AMD on the Rebound: A Chat With Henri Richard

We met up with Henri Richard, the Executive Vice President, Chief Sales and Marketing Officer of AMD to have an update on microprocessor industry developments, in particular, his views on open standards, competition ethics and to learn of AMD's new plans, products and roadmap as an integrated AMD/ATI entity.

Introduction

Two years ago, we interviewed Henri Richard, the Executive Vice President, Chief Sales and Marketing Officer of AMD at the peak of AMD's success in the microprocessor market. Our article then was titled - 'Championing 64-bit & Dual-Core' - and at the time AMD was the undisputed leader in both technologies.

Today, Intel has fought back with a vengeance and the microprocessor wars have again heated up to a boiling point. In a short span of time, there have been many surprising turns of events and industry-wide movements that has propelled AMD into the spotlight such as their bold move to buyout veteran graphics giant ATI. However, AMD itself has been keeping a low profile as of recent times in the desktop microprocessor space. The world is just waiting to see if they can bring up their game with their upcoming next generation processors, chipsets and GPUs or will they fade into obscurity as with others who have dared the wrath of the '800-pound gorilla'.

www.hardwarezone.com® met up with Mr. Richard again for an exclusive interview and this time, we discussed his views on the microprocessor market, open standards, AMD's challenges and plans for the future, plus the promotion of FUD (Fear, Uncertainty, and Doubt).

HWZ: Now that the AMD/ATI merger is behind you, what can we expect from the new AMD? Basically, a roadmap update if you will.

Henri Richard: Well, it's true that the merger has been completed and the engineers have started to get together to work on the future. We have, and I believe, very exciting products coming along and you've already heard some of then, like Fusion. Clearly with Vista coming into the market, there has been a paradigm shift in terms of how users are going to be dealing with their systems. If you look at the legacy of PCs, they were really engineering machines, designed by engineers for engineering work. Back there and then, when the first PC was created, nobody would think that millions of people would carry around a notebook, watch movies, create videos or listen to music. So if you really think about it, in a very short period of time, the purpose of the PC has changed dramatically.

And where initially what was important was computing power, today increasingly, what is important is how you relate to it - visually, through audio and of course through computing. There is a battle between the various elements of the computer that has changed very significantly since its birth. One of the reasons that we did the merger with ATI is recognizing that in the future, the necessity of a balanced platform with high performance processors, high performance graphics, great audio capability and great video capability would be as important as just having a good processor. ATI brought to the party not only leading edge graphics capabilities, but also leading edge video capabilities.

So the first level of integration that you can see from a road map perspective is just bringing some of those new technologies into the market in the form of chipsets, or what we'd like to call our motherboard GPUs, and the 690 that has just been launched is the perfect expression of that with AVIVO video technology embedded. As we continue to roll out new generations of chipsets, the next generation which should be out towards the end of this year, which is the 780, will have much more integration with the linking of our design engineers both on the CPU side, GPU side and Video side.

Long term, we talk about our view where the world has moved to a heterogeneous course where several computing units will be tapped to do different things (by the way this architecture isn't completely new since this is how game consoles are working today). We think that in the future, we'll see interesting things on the same die, an x86 instruction unit as well as a GPU. There are a lot of interesting things you can do in terms of power savings, when we have both cores close to each other. And of course also, within the same power envelop, you can do system partitioning of computing performance based on the usage model of the computer it is sitting in.

So, I think that's another dimension of the industry that is changing. It used to be one size fits all and I think that increasingly, because of the usage model moves away from one size fits all, the scalability of processors is not going to be in frequency or number of cores, but how well it is adapted to do a particular workload.

Delays and Expectations

HWZ: Talking about Fusion, some think that Fusion is more of a lateral move to create a new market segment, an evolution instead of a revolution. Can you comment on AMD's hopes for Fusion and its impact on the industry?

Henri Richard: I think that was described really well. We do believe that it will create a new category. There is a lot of intrinsic value if you can position yourself as a leader in a category. Today, when you look at the market space, there are only three companies our usual competitor, whose agenda is still very much to ask users to spend most of their dollars on the CPU, and the other company, whose both our partner and in some case our competitor, whose agenda is to say the GPU is the most important part of the system.

The way AMD is positioned is to say, Well listen, let the users choose. Let's give the market, their customers and their end-users the opportunity to differentiate their platform with a mix and match of CPU/GPU/chipset capability.

