Prior to his current role, Meyer served as president and chief operating officer (COO) of AMDs Microprocessor Solutions division, where he had the overall responsibility for AMDs Microprocessor Business, including product development, manufacturing, operations, and product marketing. Meyer joined AMD in 1995 and led engineering for the AMD Athlon microprocessor. In April 1999, he was promoted as VP, Engineering. In 2001, he was named group vice president and general manager of AMDs Microprocessor Business, and in 2002 he became senior vice president and executive officer of AMD. Prior to AMD, Meyer spent nearly a decade at Digital Equipment, where he was co-architect of the Alpha 21064 and 21264 microprocessors. As one of the worlds leading microprocessor design innovators, Meyer was presented with the prestigious Association for Computing Machinery (ACM) Maurice Wilkes Award at the 2003 International Symposium on Computer Architecture (ISCA). Meyer graduated from the University of Illinois, where he received a bachelors degree in computer engineering. He also received a masters degree in business administration from Boston University.

You have been in this industry for a couple of decades now. How do you see the microprocessor business in the next decade?
The last two decades in the microprocessor business have been characterized by a technological focus on driving processor performance up to the Moores Law curve. The corollary to that has been that the performance of microprocessor-based systems has doubled as a result of two things. One, it has been a result of an architectural sophistication to the microprocessor through putting more transistors behind the problem, and the second is that the last decade was more about driving clock speed at the expense of power efficiency. What we saw probably five years ago was a corollary of Moores Law.

So, what we saw was that the industry, led by AMD, started to look for different ways to deliver value in the microprocessor-based project. Therefore, the first step that we took was to do three things simultaneously.

First of all, we thought of adding 64-bit capability to the microprocessor and, secondly, we changed the system architecture in servers by adding a memory controller and doing some other nice things to a lot better multiprocessor in the x86 machine. The third was to architect a microprocessor so that we can very easily add multiple cores to the device.

And we think that has got some merit only for a very limited time and then we as an industry have to look for still other ways to deliver value in a microprocessor device. The way we are going to do that is by being very careful about the devices we develop.

Dirk Meyer president & COO, AMD

How has the last decade been for the processor industry?
I think in the next ten years, what we are going to see is a rich variety of integrated systems that comprise different kinds of processing elements other than the task at hand, which is really a lot different from what the industry was working on ten years ago, when it was purely performance, and performance in the classic sense. So, I think, that trend hasnt even started yet. Our Fusion processors wont be out in the market for another two years and then we will see a rich variety beyond that point in time.

One of the things that have happened over the past five or six years is that graphic processors are now being implemented using rays of general purpose CPU cores that are not actually x86 CPU cores. They are more generic and simple but thats different from the situation five years ago, when the hardware was really specialized to the graphics problem.

Now that the hardware is for general purpose, there is a movement called general purpose GPU, which is to take that general-purpose graphics hardware and in conjunction with some other software, make that hardware more generally useful. I think this is going to bring a lot more variety in the devices we create.

Do you see the consumer actually demanding and playing a larger role in your business in the next 5-10 years?
For the IT industry five years ago, technology was really driven by the needs of commercial enterprises. Now we see that PC client device consumer applications have started to drive technology. It is really the consumer usage that is driving technology, and I am saying consumer because here we are talking about media, video, and graphics that tend to be more of a consumer-oriented set of applications. We dont think very much about real-time video conferencing on PC.

Apart from processor speed, how has AMD geared up to address energy efficiency and virtualization capabilities?
Besides processor speed, one of the things you can do to address energy efficiency is virtualization. Virtualization reduces the number of boxes in needed to deliver applications, related to virtualization, architectural decisions can be made which can be of great help. Fundamentally, the fact is that we have our memory controller chip, which also virtualizes the environment.

Looking forward to the piece of technology in our systems, called IOMMU, it is another piece of technology that allows virtualizing environments to run our hardware more efficiently. So, virtualization is really more of an architectural innovation exercise. Relative to energy efficiency in general, there is no one thing other than hardware, circuit design, etc.

What is AMDs strategy for emerging markets? And where does India fit in?
First of all, our goal is to outgrow the competition in emerging markets. Emerging markets represent a huge growth opportunity for our industry. For that we must and will focus intensively on resources in emerging markets. It is also interesting to note that we think we have got an opportunity to outgrow the market simply because the market is reasonably underdeveloped. Intel, frankly, is less entrenched. So, as the market grows, we will be able to grow faster.

The market cap of AMD is about $3 bn, which is much less than the $5.4 bn that you had paid for ATI. What do you think went wrong?
Clearly, AMD is terribly under-valued. The other thing I would say is stock valuation fluctuates like crazy in the near term. So, its hard to judge the wisdom of a long-term decision by short-term fluctuations in stock price. Beyond that, I would certainly say that 2007 was not a good year for AMD, was not a good year for any of the businesses that comprise AMD. We are going to do a lot better than that in 2008, and I think over the long term, the wisdom of the acquisition will show.

What steps are you taking to make sure that you have a better year in 2008?
First and foremost, we want to get our products out. We did a lot of good things in 2007 subsequent to a very bad first quarter. We improved our gross margins in every business. We got a graphics product out in the market and received very good acceptance.

The one thing that we didnt do well, which has really kind of overshadowed a lot of the positive, was that we didnt get our Quad Core products out. But we look forward to do that in 2008.

We will refresh the multi-core desktop products in the first quarter. We will start shipping substantial volumes in the first quarter, and beyond that we are going to provide mobile platform that is intended for more mainstream usage. We will get our 45 NM products out. Clearly, a big piece of success in our business is to get products out within a timeframe and hit the markets.

Would you be able to state the single biggest reason for delays in product announcements? Is delay the reason behind lower revenues as compared to the figures in 2006?
Not so much in the following sense if you look where we really suffered. As we add customers dramatically, and particularly we added Dell in 2006, our operations got fairly well streamed through a set of decisions that were the right decisions, and, at that time, we ended up sorting our distribution channel pretty badly. And those decisions hurt us quite badly in the first quarter of 2007, and the balance of 2007 was characterized by getting out of that problem.

If you look at 2006-07, there is no question that our product competitiveness changed. I mean we just had unbelievable dominance from a product performance perspective in both services and desktops, and that situation changed a little bit in 2006-07. That changed the richness of the product. But, more generally, I would say, the biggest problem we had was that from the unit perspective we witnessed a bad first quarter in 2007, and we took a couple of quarters to work our way out of it.

How is AMD doing in the government verticals in India?
Ninety percent of the Indian government tenders specify Intels CPU as the minimum requirement. So, we have no chance to compete. We are up against a huge barrier in the marketplace. We are witnessing slow growth in this regard. Recently, the Karnataka government also modified its tender documents to allow PCs with processors other than Intels.

Sudesh Prasad and Srinivas R
sudeshp@cybermedia.co.in

Page(s)   1