What Would Joe Do?
Peggy Aycinena is a freelance journalist and Editor of EDA Confidential at www.aycinena.com. She can be reached at peggy at aycinena dot com.
PDF Solutions: The Rock of Ages
September 7th, 2017 by Peggy Aycinena
As well, CMU Professor Andrzej Strojwas [2016 Phil Kaufman Award recipient] has been PDF’s Chief Technical Advisor from the beginning — not surprising considering he served as Kibarian’s PhD thesis adviser at CMU — and Lucio Lanza has been on the board of directors for 20 years, serving as Chairman since 2004.
PDF is a company that defines stability, steady growth, and an intellectually rigorous and serious-minded approach to solving problems. It’s not a company of self-promoters or grand-standers. It’s a company of highly accomplished technologists, deeply involved in one of the toughest jobs in semiconductors: Finding out why chip yields are good, bad or ugly, and figuring out how that data might be used to improve design and manufacturing.
The last time I interviewed John Kibarian, it was 2015 and PDF Solutions had just acquired Syntricity a company with yield-improvement technology and services for the IC process life cycle.
This time when Kibarian and I spoke, PDF had just acquired several assets of Kinesys Software, including its ALPS (Assembly Line Production Supervisor) software, “designed to enable complete manufacturing traceability, including individual devices and substrates, through the entire assembly and packaging processes” – capabilities which PDF plans to integrate with their Exensio big data analytics platform.
WWJD: I’m particularly interested in your Dutch software acquisition that will allow the manufacturing history of a chip to be married to the device in which in resides, throughout the shelf life of the device. Why is this important?
John Kibarian: It is a good fit with our analytic software for chip manufacturing.
For fabless and system companies, a lot of devices have multiple chips in the package. For instance, a microphone, or a complex image sensor, or any number of different systems. If you’ve got a field return, you want to know exactly where each of those chips came from [to help diagnose the problem].
If you have a quality problem, you need to go back and trace.
The whole battery problem is [indicative]. We want to know which manufacturing facility did the bad battery coming from. To chase down all that type of information for chips is becoming mission critical. If the software stops running [on a chip], we must be able to help the customer understand why.
WWJD: What’s been the focus of Kinesys Software?
John Kibarian: The company’s first customers were people who made hearing aids, and then pacemakers. Then they worked in automotive, and now consumer devices like microphones for phones.
There’s been a need over many years for manufacturing traceability, but it has grown much more significant because of system in packages. The Kinesys software enhances the piece that PDF [contributes] by addressing everything that goes on, on the manufacturing floor.
WWJD: How has your relationship with the foundries changed in recent years, particularly with introduction of FinFET and SOI technologies, or is it just a more intense relationship?
John Kibarian: The Kinesys product doesn’t touch on the foundries, because they don’t do a tremendous amount of assembly. It’s more relevant to our Exensio platform.
For the foundries, the first important thing is the electrical characterization information. What differentiates this [from assembly] is the need to have information in several dimensions, more and more tight electric characterization, full parametric characterization, RF characterization, and noise characterization.
And number two for the foundries is Design for Inspection, in-line inspection info that is electrical. Originally it was a visual [process], but that’s no longer where the question is. Now it’s a question of leakage and reliability: Is this a partially open path, and not as good a conduction path as we had expected?
We [are working] to make the manufacturing data more design-relevant, to increase the type of electrical data that we can control.
We are going to the fabless companies and having them include in their dummy-fill areas, on-chip instruments that look just like parts of their products, which make that inspection process that much more efficient. And with [today’s machine], it is possible to measure those structures.
It’s very different from conventional inspection, where the designer had no input on what or how the [manufactured design] was inspected. Now, we are giving the designer a way to impact [the inspection].
WWJD: Do you have any competition in this area?
John Kibarian: I think we’re pretty unique in this area. Employing the exact technology is our expertise.
WWJD: EDACafe’s Mark Gilbert is very bullish about opportunities in EDA. Can the same be said for your industry? Do you have access to young talent? Are they coming to you looking for opportunities?
John Kibarian: There’s always need for help.
Yet there are CEOs in the industry who say Moore’s law is dead. Why would a PhD come here when Moore’s law is dead? The people who say it is dead are close to retirement.
