DAC Monday 2017: Joe, Lip-Bu, China, Under 40s, Smurfs, and More

It's Monday, the first real day of DAC after the kickoff on Sunday evening. See Guide to Austin for Newbies, and DAC Sunday Kickoff for more details of yesterday.

In Memoriam

In his introduction, Mike "Mac" McNamara put up a slide with some notable people from the EDA industry who had passed away in the last year. You can see their names on the above slide. I had not heard that Mark Templeton had died. He was one of the co-founders of Artisan Components, which lives on as the physical library division of ARM, who acquired them. It seems that Mark was killed last summer in a tragic kayaking accident on the Rogue River in Oregon. I am always especially sensitive to people killed kayaking since I did a lot of whitewater kayaking myself in my 20s. I was President of the Edinburgh University Canoe Club and spent a fair bit of time driving minbuses with trailers full of canoes around Scotland during winter and Austria during summer (although the driving age is 17 in UK, you have to be 21 to drive a minibus with more than some number of seats, which makes graduate students in high demand).

Another was Charles "Chuck" Shaw. When I ran strategic marketing at Cadence, he reported to me and ran the technical advisory board and various other things. At the time, he was the oldest employee of the company and one DAC, in Vegas around 2000 or 2001, he had his 75th birthday. I arranged a small ceremony and Ray Bingham, the CEO, presented him a bottle of expensive Champagne and a hastily created certificate.

Sean and Paul's Excellent Advenure

Unlike the 1989 movie Bill and Ted's Excellent Adventure, this one did not involve time travel. But on Sunday during setup, we went to a few booths of interest to see how the 54th DAC was shaping up. Since this was done under the banner of ChipEstimateTV, we started on the ChipEstimate booth. But there were other booths that were more interesting. In the IC Manage booth, they had a Lamborghini. I happened to see on Facebook that the fire marshall would not allow the car on the show floor unless it was almost empty of gas, so several people had to spend their Sunday afternoon burning off a quarter of a tank of gas on the roads outside Austin. It's tough work, but someone has to do it.

Next, we went to ANSYS where Scott and Katrine were doing the routine as usual. I realized talking to Scott that when I hired him to present our products at VaST Systems Technology that it was the first time he had ever done a show at DAC—and he has had a gig with someone at DAC ever since, for many years with Apache Design and, post-acquisition, with ANSYS. It is hard enough to keep your concentration and talk naturally with a camera on you, but more difficult still when juggling clubs are flying back and forth inches from you. The question everyone always coming to DAC if they have kids (or if their name is Tiffany) is "What animal toy does ANSYS have this year?" Answer: bulldog. Yes, a cuddly bulldog.

The most interesting looking booth this year is OneSpin's "Austin Café". Since they are based in Munich, I think it is more like a biergarten than a café, but this being DAC it is a biergarten with no beer (or even bier), but there is coffee, so despite the beer barrels, that's what we will call it. You can watch the whole video:

Joe Costello Keynote

The opening keynote was by Joe Costello of Enlightened. As usual, Joe was spellbinding and had some interesting insights into what IoT businesses are likely to work out (huge markets) and what will not (internet-enabled dog feeders). One area is medical, in particular keeping track of medical equipment. Here's a statistic that I found mind-blowing. In most circumstances, you can't get discharged from a hospital except in a wheelchair. For example, when my son was born (in Scotland), my wife was discharged in a wheelchair to keep with the protocol. But, true story, the car would not start in the hospital car park—but we lived only about a quarter mile from the hospital so we just walked home. But that's not the mindblowing statistic. It is that a senior hospital administrator told Joe that hospital admissions are 9% higher than necessary due to inability to find wheelchairs to discharge patients in a timely manner, sometimes meaning they end up staying an extra overnight.

I will cover Joe's keynote in its own post on Friday.

One-on-One with Lip-Bu Tan

Ed Sperling interviewed Lip-Bu Tan, who I assume I don't need to point out is Cadence's CEO. I will cover this in its own post some time next week.

