Intel Developer Forum, Fall 2001

Louis Burns
Frank Spindler
Ron Smith
San Jose, Calif., USA
August 29, 2001

LOUIS BURNS: Welcome to the third day of IDF. Hopefully, it's been worth your time, and travel and money so far. I'm going to try to make sure it is again today. What I want to spend my time on today is talking about how the environment is evolving and how that applies to clients. And when I say "clients," we mean desktop, mobile, and handhelds, be it handsets or PDAs, whatever that device is. At Intel, we think about clients in that fashion.

Let's get started. The Internet effect has been pretty well written about, but just want to put this up as a backdrop. The Internet effect has really taken people from working as individuals or thinking about themselves as individuals and really allowed them to come together as a community around the world, removing time and space as barriers to interact in ways they've never interacted before.

In business, from standalone businesses that did exchange with other companies in a very archaic way, now to fully integrated supply chains between suppliers, customers, and factories, literally around the world.

The Internet's enabled that and changed business forever in that aspect.

Computing, from the aspect of centralized computing in an enterprise, to truly distributed computing, computing both inside the corporation and outside that corporation, literally spread around the world, so distributed at an extreme level in ways that we hadn't thought about before.

And the last area is communications, how do we get this all to work. And the way that happens is to move from relatively proprietary communication standards to very open standard-based to enable all this.

Yesterday, Paul Otellini talked about a car analogy, referencing Henry Ford's comment, you can have any color you want as long as it's black, somewhat of a rigid description. In the computing environment, if we look at the way it was, it was a pretty traditional view of clients and servers and how they interacted with each other.

The way it is now is radically different. What it is about is clients, as in desktops, mobile, and different handheld devices working as equal members in this environment, this network. The ability in a peer-to-peer interchange between a handset and a desktop, the peer-to-peer interchange between a mobile wireless client connecting back to a server. The peer-to-peer interchange between all of those devices literally spread around the world.

This also talks to data, and how people want to work with it. They want to work with it in a personalized way. They want to interact in a way that fits their needs and their mode of working.

The bottom line on this, as we think about it, success in the future, successful solutions, be they hardware, be they software, are really going to have to facilitate the collaboration of all types of devices. The concept of only working within a feature set of two or three of these is not going to mean success. If I'm a business that's having my field people for some reason outfitted with a PDA in a wireless environment and they need access to the SAP database information to know the status of a customer order, that has to happen.

You can think of thousands and thousands of ways that people could interchange and exchange information spread around the globe.

So it's changed dramatically. At Intel, we've thought about this from the architectures that we're putting into the marketplace.

The other way of thinking about this is, how has it evolved over time. If you go back from the simple concept of Web sites, where we used to use the term "screen-scrapers" and "dead data," it's not really very active, it's really moving from Web sites to Web servers.

Jim Allchin of Microsoft talked about this yesterday, the concept of live data, interactive data, services that exist around the network, that devices will provide services and also use those services as equal members of that environment.

The technology's evolved tremendously to enable that. From simple TCP/IP to HTML, which is a screen-scraper concept, to XML, which brings applications and documents to life moving out over the horizon in the near term to UDDI, SOAP, and RosettaNet*, the very important standards to allow corporations to work with each other. Let's look at how we used to do a simple connect. And it was pretty good when you could just connect. Moving to a browser, which was some interesting breakthrough technology a few years ago. To how do we integrate and transact in that environment, to moving much more to an automated, collaborative environment.

Machine to machine making exchanges, collaborating with machines.

This is a fundamental shift facilitated by the technology and the standards in the industry. We've put a line on where we think it is today, which is kind of into this integrate and transact area with XML hitting the marketplace in a big way. In the very near term, we'll move to the right as Web services or as service-based architecture becomes a reality.

Now, you've seen Craig Barrett present this a number of times in different forums. Intel really has four architectures that we're driving into the marketplace, working closely with the industry, trying to enable these capabilities. You'll see later this week a great set of demonstrations and updates on the Intel(r) Itanium(tm) processor family and the overall server architectures that Intel is pushing into the marketplace. You'll see tomorrow from Sean Maloney what's going on in our communications environment where we're making great inroads from an architectural and a technology point of view.

But what I want to do today and what this morning is about is really focusing on the client.

