Warren Savage: Optimism for IP at DAC and beyond
May 5th, 2016 by Peggy Aycinena
Warren Savage, CEO at IPextreme, is willing to address questions regarding IP content at DAC 2016, enthusiastic in fact. That’s not surprising, given that he serves on the IP Track Committee that reviews the content.
“I think the content’s very good this year,” Savage said in a recent phone call. “We’ve been working on the IP content at the DAC for 3 years, and continue to make progress. I would say the biggest thing [we struggle with] is insufficient time allocated for IP.
“In comparison to previous years, however, the IP and Design tracks have been merged and all put under the same track – something we recommended against, because design-related submissions generally are different from IP-related submissions.”
“But as an IP consumer,” I asked, “isn’t the biggest problem knowing if the IP I’m about to buy is any good? Isn’t that the thing that should be addressed at DAC?”
Savage responded, “Over the past 20 years, there’s been considerable evolution in how people think about IP and I’ve been the megaphone for that change: IP is a product business and not a services business. It’s an idea that we felt was important back when I worked at Synopsys.
“In the early days of IP, people would try to sell some code labeled as IP. The customers tweaked and customized it for their own use, but it had no silicon provenance. What’s really changed over the last 20 years, people now have the product idea of IP. Customers have to have used it, so there’s certification.
“These are the kinds of things that potential customers look at. They need to be able to say: I trust this supplier and how this IP has been developed, and how it’s been verified. It’s this attitude that has raised confidence levels about using IP.”
“Even if IP has been certified by previous users,” I asked, “how do I know the block will perform well in my design, an environment that differs from that of previous users?”
Savage said, “There are two categories of IP, digital soft IP and hard IP. The latter does not change. Customers can be assured, for instance, that a microprocessor [verified] at 28 nanometers will not change.
“Soft IP, however, is somewhat more variable, and can change depending on the underlying libraries. A synthesizable processor core may run at certain speeds with certain libraries from ARM, but will behave different with Synopsys libraries at a different geometry.”
“I’m hearing great confidence today in third-party IP,” I commented.
“Yes, it’s quite stable right now,” Savage replied. “Over the last several decades, there was a first round of shake-ups in the industry and [the demise of] questionable IP companies after the dot.com crash.
“But the industry rebuilt itself in a structurally sound way after that. As a result, most IP companies even survived the 2008-2009 downturn.”
“If that’s the case,” I asked, “why is there so much angst about the IP industry, and the quality of IP, in the descriptions of panels that will be presented at DAC in June?”
“I don’t know,” Savage replied. “If you actually attend those panels, you’ll hear a far more optimistic attitude from the speakers. From my point of view, there’s nothing but optimism to think about in the IP industry.
“It’s by far the Gold Star market segment in the semiconductor industry. It’s growing faster than everything else, has a lot of longevity, and unlike EDA and some of the other market segments, it’s not mature and is still changing.
“That’s definitely different from the business of EDA today, where the design flows and tools used to make semiconductors are moving at a decelerating pace. We’re now working around the edges of [current design challenges], increasing tool capacity and cleaning up insufficiencies in parts of the flow for the smaller geometries.
“Also unlike EDA, IP is a content business driven by the consumer. Where EDA is driven by the requirement to manufacture the device, the IP industry is completely application-driven. In addition, IP is being driven by emerging standards, those for the auto industry and other new applications.”
“Are rumors of consolidation in the IP industry true or false?” I asked.
“That’s a myth,” Savage replied. “I know one of the great initiatives at large companies [is based on the idea] that there’s so much consolidation, including in the semiconductor industry, that the customers are saying, We just want to buy our IP from one or two companies. But that’s just the dream of purchasing agents, because that [strategy] doesn’t work in IP.
“The big IP companies work on collecting IP [for their portfolios] with the most revenue. But if you’re a semiconductor company and will be using IP for 90 percent of your design, the big companies can only supply your engineers with 80-percent of the chip. You will still need 10-percent from the smaller companies, the kind of IP the big guys haven’t yet seen the value of adding to their portfolios.”
“Seems like good news for the small IP companies,” I said.
Savage agreed: “Yes, as long as you have some differentiation, you’ll always be one step ahead of the big guys. That’s the unique thing about IP, it still offers market advantages to smaller companies.
“And there’s one other pitchy thing: The semiconductor industry has changed quite a bit over the last 10 years, VC money has slowed to a trickle. I have a lot of friends who [in a different time] would have left Intel, for instance, and tried to raise some money to start a semiconductor company.
“But it’s so expensive today, instead they change themselves into an IP company. That’s where all the great intellectual power of the semiconductor industry is going.”
“How about the changes at EDAC,” I asked “How’s the morphing of the organization into ESDA, the ESD Alliance, being received in the IP industry?”
Savage endorsed the change: “Historically, most of these organizations have been targeted at the semiconductor industry, GSA for instance, or EDAC for the EDA players.
“However, the ESD Alliance is [emerging] as an organization willing to put some of its energy towards helping the smaller IP guys, which is most of the IP community. There are 4 or 5 big companies in IP, but 400 or 500 smaller companies, where the innovation actually happens.
“In my view, getting the smaller IP companies into an industry-wide organization is something that’s been missing. And these small guys do need help. I’m working closely with [ESDA Executive Director] Bob Smith, because we both know there are a lot of things that can be done.”
Tags: DAC 2016, ESD Alliance, IPextreme, Warren Savage