EN
Back to the archive

The encyclopedia · Engineering & Operations · Technical decision · 1969–1996

Berkeley's public-domain SPICE became the standard way to simulate chips

A student project released to the public domain in the 1970s became the simulator every chip company still uses.

University of California, Berkeley

the move

UC Berkeley had no semiconductor fabrication facility, so Donald Pederson pushed the opposite direction: simulate circuits on a computer. A 1969-70 class project under Ron Rohrer — with Laurence Nagel writing the code — produced CANCER, then SPICE, a simulator whose built-in device models meant users supplied parameters instead of FORTRAN routines.

SPICE was announced to the world in April 1973 at the Midwest Symposium on Circuit Theory. Nagel credits its success to Pederson's decision that the program be public domain from the beginning: engineering schools adopted it as a teaching tool, and when graduates entered industry they wrote to Berkeley for copies, so usage proliferated without marketing or a standards committee.

By the 1990s virtually every CAD tool that dealt with circuit netlists understood SPICE 2G6's format, and a multi-hundred-million-dollar industry of SPICE derivatives served integrated circuit design. Berkeley's policy also required users to send bug fixes back, so the whole industry improved one shared tool.

why it works

  • Public-domain release meant no licenses, no paperwork, no permission needed.
  • Universities taught it, so every new engineer arrived already knowing the format.
  • Built-in device models made simulation accessible to students and firms alike.
  • The 'send fixes back' rule let the industry collectively improve one tool.
the payoffGive the tool away; the industry pays back in fixesinspired

what transfers

The strongest adoption engine for a tool is a public-domain release into universities: students learn it, then demand it at work.

what came after

SPICE derivatives now sit inside every chip design flow, and the netlist format survives in tools from every major EDA vendor. IEEE named an award after Pederson, and SPICE's device models, from Shichman-Hodges onward, remain the shared language of chip designers.

references

spotted an error? The archive wants to know.

same kind of clever