SANTA CLARA, Calif. – EDA has got a challenge on its hands, but that is one of the things that attracted John Bruggeman to join Cadence Design Systems Inc. as chief marketing officer. He was previously CMO with embedded software company Wind River Systems.
Bruggeman, speaking to EE Times at the DesignCon conference here, reckons there is a lot of churn in the EDA market as difficult, expensive or even failing projects are causing semiconductor companies to throw out incumbent suppliers and try a different design flow. The complexity of design at 40-nm, 28-nm and down to 20-nm is also driving a transition from a pick-and-mix best-in-class approach to a single supplier or few supplier strategy, Bruggeman said. And, of course, such changes and the negotiations associated with them, are one of the mechanisms by which the chip companies bear down on the pricing.
Bruggeman likes to talk of the accounts Cadence has picked up from rivals but he acknowledges that for the industry as whole the churn leaves a suggestion that EDA is struggling to do its job.
"It's fairly damning for EDA. It also creates problems for the startups that have been fueling innovation [in the EDA industry]. And semiconductor consolidation means less customers and less designs, albeit going to higher volume."
In short EDA remains, as it has always seemed to be, "an industry in jeopardy."
Cadence's response a year ago was the EDA 360 initiative. Essentially this is a corporate vision that wraps up most of the wisdom that have been extant for some time; that designers need to move up in abstraction to the software and systems levels, while also contending with the complexities of billion-transistor designs and the additional issues of deep sub-micron physics as designs move to finer geometries. It's easy to say but not easy to do.
But while much of the EDA industry has focused on the significant problems of just getting a chip out the door, the customer base has evolved, asserts Bruggeman. Many of them are now software companies that use silicon as a means of execution and as means of protecting their software and making sure they get paid for it on a per use basis.
EDA has to address needs that start with application and system considerations. Designing and making the chip is just part of the job. If it is not done in a way that is integrated with the system issues and the project fails, nobody makes money.
I think that Bruggeman's going to come up with something. He doesn't know exactly what the solution is, but he has a good idea generally where to look. He understands that yet another "me too" product is not going to save EDA.
Without speaking for John Bruggeman I would give the examples of Spreadtrum which has, with TSMC, developed a 40-nm implementation of the TD-SCDMA baseband processor. A significant part of Spreadtrum's added value is in the communications protocol. The chip is the means by which it is delivered and executed.
Similarly Mirics Semiconductor develops software modems for multimedia and TV-on-PC applications. It happens to make chips as one means of delivery.
And ditto Icera Inc. which makes software modems for HSPA and LTE and provides the chips on which those soft modems can run efficiently.
Can someone give some examples on this use case?
"Many of them are now software companies that use silicon as a means of execution and as means of protecting their software and making sure they get paid for it on a per use basis".
In refernce to the first paragraph - OpenAccess is completely owned and licensed by the Silicon Integration Initiative (Si2). http://www.si2.org/?page=69
Cadence is authorized by Si2’s OpenAccess Coalition to maintain and expand the database, under the guidance of the Coalition’s Change Team. Other enhancements to OpenAccess come from community contributions.