We've got our IA-32 architecture, which we've been pushing in the marketplace for many, many years. And you're all very familiar with it. We've also got an architecture called PCA, which is a fundamental breakthrough in how to enable hand sets, be they cell phones, PDAs, or other devices, in an environment that is logical, that makes sense to drive into high volume, a standards-based environment.

So what you're going to see the rest of today is about the client, and the client at Intel is defined, as I said, from handhelds all the way through desktops, including mobile.

I'll spend the next 20 minutes talking about what's going on in the platform architecture, and the standards that are emerging there.

Frank Spindler will come out and talk about what's going on in mobility at Intel in the industry, and then Ron Smith will come up and talk about what's going on in handhelds and the PCA architecture. I think what you'll see by the end of this morning is a well thought-through, logical set of products and architectures and standards that are about enabling this Web services environment for people and companies to work in the way they want to work.

So let's talk a bit about the platforms and the standards that are evolving in those platforms.

You've seen over the last couple days, starting with Monday, and then moving through yesterday with Paul Otellini, is we think that CPU performance is a given. We don't take it for granted. We're working very hard on it. What we've been able to show you, I hope, in the first couple days is that we're in high-volume production of 2 GHz Pentium(r) 4 products. Paul demonstrated publicly a 3.5 GHz product yesterday, based on our 0.13-micron technology that will come out later this year. It's based on a fundamental architecture base of the Intel(r) NetBurst(tm) microarchitecture. We have a clear understanding, we believe, of how to extend this up to 10 gigahertz.

So we think we understand how to do this. I think we've got a good track record. We think the CPU side of this is done.

Moore's Law will continue. In the spring this year, Craig Barrett showed you some very interesting technology and what we're doing in our process technology.

Transistors that are a mere three atoms thick that deliver incredible speeds and increase the density. And Moore's Law is very healthy, very alive, and at the basis of our industry that provides silicon.

But the part that continues to need work is in the area of I/O. How do I provide to the marketplace, how do we provide to the marketplace an I/O capability that keeps up and delivers the CPU performance in a fashion to solve customers' problems?

That's where I want to spend our time today in my section. We're going to talk about three important serial interconnects. The first is I want to give you an update on 3GIO. I stood up here in the spring and made some pretty bold commitments to you, and I want to give you a report card of where we're at.

The second is on USB 2.0. Tremendous progress in this area, and I want to give you an update of where we're at and where the industry is at on USB 2.0, and the last is Serial ATA.

The bottom line on these capabilities is about delivering innovative platforms into the marketplace. These will change and enable architectural breakthroughs and architectural partitioning that we've not seen before.

So it's about, in these platforms, enabling that innovation.

Third-generation I/O. You may or may not remember this foil from last year and what we really tried to point out from last spring. What we tried to point out is the need. If you go back and look at our industry both from a communications and a compute industry, you'll see it's been based on a standards environment.

If you go back in the PC environment, you'll see in the '80s it was ISA and ISA served us well. It was an industry standard where thousands of people could innovate around it.

If you look in the communication area there was multibus. Those standards served us very well, but after about a decade in most cases, they run out of steam and need a new generation. PCI became that. If you remember, PCI was originally put in place to improve graphics performance. That evolved over time and there were extensions to that architecture to enable new products and new capabilities.

As we looked at this last spring, it became apparent that we needed a new evolutionary I/O standard. Some of the characteristics that were also clear to us, working closely with our customers, is it had to have some important characteristics.

We saw that the limit of running high-speed parallel buses was starting to run into a problem. We saw, as you looked through the next decade, the speeds at which we need to operate this I/O capability at. When you looked at that, the following characteristics or requirements fell out.

It has to be full serial. And I do mean full serial, with integrated clocking. The ability to scale at frequencies and performance that are really limited in any other method.

It has to enable point-to-point in a very high-speed capability. It has to scale to greater than 10 GHz. It has to support multiple market segments. It can't just be a PC interconnect. It has to provide great flexibility and compatibility with some of the existing standards.

So what I'd like to do now is give you a report card. In March of this year, I committed to you on this stage that we would work with the industry to develop this next spec, this next interconnect. Quite frankly, after we read a number of the articles, I know a lot of you didn't believe me. We were very serious about it and here is where we're at.

The initiative structure is in place and working very well. We have tremendous support for this in the industry, and I'm pleased to announce again -- this came up a couple of weeks ago -- is the PCI SIG has adopted this and will drive it as the next evolution of the PCI environment.

