MOSIS or CMP or something 20 years ago, or was this more an FPGA thing?
- sponaugle 4y agoOh no.. the microprocessor stuff is purely a technical hobby, although it was what I specialized in undergrad EE, and I worked at Intel for 3 years right out of college. I have done FPGA designs, simulated designs in HDLs, and even a few really cool 74X digital logic based designs and implementations... and so many on paper experiments during occasional long boring meetings. The interesting aspect of doing microprocessor designs is the journey from the ISA concept through the implementation, discovering along the way all of the fantastic mistakes you have made. It really makes a better software engineer out of me. Doing deeply technical hobbies has improved my view and skill in so many areas, as well as helped me understand areas that I really suck at... which there are plenty!
- kragen 4y agoNice! What approaches to microprocessor design do you think are underexplored? What surprising mistakes have you made? What design possibilities are just starting to open up? What do you think of Alan Kay's claim that modern microprocessors are unnecessarily slow at running late-bound languages like Smalltalk because they're optimized for C?
- kragen 4y agoI guess I should clarify that I've been looking at weird things like Chuck Moore's Forth chip designs, the Smalltalk-80 virtual machine, the HP 9100A, the Burroughs 5500, and SeRV, but although I've implemented a virtual machine or two in C and carried some design sketches down to the gate level, I haven't run them on an FPGA or a breadboard yet.