Last night, Bob Smith, the executive director of what was EDAC, announced the new name and the new mission. The organization will henceforth be the Electronic System Design Alliance. When EDAC was founded in 1989, chip design was all about EDA. There was no significant IP business, no embedded software since it wasn't really possible to embed a microprocessor yet. EDA was a hot area with lots of interest from VCs funding startups and investment banks taking them public. One other motivation, Bob told me, was that DAC was very important but the EDA industry had no say in the show, it was still run as a pure academic conference. EDAC alllowed the industry to represent itself with one voice. But that was then and this is now. An electronic system today requires not just EDA but also interface IP, processor IP, embedded software stacks, advanced packaging, and more. At Cadence, we call this system design enablement, the notion that a system consists of a lot of different layers that all interact and need to be considered concurrently, from libraries (or even sometimes process) all the way up to system and application software. Electronics is a multi-trillion dollar industry, depending on how you count. Semiconductor is $350B or so. EDA is about $8B, but the whole design ecosystem is a $15-20B industry. The above graph, drawn from the EDAC market statistics service, shows the importance of IP (the orange line), growing from zero at the left edge of the chart to overtaking CAE, which is the largest EDA category. I don't know if it is because EDAC is in the same building as SEMI, but in the same way as SEMI represents the manufacturing ecosystem with everything from $100M+ EUV steppers down to gases and metals and wafer blanks, Bob sees his organization as representing the design ecosystem. Moore's Law-type scaling is running out of steam: how many companies will be able to afford to do 7nm designs and have the product volume to justify it? So what has become known as More than Moore is increasing in importance: advanced packaging, and so on. Plus, while hardware gets more complex all the time, the amount of software in systems is exploding. As a result of these trends, EDAC is broadening its mission to cover all of design, not just EDA, and is changing its name to the Electronic System Design Alliance to reflect this. Their new website will be esd-alliance.org (I'm sure many people in the industry are going to accidentally discover that esda.com is a German manufacturer of women's stockings). Bob told me that the ESD Alliance is in the process of kicking off three special interest groups: IP, headed up by Warren Savage, will be working on how to add fingerprints to IP and thus make it easy for design groups and foundries to automatically generate a list of all the IP contained in a design. More and more large companies are getting audited for IP royalty compliance and they need a way to look at designs without having to look at all the files. After all, a modern design might have hundreds of thousands (or even millions) of design files. 3D, headed by Herb Reiter. He is creating a working group to bring together design and manufacturing. For IC design, we have PDKs to form the link between foundry and design tools. We need something similar for 3D so that it is possible to provide more automation. Embedded softrware. This is slower to get off the ground. The focus is on software close to the silicon, of course. Once software is at a high enough level, it doesn't interact closely with the hardware, and is often even independent of the microprocessor on which the code will run. Low-level software is not like this at all, interacting with hardware registers, power management, offload processors, and more. I asked Bob about broadening the geographic scope of the ESD Alliance, since it still seems very Silicon Valley-focused. This is in comparison to SEMI where the largest conference is not SEMICON West in San Francisco but SEMICON Shanghai with over 50,000 attendees. Bob is in complete agreement, but this is not something that can be fixed instantly. In fact, Bob says the biggest challenge is to rebuild the membership. At its peak there were 200-300 members and now it is under 100 (which reflects a lot of M&A activity as much as anything). That will give enough financial muscle to broaden geographically, have a presence at more international conferences, and so on. For smaller companies, there are small company workshops. For the large companies, things like the anti-piracy, export rules, the market statistics service. Further, EDAC in the past and the ESD Alliance going forward are one of the three sponsors of DAC (along with iEEE-CEDA and ACM). Members get a discount on booth space. An example of where the whole industry can work together and save a lot versus working separately is the latest issues with ZTE, who were discovered setting up shell companies to evade export restrictions to Iran. In general, ZTE will not be able to ship equipment containing a significant quantity of US parts. This affects a lot of customers of EDA companies who will not be able to ship parts to ZTE because ZTE will not be able to sell the end equipment. EDA is just a tool that is used in the process of creating chips, but the situation is unclear. Larry Diesenhof is head of the export committee (and works for Cadence, but in this context that is a coincidence) and is driving to get clarification for the whole industry from the US Bureau of Industry and Security (BIS). So their new mission. Straight off Bob's last slide: The Electronic System Design Alliance (ESD Alliance), an international association of companies providing goods and services throughout the semiconductor design ecosystem, is a forum to address technical, marketing, economic, and legislative issues affecting the entire industry. It acts as the central voice to communicate and promote the value of the semiconductor design industry as a vital component of the global electronics industry. Previous: Memory Standards and the Future Next: Blue Gecko, Designed with Cadence Mixed-Signal, Low-Power Flow![]()
![]()
↧