Defining the third-generation I/O architecture is well underway. If you step back for just a second, if you look at the health of our industry, one of the things that we've done as an industry for years is the ability for competitors to step back, establish standards for the good of the platform, and then reengage in a competitive environment. 3GIO is an outstanding example of that in very intense competitive times.

We said we'd have a spec ready for you in the fall IDF. The preliminary spec is completed. The spec is under detailed review by the 3GIO Working Group. The PCI SIG full membership will review it in Q1 and the full spec will be available to the industry at large by mid '02. So the work is underway with the working groups, we're getting great input as an industry, and the bottom line on this one is the industry now is driving 3GIO, which is an important standard for the industry to adopt.

Tremendous progress in the first six months of this, and again, a sign of health of our industry.

This is a list of the companies involved. It comes from a wide variety of industries and a wide variety of different products in the marketplace. 27 companies, very active, very engaged in the definition and all the details of the next-generation I/O spec.

What I'd like to do now is run a quick animation to give you a little bit more detail about what 3GIO is about. So if we could start the animation, please.

(Video playing):

On its physical layer, third-generation I/O architecture supports point-to-point digital signaling over a transmit-receive pair. Data on these uses an embedded clock scale. The protocol layer features a packet IA structure that includes a dedicated message space replacing the many side band signals used in today's I/O buses.

This has the benefit of reducing the pin count between devices enabling a high bandwidth per pin for extremely cost-effective scalability.

Third-generation I/O architecture has other important benefits, it supports scalable bus widths, has low power requirements, and has a flexible I/O implementation, including a high-speed in-system chip-to-chip interconnect, a peripheral interconnect for compact form factor modular add-in cartridges, and an internal I/O attach for external I/O technologies, including IEEE, USB 2.0 I/O, InfiniBand*, and Ethernet.

In board layout, this means fewer traces, offering wider capability in the way components can be placed on the board. It also enables a new level of modulator and system design. By mid decade, innovative designs may feature an I/O compute core that can scale across many product designs. Third-generation I/O architecture supports greater ease of use, providing a new fully serial, plug and play hot swappable I/O interconnect for systems used in the home, in the office, and in the datacenter.

Whatever the usage model, third-generation I/O architecture enables easier upgrades with cross-platform interoperability. Of special importance to the industry, the transition requires no software changes. Operating systems that currently support PCI will also support third-generation I/O, as it supports the PCI software driver model and the PCI standard for enumeration and configuration.

With third-generation I/O, the industry can use a single set of design rules to flexibly implement a highly scalable, general-purpose I/O interconnect with the cost effectiveness required for desktop PCs, the power efficiency required for mobile systems, the performance needed of server applications, and the features required in future generations of communications devices.

LOUIS BURNS: I want to reemphasize the point once again. It's an outstanding example of this industry ability is to step back from day-to-day competition, make the right decisions to establish the platform for the next decade, and then turn around and go back to competing. A great sign of health.

There is a really comprehensive interconnect strategy underway. But if you start on your left, we like to think about this in terms of site to site or building to building, moving from box to box, to blade to blade with line cards, and then moving from chip to chip or chip to add-in cards. If you think about that as four basic categories, what are the interconnect strategies or capabilities that drive that.

If you go back historically, what you'll see is Ethernet has been around a long time, was a standard established by Intel, Xerox, and Digital Equipment Corporation, and it was really targeted at box to box in the beginning.

What you're seeing happen, though, is it's starting to move in a number of directions. It's starting to move and look at the metropolitan network area capability. It's starting to move into the storage area and really starting to move into some low end communications back planes.

So that evolution is underway. The next one is InfiniBand, incredibly important capability established three years ago in the marketplace. It was really focused on high performance server clusters and back planes. But that's evolving and moving into a data center in a much bigger way. Also moving into the storage area and working into a robust communications back plane.

The last one is 3GIO, which is just in the process of being established in the marketplace. Really focused in the beginning on chip to chip and chip to board. As we look forward, we'll see that evolve into things like communication back planes, but we see this as a tremendous complementer to the InfiniBand architecture.

So I'm going to move now into USB 2.0. The press on this has been tremendous, and this is another example of the industry at large working together to advance the platforms.

We introduced the spec, it's probably about a year-plus old. Products started to ship earlier this year in the USB 2.0 area, the support for this in the industry is excellent, and the press is picking up on it. Another congratulations to the industry at large to be able to drive this capability.