But there are young people coming into the industry, and they are finding more and more interesting problems. It’s no longer about geometric scaling, it’s about electrical scaling – that is what’s driving innovation. There is a tremendous amount of ideas.
To say Moore’s law is dead to young people, and to us, is a very great dis-service to the industry!
Today, Design for Inspection is a phenomenally interesting technology. At PDF now, we have over 40 people working on this product, something that we did not have even just a while back.
Some of these people came from equipment companies, some came from EDA, some from analytics, some from drug discovery. They all come with different skills, and bring new and innovative perspectives to the problems [we are working to solve].
My point is: We and the industry have the need for new people with new ideas, as well as people who have been in the industry for a long time and have been burnt [by disappointment]. It is this type of diverse group that gets you real innovations.
We should all be looking at where the new opportunities are, and encouraging young people to pursue [those things].
WWJD: You’ve obviously been involved with the ESD Alliance for a long time, serving most recently as Co-chair of the Board? How do you see the benefits of membership?
John Kibarian: First of all, the organization changed its name from EDAC to the ESD Alliance, because the nature of the industry is changing.
Today, there are system worries about how chips are going to behave, and we can no longer dis-aggregate parts of the industry and expect to get the required performance benefits.
The problems are becoming richer and need a more diverse group of people. The larger issues in design automation have a lot more elements.
There is hardware/software co-design, issues of verification, and then there are folks like us at PDF Solutions that are not in conventional design automation at all. The vehicles that design automation provides for system development, means that now that we all touch the conventional design community in many different ways.
So why be part of this association or consortium? Well, there are a couple of things.
First, we all work really hard on building our products and our companies, and on supporting our customers. But we all need the larger perspective, and that comes from other people who are roughly in the same business as you are, getting together and learning together.
Second, there is the opportunity to understand the piracy problem together, or understand export law together, or the operating system issues together, because we all deliver software.
It’s really powerful to be able to communicate with the other companies in the industry, to understand what they are seeing and for them to understand what we are seeing.
It allows us to be more efficient for our customers, and to know what the bigger trends are — more so, than when you’re just driving your own [agenda].
There is a joke specific to EDA: If you look at stock prices as an overall proxy for health, the correlation coefficient between the design automation companies is much stronger than the variance between them.
The rising tide raises all boats. What we are doing is relevant to everyone in the Alliance.
WWJD: Has that success been obvious throughout the history of EDAC, the ESD Alliance.
John Kibarian: First, the stuff they did on piracy was phenomenal.
At board meetings or in general members meetings, you’d could see it in the eyes of the audience as they realized what was going on. We were all in the same boat, with everyone afraid to put “Call Home” into their products.
But the discussion within the EDAC team – an effort lead by Kathryn Kranen – and the way we positioned this together [produced a message] to our customers: You have to pay more because there are folks stealing the software. You should support our efforts on this as a paying customer, so everyone who should pay for the products is paying.
That transformation in thinking about piracy alone, was worth the membership.
WWJD: You’ve always maintained that you’re not an EDA company. Is that still true today?
John Kibarian: Yes, we’re not an EDA company. But remember that EDAC was renamed the ESD Alliance to represent a broader constituency, which includes the equipment companies.
For instance, today it was announced that LAM [wafer fabrication equipment], a capital equipment company, has bought Coventor [simulation and modeling tools for semiconductor process technology, MEMS, and IoT].
This is all very interesting and an indication of where the industry is going.
WWJD: This is a very global industry, but recent trends in regional protectionism possibly undermine that situation.
John Kibarian: My talk about Andrzej Strojwas at the Kaufman dinner last fall addressed that issue, and is why I wanted Andrzej to win the award. He is someone who doesn’t look at borders, and that’s why innovation has moved so far and so fast.
This is a global industry. There is no one country that has a corner on innovation in electronics, or in systems in general.
Our industry is really about being able to process information with less Watts and less dollars, while increasing the productivity of human beings and adding to quality of life improvements.
I chose those words in honoring Andrzej, because that is something about him that is really true. There is a broader lesson in his life, and one I hope that we do not forget.