So, we believe that is the leadership position. We believe that the competition will have to follow. We believe that the product of Fusion not only changes the game and creates a new category, but frankly continues to provide the market with a cost effective solution. Fusion will not only be a high performance product, but also can be a cost effective product because there is a certain amount of simplicity that comes with combining the various technologies.

HWZ: With the merger, it is understandable that certain products have been realigned, but could we have an update on the R600 delays?

Henri Richard: The R600 will be out in the second quarter. The reason we decided to delay the launch was that we wanted to have a complete DX10-enabled solution top to bottom. A lot of people wrote that the reason it is delayed is because of a problem with the silicon, but there is no problem with the silicon. We are demonstrating it. We can ship it today. But if you think about it, looking at where the market is at, the volumes are going to be in the R610 and R630, so it makes sense for us to do a one time launch of the entire family of DX10 enabled products. That meant delaying the R600 for a few weeks, but frankly it doesn't make a difference in the life cycle of the product and talking with our customers and partners, they felt that it would make a bigger impact with one full launch. So we decided to do that.

Also increasingly in particular with Vista, as we've seen with the competition. It doesn't matter if you're shipping the silicon if the drivers are not stable. There is nothing more frustrating than having bought a new graphics card and having your system crash repeatedly because the drivers are not ready. Although we today, even by Microsoft standards, have the best and most stable drivers in the entire industry. The few weeks will give us even more time to continue improving the drivers. Again, the decision lies in the fact that we will have a top to bottom DX10 offering with drivers that will have a very very high level of stability and the only difference is a few weeks. So it seems to make a lot of sense to do it that way.

HWZ: Will we see a revival of AMD as a chipset manufacturer in full force? Will this impact your relationships with close platform partners?

Henri Richard: Absolutely. Though that does not necessarily mean we are not going to keep a completely open ecosystem. We do. We are very, very satisfied with our relationship with NVIDIA. In the commercial space, our customers are looking for a one stop shop, and although we don't have any aspirations to have a 100% share of the market, we believe that there are a lot of customers who will want to have AMD products on AMD chipsets. That is logical. So we are re-investing in that business, we are making sure that the chipsets we bring to market are best of breed. But, in the same time we will not do anything artificial from keeping our partners like NVIDIA from being successful.

HWZ: Will AMD be developing complete integrated solutions ala Intel Centrino then?

Henri Richard: Yes, we will. Absolutely. Though we probably won't bundle them the way Intel did because we want to welcome competition, but we will have full solutions absolutely.

HWZ: Wouldn't it be counter productive to your efforts in fostering an open ecosystem?

Henri Richard: I'm a big believer that the best way to have great products is to have competition. So if I start to create artificial barriers for my chipset division to not be in tune with the real competitive nature of the market, I take a risk of having inferior products. By fostering competition, I force my people to have the best product. And if they do, my partners are then forced to innovate as well and their innovation breeds further innovation. I believe in the virtuous circle of competition. At anytime you put artificial barriers on competition, you really are hurting yourself more than you are hurting your competition.

On Open Standards

HWZ: What is AMD bringing to the table then to further open platforms?

Henri Richard: Well, the latest one that is very exciting is the DTX, that I'm sure you've heard about. It is our new small form factor. It is very exciting, with lots of great support from our partners. DTX will enable much smaller form factors, while at the same time reducing the cost of the solution. It is an interesting combination where we are taking a very customer centric approach, allowing an infrastructure which is more in line with what people are looking for today in terms of the footprint of their systems, while also addressing the cost structure of the industry in trying to bring the same level of quality and a more cost effective solution. I think that is why it is getting so much good press and support because it is both a good standard and cost effective standard.

HWZ: Do you think DTX will be a success then? The last industry platform revamp (Intel's BTX) didn't really take off at all.

Henri Richard: BTX didn't fly anywhere because BTX's objectives was for Intel to be able to deal with the Netburst out of control power output. So, BTX was never meant to create a smaller chassis or create a more cost effective solution. BTX was a way for Intel to deal with the fact that the Pentium 4 was out of control. That's why it failed.

DTX has nothing to do with that. It is built from the ground up for 45-65W power envelops, for energy efficient processors, and built to be more cost effective with ATX. It is backward compatible as well, so it is a very seamless transition. It is very clear today that every single business partner in the world, large OEMs and global system integrators are looking for smaller chassis and we're bringing to them exactly what they want.

HWZ: Open standards sound great to consumers, but how does it relate to the enterprise?