I want to remind you what the capability and importance of USB 2.0 is all about. It delivers 40x performance over USB 1.1.

It's compatible with USB 1.1 devices, which I think is critical. 300-plus companies are active in this spec and driving this capability. And you're starting to see a large number of USB 2.0 peripherals show up in the marketplace across a wide variety of capabilities.

So what I'd like to do now is bring Bill up on stage here and give you a quick demonstration and contrast USB 1.1 and USB 2.0 and also a pretty important product announcement. How are you doing?

BILL: Good, good. So for today's demonstration, I'm going to be using a Pentium 4 platform. Connected to that platform, I have two identical Fujitsu hard drives connected to the system. Both are USB 2.0 capable.

I have them set up -- one of them as a USB 1.1 device, the other as a USB 2.0 device. I'm running the same AVI file on both drives and, as you can see on the screen, I have them displayed in separate windows.

Taking a look at the USB 1.1 file, you can see the frames are locking up, we're dropping frames and the bandwidth is not sufficient to display the AVI.

When we look at the USB 2.0 device, you can see that no frames are dropping, we've got sufficient bandwidth to get the data to memory for display.

LOUIS BURNS: So it's a great example of the power and performance of a Pentium 4 platform in getting an I/O capability that really helps deliver on that performance in this case.

BILL: Exactly right. You can use the same example if you were doing file transfers. The USB bus would not be a bottleneck.

LOUIS BURNS: That's been interesting to show side by side. But I know there are other pretty cool things happening. And I notice you have this big ugly green screen here. What do you have here?

BILL: We have a USB 2.0 camera from Opteon Corporation connected to the platform. As you can see, there's sufficient bandwidth to display the realtime video. And it's all uncompressed.

LOUIS BURNS: So what I've noticed here is a couple things, it's great quality, and I've also noticed a desire on your part to appear on David Letterman Live, because it looks like a New York background. So I notice you've got a motherboard up here. What's that all about?

BILL: Before we go to the motherboard, I'd like to say, on this display, what we have is made possible by the USB 2.0 bandwidth to provide the realtime video uncompressed to the system. We're also using the performance of the Pentium 4 processor. And also Serious Magic's Real Effects visual -- Real Effects software engine to display this virtual environment I appear to be in.

LOUIS BURNS: Okay.

BILL: Now, on to this motherboard that you are so antsy to see. This is the heart of our technology, a Pentium 4 motherboard. The unique part of this motherboard is we have a core logic chipset that has an integrated USB 2.0 controller in it.

LOUIS BURNS: What you're seeing here is the first time a public demonstration of integrated USB 2.0 in an Intel chipset?

BILL: Yes.

LOUIS BURNS: When will that be available? BILL: We will begin sampling in Q4 of this year. And we plan to go into production first half of 2002.

LOUIS BURNS: Great. Thanks for coming out.

BILL: Thank you.

(Applause.)

LOUIS BURNS: So what you've seen in a pretty straightforward example is the performance improvement delivered by USB 2.0 to the marketplace in a truly compatible fashion with USB 1.1. This is important as we move these platforms forward to keep the balanced I/O with the performance of the processor. What you've also seen for the first time is a public demonstration of USB 2.0 integrated into an Intel chipset.

What that means is, in the marketplace, we'll start making USB 2.0 ubiquitous as we deliver millions and millions of chipsets next year with USB 2.0 integrated in it, fully compatible with 1.1, and the total performance of the 2.0 as those peripherals emerge. So it's integrated, ready for broad deployment into the marketplace and another example of moving the platform forward with an industry standard driven by the industry to improve overall user experience.

Want to move to Serial ATA. Serial ATA is a very important standard that's moving forward, because it also addresses performance, cost, and flexibility of that box. If you think about it, it's really scalable again in a serial fashion for over the next decade. It helps in a very big way as people start defining their repartitioning boxes. One of the more profound visual impacts is in cabling. I know all of us have spent our weekends opening boxes up and moving these cables around as we've tried to add in boards or new disk drives. (visual demo) This is a flexible flat ribbon cable. You can bend it in a lot of different ways. If you don't get it quite right, the cover won't go back on. But this has delivered some relatively good performance to the beginning of our industry. What we're moving to is a very simple Serial ATA cable that will allow people to ramp in the performance but also simplify the inside of that box. And you won't have to do all those interesting folds and bends as you put the product together. It's 100 percent software cable compatible, with the existing ATA again playing close attention to the needs of the marketplace in an evolutionary way, not to leave the past behind in the sense that you have that evolution. It provides hooks for hot plug. And the cost structure is very similar to parallel ATA.

