• Skip to main content
  • Skip to search
  • Skip to footer
Cadence Home
  • This search text may be transcribed, used, stored, or accessed by our third-party service providers per our Cookie Policy and Privacy Policy.

  1. Blogs
  2. Breakfast Bytes
  3. How Did Cadence Get to Be Good in Analog?
Paul McLellan
Paul McLellan

Community Member

Blog Activity
Options
  • Subscribe by email
  • More
  • Cancel
cadence
diva
composer
analog
Virtuoso
analog mixed signal
OpenAccess

How Did Cadence Get to Be Good in Analog?

5 Nov 2015 • 4 minute read

jim solomonWe recently held the annual Mixed Signal Technology Summit. I blogged about it and Christine Young covered the keynote. Cadence has had a leadership position in analog and mixed-signal design for a long time, even though originally the companies that were the ingredients to Cadence were purely digital. I called up Jim Solomon to find out the story of how it all started. But first you need to get a bit of Cadence history to give you some context.

Jim Solomon left National Semiconductor, where he was director of IC design for analog and mixed-signal, to found SDA (which, depending on who you talk to, either stands for Solomon Design Automation or Silicon Design Automation). He funded the company, half coming from VCs, and half from four companies (including his former employer), getting them to provide an investment in return for being able to send two engineers to work there and access to the software that resulted. At the time there wasn't really an EDA industry and top-tier semiconductor companies developed their own tools, leaving companies that had fewer resources without many good options. In that era, funding a company for access to design tools was attractive.

SDA only produced digital tools such as schematic entry, layout, and standard cell place & route. The original tools had "edge" in the name, and were actually named after a long-gone ride at Great America, which could be seen from the conference room window. In 1987, SDA filed to go public but unfortunately the day that their IPO was to take place turned out to be "black Monday," the biggest one-day fall in the stock market ever. So the IPO never took place. The following year, 1988, SDA merged with another EDA company, ECAD (which produced the physical verification tool, Dracula). Since ECAD was already a public company, this was a sort of back-door way of going public. The new company was called Cadence, Joe Costello became the CEO, and Jim moved upstairs and became the chairman of the board.

 Jim found that he really had little to do, so he created an analog division using the same model as he had used for SDA, getting partners to invest money and people. This was originally created as a separate company, a subsidiary of Cadence. Jim was president, Charlie Janac (now CEO of Arteris) ran sales and James Spoto (now CEO of Integra Devices) ran engineering. Steve Lewis was a standup comedian who was hired to do training (and is still at Cadence today). He did the opposite of a friend of mine from VLSI Technology, Don McMillan, who was an IC designer who won Star Search and became a full-time stand-up comedian. They created Analog Artist, mainly using the SKILL language (in another weird coincidence, SKILL was partially the creation of Graham Wood, who shared an office with me in Edinburgh when we were both PhD students). Another important early contributor was Joe Santos, who developed the custom layout editor and later was architect and team leader for OpenAccess, the database that underlies many Cadence tools today.

Jim told me that one big issue was that the partner model didn't really work nearly as well as he hoped, since it was hard to get the partner companies to send good people, especially to send real designers, which he wanted. They sent CAD people instead. One surprising fact is that Bell Northern Research sent Rajeev Madhavan who would go on to found Logicvision (acquired by Mentor), Ambit Design Systems (acquired by Cadence) and then Magma (acquired by Synopsys). Actually not getting real designers wasn't that big a deal since they did get money and customers, and they could talk to plenty of design engineers when they made on-site customer visits.

The first version of the Virtuoso tool was actually a microwave custom design editor. But soon it became a key part of Analog Artist. The key components were schematic capture (called Composer), layout (Virtuoso), interactive DRC and ERC (called Diva). The name Cadence came from a mixture of CAD (computer-aided-design) and a tie in to music. The early products of that era all had names with at least a weak connection to music (composer, virtuoso, diva). Initially there was SPICE for simulation but Ken Kundert joined the team and Spectre was born (not sure how that ties into music, though). Back around 2000, when I ran Cadence's Custom IC division, the musical names all still existed. Now I think only Virtuoso survives.

Obviously, in the years since, a lot of development has happened with all these products, but those early versions are the acorns from which today's oaks have grown, and continue to provide market leadership for Cadence in this area.