Henri Richard: When you look at enterprise user, they are looking for stability in a platform and we agree that there is a conflict of interest there between stability and open platforms. But enterprise users also want cost efficiency and competition. It is difficult to compete on proprietary architecture and if you look at it, the best example that open platforms is the right thing for the enterprise is the success of industry standard server and how the x86-based server has taken out bit-by-bit pretty much all the proprietary infrastructure.

There are only a couple of proprietary infrastructures left in the market and they're losing market share every year. The fact of the matter is that the economics of scale, the dynamics and the level on innovation on the platform is very difficult to beat. Yes, there is a downside to it with less stability, but it seems to be that the market is willing to take the lack complete stability from open standard platforms as a fair exchange for the amount of innovations, competition and dynamics that exist in the open platform. I don't think that is going to change.

On Quad-Core Computing

HWZ: The move to quad-core computing has accelerated at a much faster rate because of the pressure Intel has put on...

Henri Richard: Yes, the quad-core situation is kind of interesting. In the server space, in certain workloads, customers will really love it, particularly our solution which doesn't have the FSB limitation.

In the desktop, I'm a lot more skeptical. Of course there is always the marketing angle and you can convince people that four is better than two, but the fact of the matter is that there is no application that will take advantage of four processors, except for a very, very small sliver of the market space. And so, we have to be careful as an industry and not get trapped into the state of shipping a lot of technology that gets completely unused.

So if you ask me, today, in the desktop and notebook environment. Instead of spending more money on two cores that don't get used at all. I would advise customers to spend money on two cores that will get used and a good graphics processor to get the full benefit of the system.

Again, I think this is an area where our strategy differs from Intel, because we don't want users to spend money on technology that doesn't get used at all. Yes, there is a small percentage of desktop users that will see benefit from quad-core. That's why we made the Quad FX platform. That is a very niche and elitist enthusiast group of people that represent less than 1% of the market space. For the rest of the market, driving the message of quad-core knowing that people will never get benefit from it is really irresponsible.

HWZ: Interesting thing about the Quad FX. AMD's Quad FX platform has been received with tepid response as a hunkering, power hungry beast. Will Barcelona deliver the victory AMD is looking for?

Henri Richard: Well, there are two ways to look at the Quad FX platform. People who are looking at it and think that it is an attempt at better benchmarking are wrong. That's not why we did it. We did it for two reasons.

First, I believe that the ATX form factor is completely out of runway, given the parallel roads now needed inside a system. The ATX motherboard from factor was never conceived to put four graphics cards and two processors. And we found that it was time to move to the next level and that's really what the Quad FX form factor is about. It is a motherboard that has room for more improvements. It is a product that is made form the get go as a workstation. If you look at the Quad FX, it looks more like a workstation motherboard than a desktop motherboard, that's what the Quad FX is about.

As far as benchmarks go and being a great performance product, Windows XP doesn't manage NUMA architecture. So the fact is that we have two processors with their dedicated memory not getting any benefit from the operating system. Vista does manage NUMA, so the same benchmark on XP are significantly different on Vista. Quad FX was never meant for XP, which we know is basically EOL, but it was made for Vista. None of the benchmarks today that have been published have been on Vista.

I'd like to point out that when Barcelona gets out, that will give you 8 processors on two sockets. That's something.

Again, that will only be for a small portion of the market, but for that portion of the market, its performance will effectively be untouchable, which is essentially the equivalent of having a high end Opteron server on your desktop. Quad FX will be there to stay, it will continue gaining traction. It is a very high-end solution that we're not trying to sell hundreds of thousands, but it is a direction that the market will have to go to because ATX is just out of gas.

64-bit, Same Question, Different Year

HWZ: Let's talk about 64-bit. It has been 5 years since AMD's 64-bit processors hit the scene and 64-bit adoption has still not taken off. In our previous interview in 2005, you speculated that 32-bit will be a thing of the past come Vista, but there were no reasons to go with Windows 64-bit before and there are still no compelling reasons to use Vista 64-bit today � What are the factors stopping 64-bit computing going forward?

Henri Richard: Well, it has been a disappointing experience. Of course, on the server side, everything has moved on to 64-bit. On the client side, Vista was - is a great opportunity to move to 64-bit. Microsoft did a lot of work to ensure a full fledged operating system with its own drivers and all the compatibility needed. But we're thrust into a situation that is very difficult for our OEM customers.

The PC OEM market today is one where it is very difficult to make any money. There's a reason for that and that is the monopoly and tax imposed by Intel. So because it is very difficult to make any money for the PC manufacturer, it is difficult for them to invest in true innovation and difficult for them to accept the cost that comes with progress.

