On the record: Gary Smith
The Oracle at Delphi has nothing on the Chief EDA Analyst at Dataquest.
[Editorís Note: An abreviated version of this article first appeared on-line in EDA Vision in July 2001.]
Starting and ending with the Tao is pretty enigmatic stuff when, in the middle of the stream, you find a bass-toting, black-leather-clad blues musician fresh out of the Naval Academy living in a shack in the midst of Silicon Valley.
That pretty much summarizes Gary Smith for those who know him. For those who donít, to quote from an introduction to Gary I heard at a panel last year where he was acting as moderator, "If anyone in this room doesnít know who Gary Smith is, they donít belong in this room."
For a number of years, Gary Smith has been (and by the looks of things will continue to be) the single most important prognosticator in EDA. The industry listens to Gary, at DAC and a thousand other venues over the course of the year. They bank on his annual numbers reporting on the health of the industry. They pin his EDA Landscape poster up on the wall to keep track of which companies are which in the here today/acquired tomorrow world of EDA. They take their business plans and nascent product ideas to him and hope for his blessings. They quote him. They court him. They keep him busy. And, apparently, he loves it -- taking all of the adulation in stride with a smile and a nod. Which is what you would expect from a guy who takes Eastern philosophies seriously and incorporates them into his mindset and lifestyle.
The rest of Garyís story is as follows. However, if you believe as Gary does that "less is more," you neednít read on. Based on what youíve read, you already know him.
Chapter 1: I am born
Smith was born in Stockton, California, and grew up in the San Joaquin Valley. He claims with a laugh that he knows the whole story of George Lucasí American Graffiti as if it were his own. I pointed out that the movie was set in Modesto. "Oh well," he chuckled. "Youíve seen one Valley town, youíve seen them all."
When college beckoned, Gary was accepted at U.C. Berkeley and then received word of his appointment to the Naval Academy in Annapolis, Maryland. How to decide? It was simple, he says. He looked at a map -- 120 miles from Stockton to Berkeley, 3000 miles from Stockton to Annapolis. He opted for really going away to school and "enlisted." (He acknowledges that Annapolis is one of the most beautiful communities in the world and, if he were not so happy in Northern California, thatís where he would settle.)
During his plebe summer in Maryland, the Navy rowing team went to Rome to represent the U.S. at the Summer Olympics. On the night before their event, the team did as the Romans do, enjoyed a bit too much Chianti, and attempted to "borrow" all of the flags of the various countries in attendance. They were -- as could be predicted -- caught, jailed, quickly bailed out, and still hung-over the next day while rowing. Naturally, they came in dead last and returned in shame to Annapolis, facing a yearís worth of KP or worse. But, to their surprise, when they arrived on campus they were greeted by nothing short of a standing ovation from their shipmates and peers. Gary tells this story with relish, knowing full well that his secret is out: Garyís a rebel and proud of it.
Roger Staughbach was one of Garyís shipmates at the Academy, an indication of the depth of talent on the football team during Garyís years there. Those were good years: Gary majoring in electrical engineering (note: the Naval Academy Degree, at the time, was called "Marine Engineering" and had five sections; Mechanical Engineering, Nuclear Physics, Aerospace Engineering, EE and Computer Science), rooting for a victorious football team, and preparing for a career in the military. Although Garyís seen a few passages in life since his college years, he still sports a U.S.N.A. sticker on the back of his car. Once a Mid (as in "Midshipman"), always a Mid.
Unfortunately, reality set in and by the time Gary was stationed on a Naval destroyer for the fourth time in six years off the coast of Vietnam, he was ready to make a change in his lifeís plan and become a civilian at first opportunity. When he came out of the Navy, he went to see fellow Academy graduate and headhunter, Joe Daniels, for some career counseling. Gary already had a job offer in his pocket to work for McDonald Douglas as an aerospace engineer specializing in power design for electronic systems: "Youíre desk will be on the top floor of the hanger, tenth row down, fifth desk over," he was told.
Not surprisingly, Gary wanted to listen to Daniels who countered with a different proposal, "Iíll bet you donít really know what you want to do, but I do. You want to be in semiconductor sales. Trust me on this one."
Gary was sure the Daniels was mad: "Iím a power guy, and I donít know anything about semiconductors or sales." But Daniels prevailed, and Gary tracked down two different job offers along those lines, one with Texas Instruments in Dallas and one with National Electric in Geneva, Ill. He opted for National Electric and semiconductors became Garyís lifeís work. (Parenthetically, he admits that he didnít "feel comfortable in semiconductors until we went to MOS-voltage forms because they were equatable to tubes.")
Breadth of a Salesman
Gary thoroughly enjoyed the work at National. In those early years when he was selling SCRs, diodes, and power tubes, sales folks also served as application engineers, integration engineers, and over-all walking experts on the use of semiconductors in electronic systems. The expertise he developed at National allowed him to be recruited by International Rectifier in El Segundo, CA. Gary was with IR until he was recruited into TRWís marketing organization in Lawndale, CA, working with Q.T. Wiles, later made infamous by the MiniScribe scandal; a driven man who Gary credits with developing early semiconductor-level technology and marketing.