Now, what we do see, though, is the simple fact that because of the way these cables are done, it will actually hopefully improve manufacturing costs as people are trying to put those systems together.

70-plus companies have been involved in this Serial ATA working group. They have made tremendous progress. If you go onto the showroom floor, you'll see a large number of demonstrations of this capability running in multiple operating systems. Industry development's well underway. We expect first products in the first half of next year.

Also, today, the working group is releasing 1.0 spec of Serial ATA. It is available at the Web site below (www.serialata.org). Again, another example of the industry working together to establish an important spec to move the platform forward. Now, if you step back for a second, look at these three important initiatives, which you'll see both in desktops and in mobile computers, and you start looking at what will enable them into the marketplace, it's really about some innovative form factors. Now, the area of small form factors has been an interesting debate in the industry. I wanted to give you a quick snapshot of what Intel has been doing in small form factors. We've been active with a lot of research, R&D, and working closely with a number of partners for three-plus years in the area of small form factors. This includes research and enablement from simple things like chassis, power supplies, cooling, motherboard layouts, trying to understand how we put the best capability into the marketplace.

Some of those experiments have ended in the lab. A number of those experiments have gone into production with a number of our partner companies.

So there is work well under way. There is joint work well under way with the industry. And what I'm going to do now is bring out the most recent work that we've been doing with one of our important partners from China. What you'll see here is a system that we've been working on with Legend, China's largest computer manufacturer. It really does some very interesting things.

You'll see it's sub 6 liter in size. Kind of got cool packaging to it. It's carrying a full 2 GHz Pentium processor. The processor is capable of handling up to 60 watts in a sub 6 liter box.

The package has also taken advantage of some of the new capabilities in the I/O space and has actually repartitioned the box using USB 2.0 because of its performance to enable that.

The thing runs in the 30 dB range from a noise point of view, so what you get is a no-compromise capability. High performance, 2 GHz, repartitioned used and taking advantage of the industry standards, very, very quiet, no compromise capability. Another example of our work with the industry in the small form factor workplace with no compromise.

I want to close my section with a foil from Paul Otellini's talk yesterday. We as an industry have a responsibility and an opportunity, and if you go back to Charles Kettering's quote back in 1933, I think it somewhat comes down to this: It's business will come back when we put some products in the marketplace that people want.

The industry's working hard to deliver new I/O capability and standards to advance the platform. The industry has shown the ability to step back from its competitive place, establish those standards, and reengage in the competitive marketplace. A sign of health of our industry to improve the platform for the users.

Working closely with you as an industry to establish these standards and proliferate them around the world, we have a great chance of delivering on what Mr. Kettering asked of us.

Thank you very much for your time this morning.

(Applause.)

LOUIS BURNS: What I'd like to do now is introduce Frank Spindler who is a vice president and general manager of our mobile platforms group. Frank.

(Applause.)

FRANK SPINDLER: Thank you, Louis, and good morning. I want to talk today about things that we are doing in working with the industry to architect mobile PCs for the future. I think an important thing to look at is -- or a good way to start is, how have different value vectors in mobile evolved over the years.

If you go to the very first somewhat mobile PCs that were really luggable systems that ran on AC power, people just liked the fact that they had a computer they could take with them. So the performance had some degree of importance. Form factor was just enough to be able to carry around, lug around, but it was already conscious to people how heavy things were and there was a desire to move to lighter systems.

In the late '80s, we got to the first battery-powered systems. They were more transportable. They were getting better in performance. They had a smaller size. They were getting into eight to ten pounds and above type of sizes, but now that they ran on battery, people came to be aware of battery life and the importance of battery life.

As we got into the mid-1990s, the first Pentium( processor-based notebooks, they continued to get higher and higher in performance, they got smaller in size, more in the seven-eight pound category, battery life was getting better, but then the connectivity and the ability to put a modem, typically through an add-in card, or soon thereafter, a local-area network connection also typically through an add-in card became somewhat important.