If you're running a large call center today, it is easier to keep your customer based on 32-bit than to move them to 64-bit even if you know its a better solution than to deal with the issues that arise from compatibility. Even though with Vista, it is greatly reduced.

So the issue of moving to 64-bit is not a technical one. It is not even an issue of applications because I can show you even simple applications such as HD video editing software that works much better, a lot better on 64-bit than 32-bit. The issue is one of economics and the fact that this industry is really plagued by the lack of healthy profits that enable our customers to really drive innovation.

Microsoft did their job. There's a full Vista 64-bit OS, which is the same price as the 32-bit. We can't really ask more of them than that. Now, what we need is for the rest of the ecosystem to move forward and to do it because they know it is the right thing to do. I don't think you will meet anybody that will say 64-bit isn't a good idea, but you will meet a lot of people that cannot afford to take it to the market. That is a very sad state of affairs of our industry. And frankly I think that is where the competition is lacking in their ability to be a leader, because instead of helping to expand the market, they seem to be very focused on trying to hurt me.

HWZ: Is AMD still actively trying to evangelize 64-bit computing?

Henri Richard: Absolutely, we're continuing to do so. We have a very large majority of processors out there that are 64-bit enabled, more than our competitor. We don't ship any 32-bit products to speak of except in our embedded business for the legacy side. In the open PC market, we're shipping 64-bit products and we'll do our best to move the agenda of the industry to go with 64-bit computing, but it is not something AMD can go alone, we need to support of the industry.

Competition and Ethics

HWZ: In your opinion, what are the issues dodging the microprocessor industry at the moment and how is AMD going about it?

Henri Richard: There are two issues. One of them is finding applications that can take advantage of the computing power that is available. I'll draw a line between server and client here. On the server side, there is always a need for more performance and the sky is the limit. The biggest challenge is power and performance-per-watt.

On the client side, conversely, if you look at applications, they really haven't changed fundamentally over the last five years. It's the same Excel spreadsheet, same Powerpoint. Yes, you have HD video that is driving demand for computing power, but fundamentally, we're running out of really demographical value for the end user. I always like to say, if you press the Enter key and the answer is instantaneous, twice as instantaneous doesn't bring you anything.

So, the challenge in the client space is to deliver better user experience. And I'm convinced that better user experience doesn't mean faster processor. It may include a faster processor, but there are a lot of dimensions that are just as or even more important than a faster processor. We're at an inflection point where it is going to be a different game than in the past 10 years.

HWZ: One of the hot topics recently has been about the benchmark ethics issues. We've heard see-saw comments from both sides, but do you have anything personal to add for us?

Henri Richard: Intel went out and took SPECint2000, knowing there is already SPECint2006 and they did that only because it advantaged their platform vs. ours. There is a simple question to ask, if SPECint2006 had shown better results than SPECint2000 instead, which one would you have used?

I think on this one, they can't pull off a lie. It is just so blatantly obvious that instead of playing by the rules, they are trying to do anything they can to manipulate the public opinion. That just demonstrates that they're not a leadership company. It is a coward's approach and all we want is the truth. We lose some, and win some. We don't win every benchmark, not even in SPECint2006, but why deceive users by using a six-year old benchmark. I've read the answer of their spokesperson saying, "Well we just used whatever was available on AMD's website". That is a lame answer.

But let me get to the bottom on why this is so important. It is not so much between the battle between AMD and Intel. It is the fact that we both, both companies have the responsibility first and foremost to expand the market space and I believe that comes through the good education and information to the end user, not FUD. I really think that is the main issue.

If Intel focused on expanding a 35 billion dollar company to 70 billion and in the process of doing that AMD were to go from 10 to 20 billion, who cares. They are still No.1, they are still doing well. But instead being focused on expanding the market, they think their path to greatness is to take us and beat us 2-300 million bucks a quarter, which is ridiculous.

That's why I made the statements; that's why I say we have to stop and be responsible. This industry needs better rules. This industry has been around for 20 years and there isn't even a measurement as simple as miles-per-gallon for a car. That is wrong. If you think about the amount of money spent around the world by users buying IT technology and the lack of transparency on exactly what they're buying - it is a unique situation. There is no other industry where the misinformation is as bad as our business - that needs to be fixed.

Simply because Intel is playing PR, manipulating benchmarks, I don't think it's a right thing to do. I could do the same, but I'd rather just call their bluff.

Present Issues and the Future

HWZ: Recently, there have been rumors/predictions of cash flow problems stemming from the ATI buyout, specifically because of the huge loan taken out on it, adding in factors of your new proposed billion dollar plants, are these just pure speculation? Can you comment on this?