China's IC Industry: Today and Tomorrow

Ed didn't ask Lip-Bu anything about China's IC industry, although since he visits the country once a month and makes investments there, he would be well placed to answer. Professor Shaojun Wei presented an overview of what is going on in China's IC industry and what the country is doing about it. He didn't mention either of these points directly, but I think they are two datapoints to make you sit up and take notice. The first is that the value of the semiconductors imported into China is more than the value of the oil imported into China. The second is that there are at least 26 new fabs being constructed in China right now.

I will cover this presentation in detail in a post next week. China is not the largest market for semiconductors but it is hugely important. 60% of the world's semiconductors are imported into China and, strategically, China wants to build more themselves and is ready to invest to do so. One big motivation is that so much of the value is in the semiconductors. In an analysis he talked about, only 0.6% of the cost of an electronic device was captured through the assembly process, the rest was the cost of the components.

Under 40s Panel

Before the opening keynote, there were various award ceremonies. One was for the Under 40 Innovators Award. As you might guess, this is for people in the industry who are under age 40. There were five award winners:

• John Arthur, Research Staff Member and Hardware Manager, Brain Inspired Computing Group, IBM Research – Almaden: John designs large-scale neuromorphic chips and systems as well as algorithms to training them, including Stanford's Neurogrid and most recently IBM's TrueNorth.
• Paul Cunningham, Vice President of R&D, Cadence Design Systems, Inc.: Paul is responsible for Cadence's front-end digital design tools. Paul joined Cadence in 2011 through the acquisition of Azuro, a clock concurrent optimization company where he was a co-founder and CEO.
• Douglas Densmore, Associate Professor, Boston University: Doug creates electronic design automation inspired software tools for synthetic biology. He is a founding member of the BU Biological Design Center (BDC), head of the NSF’s “Living Computing Project”, and a Senior Member of the IEEE and ACM.
• Yongpan Liu, Associate Professor, Tsinghua University: Yongpan's research interests include design automation and emerging circuits and systems for IoTs. He designed the first nonvolatile processor used in both academia and industry. He received IEEE Micro Top Pick16 and best paper awards of HPCA15 and ASPDAC17.
• Sasikanth Manipatruni, Senior Staff Physicist/Engineer, Intel Corp.: Sasi merges physics based design with the experimental demonstration of spintronic/photonic/quantum devices for technologies beyond advanced CMOS. He has authored more than 50 scientific articles and holds 80 patents spanning nanophotonics, medical imaging, and quantum and beyond CMOS computing. He also coaches middle/high schoolers for Physics Olympiad.

Later in the day, there was a panel session, moderated by Dylan McGrath of EE Times, along with the five award winners. The format was that Dylan would ask a question and then go around each of the panelists and get their answer. Unfortunately, there was no attempt to get a discussion of any sort going between them. There was a good mixture of interests for that too, with a couple of people on the border of electronics and biology, a couple of people on novel devices and circuits, and Paul Cunningham who runs the front end of digital design (synthesis, high-level synthesis, test, power, and more), so right in the heart of what is being delivered to end-users today.

The first question to the panelists was what excites you the most. John went first and said that the slowing of Moore's Law means that there is lots of opportunity for innovation since increasing performance cannot simply be done by waiting...in particular low-power neural edge devices, which is his main research interest. Paul said that electronics is still the thing that changes our lives the most, and life is getting like living in one of the science fiction books he read as a kid. Plus, no other industry has so many complex and varied algorithms in one place, it is "an algorithmic candy store.". Doug is interested in computing-inspired biology (as opposed to biology-inspired computing), we are starting to be able to merge electronic and biological systems. Yongpan went with the slowing of Moore's Law too, since it drives his interest, the creation of novel devices like MRAM or flexible transistors. Since Sasi is from Intel, he had to say something about Moore's Law, too—as long as there is innovation there will be Moore's Law in one form or another.