Q.T. unfortunately ended up in jail, but not before putting his stamp on the industry. His management style was unambiguous, according to Gary, "You pleased Q.T. or you were history." Fortuitously, Fairchild Semiconductor then offered Gary a position that he happily accepted, as it required a move to Silicon Valley. Unfortunately, the hiring manager left before Gary arrived, so he accepted an alternative offer at Signetics (Sunnyvale, CA) instead, working in Central Marketing.
Eventually, it was time to move on again. This time he had an opportunity to be part of his first start-up as a Sales Manager for Telmos in Sunnyvale, an enterprise spearheaded by Fairchild co-founder, Jean Hoerni. Hoerni invented the planer transistor, an idea that helped trigger the development of the integrated circuit, and was one of the eight legendary scientists and technologists who left William Schockley in 1957 to found Fairchild Semiconductor. Garyís experience at Telmos was a positive one, everyone full of zeal and optimism. Gary lasted 5 successful quarters with the company before it was time to move on. He came out financially sound, however, by selling his stock upon departure.
Gary was lured south again by Plessey to serve as Business Unit Manager for ASICs in their Orange County offices. It was at Plessey that the idea of getting back into full-time engineering first came to Gary. He had lost his Chief Engineer and had spent months trying to replace him. Gary served as acting Chief Engineer in that period and discovered that he liked that role a whole lot more than he liked running the division.
Additionally, he discovered that he missed Silicon Valley, so he moved back north as Sales and Applications Manager for ES2 in Santa Clara. That led him to his work as a Sales and Marketing Manager for IMI in Milpitas, CA, and then to a stint at LSI Logic, also in Milpitas. While at LSI, Gary decided to "retire" from management and go back into actual engineering, serving as a methodologist for the company.
Concluding this part of his story, Gary points out that he averaged 3 years per job over the course of his long journey through technology, far higher than the standard metric of 2.4 years per job for the typical working stiff in Silicon Valley.
Finally in 1994, Dataquest nudged him into his truest calling -- Industry Analyst. Since that time, Gary has applied his background in semiconductors, which coincides with the complex history of the industry, to his role as guru for the most fascinating niche in the business, Electronic Design Automation. (The role of guru is often the one he describes as his "real retirement.")
Has Gary enjoyed being an analyst? "Yes. As I always say, itís great to be 'retired.' Now, if I could only get my hours to under 50 a week."
Pressed to bypass modesty, Gary tries to explain why heís one of the "leading indicators" in EDA. "Well, John Marron once said that I was right 80 percent of the time -- probably a bit exaggerated. Actually, I steal 80 percent of my ideas. But itís not a coincidence that I can make the calls in EDA. I came out of the Power User community and I know a lot of really smart people. All I really do is communicate those ideas in a half-way decent manner."
Gary feels that the take-away message from DAC 2001 in Las Vegas is clear: "Standards are finally here." Heís got the long-view with which to make such claims.
"In DMV days [Daisy-Mentor-Valid], the companies had become old and stodgy. Then, Cadence and Synopsys made the jump to RTL and they provided the tools. [Out of DMV], only Mentor survived because of its military customer base, a situation that was both good and bad for the company. [Unfortunately], the military did not make the move to RTL, and Mentorís always been good at doing what their customers tell them to do. ĎWeíre going to do gate-level [design] for the rest of our lives. Donít bother us with RTL,í was the militaryís message to Mentor. Thatís why Wally [Rhines] was brought in to turn that situation around."
And the situation in EDA today -- is there room for a couple more big players? "Absolutely," according to Gary. "Thereís market space available in the applications area. Itís a leadership issue."
Although heís not willing to bet money on the winners, he says that Monterey, Magma, and Sequence certainly all have a shot at joining the ranks of the first tier EDA companies -- Cadence, Synopsys, Mentor Graphics, and Avanti. Gary notes that Roy Jewell, new president and CEO at Magma, and Vik Kulkarni, COO at Sequence, both spent time in Avanti, their tenure at that now-troubled organization coinciding with a time when the company was "in its heyday" and enjoyed solid leadership. "At that time, I had my money on Avanti being the next big leader," Gary admits.
Making the Tough Call
Commenting on possible problems ahead for the current EDA leaders, Gary says that Cadence has never in its history had a stronger management team than it sports today, "But unless they put their R&D team together, theyíll fall behind." Additionally, he says that it appears that "Synopsys may not understand the software implications of electronic system-level design the way Cadence does [with its VCC offering]."
Business forecasts and P&Ls stand nervously in the spotlight this year across the entire semiconductor food chain. So where does Gary think EDA is headed, financially, over the next 4 quarters considering the current downturn?
Heís sticking with the ballpark figure he threw out at his Dataquest presentation at DAC 2001 in Las Vegas. "I still see 19.4 percent growth, so Iíve lowered it by .6 percent. [Clearly], the semiconductor industry is in the worst downturn in its history, the communications industry is in the dumps, and all of the other industry segments -- with the possible exception of military/aerospace -- are in trouble as well. So, the question is: How will all of this affect EDA?"