Henri Richard: Sure, but I can only say this and talk on my personal experience. When I joined AMD 5 years ago, our balance sheet was extremely weak and I remember very well talking to Robert Rivet, our CFO and he gave me a very clear plan on how he was going to fix the balance sheet and he executed every step of it. Today, when I go talk to him about our cash flow, he has a very good plan in place. And I have no reason to doubt that the same man that did a fantastic job 5 years ago can't do it again. There is a lot of speculation and rumors going on, but I'm not worried about our cash flow position.

HWZ: AMD has been operating on a very low profile in the desktop space over the past 6 months, quietly releasing new lines of processors, such as moving to 65nm, beefing up your energy efficient models and constant price cuts. How difficult has it been for AMD to maintain competitiveness and sustain growth?

Henri Richard: Well I think what's been really frustrating is that I've been going around the world and people think we've been too quiet. I take it very seriously because I think it's true, I agree. But, we are a company that's been always operating on the basis of saying the truth and not hyping things. I've been recently quoted because the competition has been really trying to use, what I consider to be unethical practices to hype their products. I'll give you two examples. First, everyone keeps quoting Intel's power envelop, but forget to include the Northbridge. If you're comparing a processor without a memory controller, which is in the chipset to our processor that has an integrated memory controller. That isn't a fair comparison, but everyone allows them to do that. We keep calling foul, we keep saying it is the responsibility of the analysts and the press to report accurate information. And comparing an Intel CPU without the NB to our processor isn't fair, but that's the way it is.

Recently, I've been reading a demonstration of their G965 chipset against a Radeon X1600 and it is clear that the demonstration was rigged. I think that this is not what we would expect from a market leader. It tells me that Intel is under a lot more pressure that they want to admit.

We're definitely going to change a little bit and probably be more vocal, more over the radar than under, but fundamentally I don't believe that two wrongs make a right and we're going to continue to sit this through.

For example, when we launched a new line of 65nm products; from an end user perspective, it isn't really interesting because an end user doesn't care if its 90nm, 65nm or 45nm. What they care about is what product will they get, will it do the job and what price are we talking about. So, on this very question, I'm somewhat divided. On one side, yes, we should make more noise on our products, but on the other side, ultimately it is important for our customers, our OEMs to position their products right, and we rely on them to do the right thing in the market place to serve the end user. I don't think we'll change that fundamentally, but I hear you.

HWZ: AMD is no longer considered an underdog in the microprocessor industry having lead through innovation over the past few years, but now that the gorilla has awaken and fought back, what lies ahead for AMD?

Henri Richard: We launched Opteron in April 2003, we're now in April 2007. That's four years. It's been a good fight, but it's now time for us to have a new architecture. Today, the Athlon 64 and Operon dual core processors are still providing very good value at any price point. You get a better processor. We used to have an enormous advantage and now we don't, but it took Intel four years to catch up. That's not bad for a smaller competitor. Now, we need to bring out our new generation of products, and that's Barcelona. It is coming this Summer. That's when we'll start to make a lot of noise. That's just the reality. My products are still very good, they're still very competitive. They're still providing the best value in the industry. But they don't have the newness of the competition, who just launched a whole new set of products, with a new name, etc. That's what you're seeing. They're on the attack and we're on defense, but as soon as we have our new products, that will change.

HWZ: So in your personal opinion, where do you see AMD heading to?

Henri Richard: Come summer, well we've got the Barcelona, R600, R610, 630 and we've got a lot of new products coming into the market. That will give us lot more competitive line-up. Customers that we have today are very different from what we had five years ago. The product portfolio we have today is different from five years ago, so people who are looking at the current situation and saying, Well this is 2002 all over again, are missing the point. Yes, there is fierce competition out there and yes, we woke up the gorilla, but the environment is very different.

We're now a worldwide player, with very good market share in almost every territory in the world. We have a good business relationship with every single large OEM in the world. We have a portfolio of technologies that goes beyond the microprocessor, we're into handsets, we're into digital TV and now graphics of course. That's not to say I don't expect two or three quarters of very tough battle, but I think that the way that AMD will re-surge will be very different than what happened a few years back.

Our vision is solid, our customers are very enthusiastic about what we can bring to the market. A lot of people understand that this industry needs balance and actually believe that AMD can bring this balance. I expect tough fighting, but I expect also that we will be able to demonstrate that within the next 12 months, we will be back on a very good competitive position. We will continue keep the gorilla on its toes.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.

Share this article