The next question was what people thought was the biggest barrier to innovation, technical or otherwise. Paul thought it was to do with people, since innovation comes from people. We have a challenge that there are so many opportunities at companies like Facebook and Apple, but EDA is really exciting, too, and we need to do a better job of getting that out there. Doug reckons that the interesting place to be is on the barrier between two fields, but that doesn't fit well with the way academia works with junior faculty. He risked a lot to move into biology, effectively starting from scratch. Sasi thinks that we need open source tools since innovation is held back by the cost of the tools. He thinks we are thinking bureaucratically and maybe should only pay once a startup starts to get revenue. "There is no open source software "as good as what Cadence has," he said. John pointed out that any new architecture is very risky, because you need an ecosystem of partners around it, and if it doesn't catch on all that is wasted.

Dylan asked how electronic design would look 10 years from now. There was a certain amount of "if all you have is a hammer, everything looks like a nail." Doug thought it would be hybrid electronics/biology. Yongpan thout that devices driven up from the physical level would be important, and other aspects driven down from the application such as AI. Sasi said that by then all design would be automated and it would be time to automate the design of the design. John agreed, wanting to be able to say he has a big bunch of data and something he wants to achieve, and have the entire hardware/software system built automatically. Paul said that he came under fire in 2000 when he did a startup straight out of school. "Everything will be programming: software and FPGA." But it is the opposite and he thinks there will be a custom device for everything. By then, designing a chip won't be much different from writing software.

And the next thing in five years? Sasi picked quantum computing. John said machine learning is just getting started, it is all going fast. Paul said getting energy/compute down, maybe quantum, Josephson junction compute, near-threshold devices. Doug reckons wearable and throwaway electronics on the boundary with biology.

Dylan pointed out that all five of the panelists had made a contribution before they were 40. What did they expect from the generation coming up behind them, say 15 years younger? John said that one risk is that as things get more and more automatic, nobody understands how it works under the hood. Paul said "hard work." Of course the ability to innovate is important, but you need to work hard and have ambition. Improving the quality of life is not just sitting on the beach. Doug went for risk taking. It will never be so easy as when you are young with no responsibilities. Yongpan thinks the key is to get the direction right, and then work continuously on it. But to be aware of where opportunities might arise (for instance, don't do research on CMOS, since most things that can be invented have been). Yongpan reckons creativity. That is the area where humans are relevant. He also suggested talking to your teachers since they are a great resource.

The final question was what was he most important lesson each of them had learned in their career. Paul said it was getting used to managing a large organization with many ways to innovate. You can't prescribe just one with a large team. It took him a long time to learn that. Doug said you should strive not to be the smartest person in the room. Find people who know more than you and be humble and ready to learn. Yongpan empasized that "hot" areas are not necessarily good choices for research, since there is lots of competition. Sasi suggested mixing up diverse application areas, even if you don't know anything about some of them at first. John emphasized again taking risks. Hard problems can lead to significant breakthroughs.

Smurfs

The Cadence booth looks great with its new video wall. Although one thing I have discovered is that you can't take a photograph of the screen on an iPhone, you just get trippy interference patterns. This year Cadence is giving away Smurfs. There are several, so come by and collect them all. Based on past years, I think you can expect Lip-Bu Tan and Nimish, at a minimum, to appear on stage during the Denali party in some sort of Smurf costume. So don't miss that. The Denali party is in a new location this year, very close to the convention center, Palm Door on Sixth, at 508 E 6th Street.

HOT party

It's hot in Austin...although I hear that it is actually even hotter in San Jose, which is definitely unusual. But tonight was also HOT, the Heart of Technology party for charity, this time for a SJSU Scholarship in honor of Gary Smith. As usual front-man for HOT, Jim Hogan, was also lead guitarist for one of the bands. I didn't get there except for the last 15 minutes so I can't honestly tell you much about it. But here's a picture of the band anyway.