Then as we look at the environment today, all of these factors are continuing to increase in importance. You have performance, people liking smaller systems down to sub 3 pound type systems, battery life getting better, and now the emergence of a new important value vector, wireless communications or wireless connectivity.

So as we look to the mobile PC platform and how to best develop solutions for that mobile PC platform, we focus on these four areas. Performance is important. People want in a notebook every bit as much responsiveness, every bit the performance as they have in a desktop system. I don't want to compromise performance solution in a mobile PC.

But other things are important, like battery life. We've had great improvement. We've seen systems over the last several years evolve from typically two hours of battery life to four hours, but people want to have systems that can go all day long and use them without worrying about finding a power source sometime during the day.

Seamless wireless connectivity. Things that are simple, easy to use, not worried about what the underlying technology is, the ability to connect anywhere, anytime is increasingly important.

And the importance also of smaller, lighter, easier to carry systems. The more I use these, the lighter the system is, the happier I am with that particular system.

So we are focusing our strategies both at a microprocessor/chipset, development and platform development areas on these four mobile value vectors.

Now, I'd like to start first by talking about wireless, because we believe that wireless communications on mobile PCs is going to have a profound effect on the way people use their systems and the way these systems are going to evolve over the years. And it's going to incorporate multiple wireless communications technologies.

One of the things that we're seeing earliest is 802.11, wireless local-area networking. Wireless LANs -- working groups were established in the late 1990s. The first rev. of the 802.11b specification was published in 1997. We saw the first emergence in some degree of volume of wireless LANs typically through add-in cards on mobile PCs in 2000.

This year we're seeing a dramatically higher adoption, not only through add-in cards but being built into the system. In fact, several of the systems that we honored as Innovative PC Awards for mobile earlier this week had integrated local-area network connections built in.

At our recent Intel(r) Mobile Pentium(r) III Processor-M introduction, every single system that was highlighted there was configured with wireless local-area networking connectivity. So we're seeing major growth in this capability on systems today.

We're also seeing the build-out of infrastructure to support wireless local-area networking. Businesses are beginning to deploy it. For instance, Intel is installing wireless local-area networking in several of our buildings. College campuses or schools are using this capability. We're seeing more people build airport infrastructure. For instance, the San Jose Airport and the San Francisco Airport both have wireless local-area networking installations. Starbucks will have over 100 locations just in the Bay Area with wireless LAN built in, and we'll see this extend to hotels. It will become prevalent in the home as well. The infrastructure is happening.

The next technology is a complementary type of technology that is more of a personal connective type of a technology, which is Bluetooth*. For Bluetooth, the first specification was published in 1999. We're seeing a number of products emerge in the second half of this year with Bluetooth capability.

For instance, Hewlett-Packard has announced a printer that supports Bluetooth. Multiple phone vendors, like Nokia and Ericsson have announced phones with Bluetooth connectivity. Sony recently announced digital camcorders that support US -- I'm sorry, support Bluetooth connectivity.

And so the way to look at Bluetooth is analogous to a serial connection. So something like a wireless USB type of connection for a mobile PC.

As we go through time, we'll see the integration of wide-area networking capability. This is difficult to tell exactly when and how this will happen, but the fact that high bandwidth wide-area networks are being deployed, things like GPRS or, further on, 3G, that's going to give communication bandwidths at hundreds of kilobits or even megabit per second levels. It's going to be a very attractive connection technology for mobile PCs.

It's important for us as a development community to realize that all of these technologies are going to be adopted in mobile PCs and to develop technologies that allow for seamless interaction between different technologies, and even seamless interaction between different types of devices, like mobile PCs, handheld, and phone types of devices, all of which will have the ability to communicate with one another.

Now, what this wireless technology is going to allow users to do is something that they wanted to do for a long time, which is take it off. So if I could get a little mood music here, I would like to do what users would like to do. They want to take it off. They want to take off the LAN cable on their system. They want to take off the serial cables on their system. They want to take off printer ports and other types of connections. And while they're at it, why not take off that other connection, the power cord. So relax. That's as far as it goes.

(Laughter.)

(Applause.)

FRANK SPINDLER: But what this will do, then, is take mobile PCs out of a semi-mobile environment, because with the Internet, communications, connection, access to information is one of the absolute primary drivers of the value of mobile PCs.

But today we select different fixed locations to get that connectivity. We plug in at the office. And because we have to find a fixed location and plug in a cable for our connectivity, it's not that difficult for us to plug in other things, like AC adapters.

