DAC 2013: Top 10 from Day 2 in Austin

The only thing most people remember about Tuesdays at DAC are the parties. You’re a success if you attended at least two, less than a success if you only attended one, and guaranteed immortality if you attended more than three.

Of course, other things happen on Tuesdays at DAC – early morning breakfasts where sincere technologists present and/or opine about somber challenges facing the industry, the plenary session, presentation of multiple awards, pavilion panels, mid-day luncheons, afternoon sessions, posters, and many, many hours logged in by booth staff talking and talking and talking to customers, potential customers, and general industry hangers-on looking for free give-aways.

Admittedly, I’m less than a success in the party area having only attended one, the ever-gracious Synopsys Media dinner. The real hipsters hung out at the Cadence/Denali party tonight, and the even hipper hipsters went to the Stars of IP. Oh well, my list of Top 10 on Day 2 at DAC has nothing to do with parties. Instead, it has to do with people I heard speaking at podiums all day at the front of various rooms. Each one of those talks deserves full coverage on its own, but here I’m just going to log in a few brief notes. If you were at parties all evening, this will be of little interest, if any. Oh well, that’s the way the tumbling tumbleweed tumbles.

No.10)  The GlobalFoundries/Synopsys breakfast. Okay, I got there late and left early, so what do I know except to observe that it was a much more somber affair than what I remember from years past. Different foundry partner perhaps? The row of dark suits sitting up in front of a room full of breakfast guests at the Hilton clearly was on a mission to explain in great detail how Synopsys and GF together will stare down and conquer the pending world of finFETs.

High points: Layout will be the key to overcoming the additional complexities of finFET; Early silicon results are validating the benefits of finFET structures over planar; The requisite design methodologies are going need close collaboration between the foundry, the designer, the software developer, and the EDA vendor. Okay, you heard it here first.

No.9) The Samsung keynote at 9:15 am in Ballroom ABC. There have to have been 800 people in there, at least 200 of them standing in the back for lack of chairs. Samsung Electronics President Stephen Woo is a great speaker and previously an EDA guy, so DAC is familiar territory for him, including a Best Paper in 1994.

His talk centered on “Smarter Mobile Devices” and had great graphics, video content, and an optimistic outlook w.r.t. the synergy that EDA and Samsung will continue to have, which will make everything that can be imagined into a reality. It was the kind of message and tone that one would expect from the opening keynote at DAC. And it was taped, so you can see it later.

No.8) The CEDA-sponsored Cyber-Physical Systems lunchtime talk delivered by UC Berkeley’s Ed Lee. The 4th floor room was packed as Dr. Lee talked about the difference between models and reality and lamented that engineers often don’t know that difference. Lee quoted Solomon Golomb who apparently said, “You’ll never strike oil by drilling through the map.”

Lee also invoked the “Hermann Kopetz” Principle: Many predictive properties we assert about systems (determinism, timeliness, reliability, safety) are not properties of the implemented system, but of the model of that system. From there, Lee talked determinate versus non-determinate models, and said software has just got to start thinking about getting time into its abstractions. Not surprisingly, there were a lot of academics in the audience.

No.7) Freescale’s Jose Nunez’ mid-morning talk, “Challenges of Integrating External IP”, one of the best tutorials I’ve ever heard on the topic. His observations were specific, and his advice for both IP vendors and IP users detailed and adamant. Get everything in writing in advance of buying/implementing IP, if you’re a user. Willingly admit there are bugs in your IP and work with your customers to deal with it, if you’re a vendor. Just a great talk which deserved to be heard in a larger, keynote-sized venue.

No.6) Tufts Prof. Soha Hassoun on “Gene Modification Identification under Flux Capacity Uncertainty.” This is the stuff of the future, if you’ve been listening at all this week in Austin. Every which way you turn, people are talking about the biological applications and implications of EDA-style algorithms and optimizations.

The subject of Prof Hassoun’s talk (“My problem formulation is simple; my goal is to maximize the rate of production of a metabolite.”) is part of that trend, and of particular importance given that she will be General Chair of DAC 2014 next year in San Francisco. Look for Dr. Hassoun’s paper in this year’s proceedings, among others from this session.

No.5, No.4, No.3, No.2) DAC Management Day, the afternoon session, with 30-minute presentations from Samsung Electronics VP of Design Technology Kee Sup Kim, LSI Engineering Director for Design Tools & Methodology J.C. Parker, IBM Microprocessor Design & Automation Manager Rex Berridge, and GlobalFoundries Director of Design Enabled Manufacturing Bob Madge.

If you read the previous paragraph carefully, you would see that these four speakers know a great deal about procuring/managing CAD tools and the people who use those tools, not to mention how to create an effective interface between tools, designs, designers, and fabs. Think you know a lot about this stuff? These guys know even more.

Kee Sup’s an expert on negotiating aggressively with tool vendors; J.C. can talk up a storm comparing Best-in-Class tools from small companies with Monolithic solutions from the big guys; Rex knows how to create the right mix of internally-generated solutions and those from third-party suppliers; and Bob knows it takes exquisite collaboration between the yield engineer, the test engineer, the reliability guy, the fab manager, the production planner, the designer, and the customer to make everything come together with guaranteed quality, on time and under budget.

All told, theirs is the kind of battle-hardened wisdom that every EDA vendor on the planet should be seeking out. The slides from the talks will be online as soon as Management Day Organizer Yervant Zorian can get it done. Maybe after this week when his responsibilities as DAC 2013 General Chair are complete.

No.1) Aart de Geus’ visionary talk at the outset of this morning’s plenary session. What a romp through the history of technology, and the history of history. Printing presses, steam engines, books, popes, ink, mills, factories, cars, boilers, pistons, transistors, metallurgy, computers, machine learning, implantable devices, sensors, dust, brains, biology, and Arthur C. Clark. All that, and EDA, in just 10 minutes!

Take-away? Disruptive technology is the push that starts the cycle; unforeseen applications of the technology is the pull that takes it to completion. Most cycles take 100 years; the current one we’re half-way through may take even less. Buckle your seat belt. It’s going to be a wild ride!

********************

Tags: Aart de Geus, Bob Madge, Cadence/Denali, CEDA, Ed Lee, Freescale, GlobalFoundries, Gutenberg, Hermann Kopetz, IBM, J.C. Parker, Jose Nunez, Kee Sup Kim, LSI, Rex Berridge, Samsung, Soha Hassoun, Solomon Golomb, Stars of IP, Stephen Woo, Synopsys

This entry was posted on Tuesday, June 4th, 2013 at 10:50 pm. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.