I have just been updating myself on the current state of play in the analog EDA world. Whilst my efforts may or may not have been worthwhile, time will tell, one consideration struck me; just what made SPICE so enduring and successful?
I have just been updating myself on the current state of play in the analog EDA world. Whilst my efforts may or may not have been worthwhile, time will tell, one consideration struck me; just what made SPICE so enduring and successful? Plenty of you will have some strong and insightful views on this (and please feel free to share them with me) but having done a little research, I have a couple of theories of my own.
SPICE can be traced back as early as the 1960s to Professor Ronald Rohrer and his students at the University of California, Berkeley. Developed under the somewhat uninviting moniker of CANCER, the circuit simulation program was rewritten and renamed as 'Simulation Program with Integrated Circuits Emphasis' (SPICE) and released in the public domain in 1972. Theory one: SPICE is nice!
Instrumental in that process was Professor Donald Pederson, an early exponent of the open-source concept, who believed that the technology should be shared. Theory two: SPICE had a promoter with great foresight, as well as long term funding support from the Army Research Office.
So it is that many thousands of engineers have made SPICE what it is today, including various industry luminaries who studied under Pederson. Indeed, many people owe their livings and their design success to SPICE. Theory three: So generations of engineers contributed to SPICE's evolution, huh? Yet, there were similar programs under development at that time, including ECAP and Spectre. I believe that nature biases us towards something we have personally created; call it 'parent syndrome' if you will.
Over the years, SPICE has evolved from a single analysis program to a plethora of simulation and characterization tools. And so it is that today, most analog designers are still using fundamentally the same analysis tools that were used to design circuits some 30 years ago. For me, this strikes a chord. It's strangely satisfying to think that if you peel back the layers, some of the original code that Pederson and his students created all those years ago lies behind current versions.
As analog EDA vendors currently debate topics such as interoperability and the openness of parameterized cells (p-cells) it is worth remembering what made the industry such a success in the first place.
Vanessa Knivett is editor of Analog DesignLine Europe