But with wireless communications, then, it's going to drive to an incredible extent the other value vectors for mobile PCs.

So for example, now I don't have wires. I can carry this around more. Well, battery life becomes more important. I don't want to have to access and plug in. I want this all day. Size and weight becomes more important, because I'm going to take it with me, I don't want to have to feel like I lug it around. And almost most significantly, performance becomes more important.

And so what this will drive, then, is a cycle, a cycle of technology, wireless communications driving new users and uses for mobile PCs which increases their expectations on other things like the battery life, the size and weight, and the performance that that particular system is going to provide to them.

Why performance, you may ask. Well, the more that we ask of the system in terms of wireless communications, the more that the processor has to do in being able to handle wireless communication streams, to be able to handle background security functions, to be able to handle the background compression and decompression of images that are transmitted wirelessly.

So as one example of a major step we're taking to extend performance, first we have just introduced the Intel Mobile Pentium III Processor-M. It's the highest performance mobile processor on the planet today, runs at 1.13 GHz, had several performance enhancements that allow it to significantly increase its performance over the existing Pentium III processors. We're going to extend that family with higher speed versions, we're going to offer low volt and ultra low voltage for smaller form factor versions later this fall, and those systems are ramping in volume production today.

The next step for us will be to bring the Pentium 4 processor to mobile. Now, the mobile Pentium 4 will incorporate all of the architectural features built into the desktop Pentium 4 systems. It will support rapid execution, improve cache subsystems, hyperpipeline technology, and other system features like fast system buses, and the new instructions, the streaming SIMD extensions, too.

But bringing a processor to mobile is much more than just bringing the core architecture and performance features. It's also incredibly important to support the mobility. And we will build into Pentium 4 the full set of mobile features that we are offering in the Intel Mobile Pentium III Processor-M today. So for example, we'll take advantage of .13 micron process which allows the technology to run faster but also consume less power, also specialized packaging so it can fit into thin and light systems, we will support enhanced Intel(r) SpeedStep(tm) technology that allows the dynamic adjustment of voltage and frequency and automatic adjustment between a maximum performance and battery optimized mode dynamically on the system.

And we'll take advantage of low power states such as deeper sleep and other what we call alert states that move the processor to an extremely low power consumption level even in the midst of running applications.

So what I would like to do now is showcase the first-ever public demonstration of a mobile Pentium 4 processor. I'd like to ask Jeff to come out and help me. Good morning, Jeff.

JEFF: Thanks, Frank.

FRANK SPINDLER: What do we have here?

JEFF: What I have here is a customer reference board running a prototype of the Intel Pentium 4 mobile processor. As you can see, it's alive and kicking. What I'm going to show you today is a demonstration of the power that you can see on this upcoming platform.

FRANK SPINDLER: Okay. Well, let's take a look.

JEFF: I have this application called Magix Audio and Video Office*. And I'm using this to show you how the Pentium 4 mobile processor, with the Intel(r) NetBurstTM microarchitecture, enhances your ability to create media-intensive content anytime, anywhere.

So in the interest of time, I've preloaded the application with a project I've put together from this library of media clips below. And as I play the project, I'm going to show you how easy it is for me to make realtime edits to the project and apply realtime effects.

FRANK SPINDLER: Okay. So you've basically set up a bunch of audio tracks. Now you can set up video tracks and other types of effects, so create a combined audio-video type of experience?

JEFF: Absolutely right.

FRANK SPINDLER: Okay. Well, that looks pretty good. Now, I think we want to get some kids to want to buy the Pentium 4 mobile systems, too, or make them attractive. Can you give me a little more of an MTV experience here?

JEFF: I think I have something just for that. I can apply some chromakey effects here to show some really rich graphics here. And there we go.

FRANK SPINDLER: Okay.

JEFF: And once we're done with the project, the final step is to encode it into a media file, into a video file. And I'm going to go ahead and use the MPEG-4 codec. And this is where you see the power of the Pentium 4 mobile processor. And as you can see, it's encoding at a really fast speed.

FRANK SPINDLER: So, basically, you've then done that complete encoding in a matter of seconds, which was dramatically faster than would you see on a Pentium III. Now you get a feel for the final product. And how long did it take you to put this together?

JEFF: Just a couple of minutes.

