A silicon compiler is a software system that takes a user's specifications and automatically generates an integrated circuit (IC). The process is sometimes referred to as hardware compilation.
Silicon compilation takes place in three major steps:
Silicon compilation was first described in 1979 by David L. Johannsen, under the guidance of his thesis adviser, Carver Mead.
Johannsen, Mead, and Edmund K. Cheng subsequently founded Silicon Compilers Inc. (SCI) in 1981.
Edmund Cheng designed an Ethernet Data Link Controller chip in 1981-1982 using structured design methodology, in order to drive the software and circuit-library development at SCI. The project went from concept to chip specification in 3 months, and from chip specification to tape-out in 5 months. Fabricated using a 3-micron NMOS process, the chip measured 50,600 square mils in die area, and was being marketed and manufactured in volume-production by 1983 under license from SCI.
John Wawrzynek at Caltech used some of the earliest silicon compilers in 1982 as part of the "Yet Another Processor Project" (YAPP)
In 1983-1984, the SCI team designed and implemented the data-path chip used in the MicroVAX in seven months. MicroVAX's data-path chip contains the entire 32-bit processor, except its microcode store and control-store sequencer, and contains 37,000 transistors. At the time, chips with similar levels of complexity required about 3 years to design and implement. Including those seven months, Digital Equipment Corporation completed the design and implementation of the MicroVAX within one year.