He answers his own question at length: "In a market as complex as EDA, there are many sub-markets, all with their own business cycles. If you take a quick look at EDA, you can identify 54 of them. [Meanwhile], today there are two cycles that are driving the industry, the Semiconductor Cycle and the Semiconductor Process Cycle. The Semiconductor Cycle probably should be renamed. In essence, itís the Idea Cycle. The User Community running out of ideas, or at least not putting them on the market, causes any large, broad-based semiconductor downturn."
"Any really hot market, [like we saw in 2000], tends to slow down product innovation. The answer then is to design yourself out of the recession. This [will cause] an upturn in EDA spending for around two design cycles, two to three calendar years."
"The Semiconductor Process Cycle is driven by the Semiconductor Roadmap. Today weíre building .13-micron semiconductors. EDA tools, especially IC layout tools, need replacing every two process cycles. We are presently late with the tools needed for the .13- to .10-micron processes, which has kept demand for EDA tools extremely high, despite (or really because of) the present downturn."
"That being said, there are some things to watch for. Obviously certain sub-markets are affected by the downturn, the emulation and acceleration sub-app being the most obvious. These are large capital expenditures and always are closely monitored by finance. Also, consulting projects are almost always cut [in times like these]. If the recession goes on too long, youíll see companies starting to cut into critical parts of their organization just to keep afloat. Thatís when engineers go, usually after a long recession lasting a year to 18 months. This one is 10 months old as of now. Also you will see EDA spending cut at companies that use the component-based design style. For those companies, design isnít as critical a component to their corporate survival and that could impact PCB and FPGA tool sales."
"The EDA forecast is based on the assumption that this recession will not last into 2002. However, if [it does], the numbers will be impacted [accordingly]. But for now, with the demands caused by the Idea Cycle and the Semiconductor Process Cycle, the EDA market is [actually] in an enviable position."
Garyís nothing, if not an optimist.
The Music Man
Heís also a musician.
Back in Stockton, Gary was a drummer until he hit the 8th grade. Thatís when one of the bass players in the school orchestra "accomplished the impossible" by actually breaking the "unbreakable" E string on the instrument. Gary guesses that the kid was probably injured by the destructing string, which would explain the boyís sudden departure from the middle-school orchestra.
Gary was asked by the conductor to step in and assured Gary that playing the bass was easy. "Iím only teaching you three notes on the instrument," the conductor told Gary. "When I point at you, just go bam, bam, bam." Clearly, there was more to playing the bass than these instructions would indicate, but the conductor said that, for the remainder of the performance, all Gary had to do was to fake it, to appear to draw the bow across the strings without making contact. That way it would appear that Gary was playing, but could instead be staying out of harmís way, musically speaking. The ruse was successful: During the performance, Gary played his occasional "bam, bam, bam" as instructed and the rest is history.
Garyís a member of the Los Gatos Blues Band and sometimes-bass player with the industry music standard bearers, the Porch Dawgs. Check outwww.losgatosbluesband.com. There youíll find, among other useful images, two photos, dubbed by Synopsysí Karen Bartelson as ScaryGary1 and ScaryGary2, of bass-slung-low Gary "Marlon Brando" Smith at his best and baddest.
Meanwhile, a new band debuted at DAC 2001 in Las Vegas, The Full Disclosures, a 5-man assembly including Gary on bass, Aart de Geus on lead guitar, Don MacMillen on harmonica, Arne Lang-Ree on drums, and Grant Pierce as singer and guitarist. Gary says their successful gig in Las Vegas validated one of his longest-standing theories: The singer always gets the girls. Gary served as lead singer on several numbers at DAC in Las Vegas including "Every day I have the Blues" and was sought out, after the fact, by a groupie specifically referencing his fine rendition of that number.
In the End
Gary is open about his life-long interest in comparative religions. Heís widely read on the topic and has a remarkable range of knowledge and familiarity with a number of different sacred disciplines. When asked to provide a few pearls of wisdom with which to draw our conversation to a close, he offered these:
"There are three illusions: perfection, control, and knowledge. The best rule of conduct is to laugh at everything."
"In order to know the Tao, it is necessary above all to not think. Man is born with free will. Donít use it. Follow the Tao."
So donít be surprised if the next best seller to come out of Silicon Valley is a philosophical treatise: The Tao of ASIC Design. Donít hold your breath, however, for The Tao of FPGA Design. Remember this is an ASIC guy whoís publicly predicted the pending demise of the FPGA and only time will tell if Smithís right.
Meanwhile, itís clear that the Oracle at Delphi has nothing on the Chief EDA Analyst at Dataquest. After all, the Oracle never belted out the blues on bass, clad only in black leather and a smile.
[Editor's Note: Gary Smith is engaged to be married to Verisity's Lori Kate Calise in July 2004. Congratulations to the bride and groom!]
May 21, 2004
Peggy Aycinena owns and operates EDA Confidential. She can be reached at email@example.com