FRANK SPINDLER: A couple minutes on the fly in mobile. So that looks great. We introduced a Pentium 4 2 GHz on the desktop earlier this week. And I know we're at 1.13 today on the Intel Mobile Pentium III Processor-M . How fast is this thing running?

JEFF: Okay. If you take a look, we're running at just about 2 GHz.

FRANK SPINDLER: 2 GHz. Okay. Well, that looks great. Thanks, Jeff.

JEFF: Thanks a lot.

(Applause.)

FRANK SPINDLER: Now, I assure you the system is actually running at 2 GHz. There's just some rounding errors inside of the display system there.

(Laughter.)

FRANK SPINDLER: What you're going to see, then, is this type of a capability that's truly mobile. I could just imagine a disc jockey at a dance just taking a mobile PC with a Pentium 4 inside it and being able to develop complete unique multimedia-style presentations, bring a full library of MP3 audio clips to use, and really create a full experience.

Just some examples, I think, of how this type of performance in a notebook is really going to extend the uses of notebook systems.

We have been incredibly pleased with the Pentium 4 mobile processor and its status. We are exceeding the performance expectations we had for the processor originally. In fact, when we introduce the product in the first half of next year, we will exceed 1.5 GHz in performance. I showed you 2 GHz as a demonstration here. We will reach 2 GHz in production-level Pentium 4s in notebooks by the end of next year. We're sampling the processor in volume today.

We're also seeing very, very positive power characteristics and thermal characteristics for this processor. In fact, you will see the Pentium 4 in mainstream two-spindle, five- to six-pound type of notebook systems at introduction next year.

Our 0.13-micron process is ramping like a hose. We'll be able to build these in very, very high volumes. And we'll be able to support Pentium 4 at a broad variety of price points at its introduction. So it will move very rapidly, very quickly into the mainstream of the notebook market segment in the first half of next year. So we're very excited about the prospects for Pentium 4 in mobile.

Let me move to the next category of extending battery life. We've spent a lot of time talking at our processor introductions and at IDFs in the past about how we have aggressively managed our microprocessors, that we're using a whole range of power management techniques, we drive processors into low-power states. And as a result, we get processors which average as low as one-half watt of average power. For instance, an ultralow voltage Intel Mobile Pentium III Processor-M will average that kind of power characteristics.

So the CPU has become a relatively small percentage of the overall power consumed inside of a laptop.

Here we did an analysis of a typical thin and light mobile PC based on a Pentium III that was available earlier this year. And you get a feeling for the fact that there are many areas of the platform, for instance, the LCD panel, hard drive, that are significant power consumers.

In the chipset and graphics area, we'll see significant reductions in these numbers as we see the 830 chipset deployed on the Intel Mobile Pentium III Processor-M -based system. The 830 will provide both chipset level functionality and integrated graphics functionality. And so as that evolves into new platforms, we should see the chipset and graphics portion of this power consumption decrease significantly.

What I would like to do now is just give you some idea of how the Intel Mobile Pentium III Processor-M performs on new types of applications. And I think it's very important as an industry that we not look and concentrate only on old models, old usage models of productivity inside of notebooks, but that we look at the way people are going to be using systems in the future. And, again, this is one where wireless is going to have a very profound effect.

At Intel, we are actively characterizing both processor solutions and wireless solutions for power efficiency, and, in fact, wanted to give you an idea of how the new Intel Mobile Pentium III Processor-M -based system performs on these types of applications.

So, Jeff, what do we have here now?

JEFF: Okay. We have a Mobile Intel Pentium III Processor-M notebook here, and a Pentium III notebook here, both running in battery-optimized mode. And both systems are streaming video to each other at a bit rate of 300 kilobits per second over an 802.11b connection. And both systems here are instrumented so that we can take realtime power measurements on this host system here.

FRANK SPINDLER: Okay. So, actually, what you're putting up on the screen now is something we call a power analyzer. The -- it's showing realtime the power that's consumed by the microprocessor in each one of these systems.

The red indicates the Pentium III processor. The blue indicates the Intel Mobile Pentium III Processor-M. They're both doing the same sort of things.

The one thing I would highlight is how this illustrates the extremely active and aggressive dynamic power management of all of the Intel mobile processors. The fact that we are in the midst of this pretty performance-intensive application, we are streaming video out, we are receiving video in, wirelessly, we're displaying it realtime on the system. But even in the midst of that, you can see that the processor is moving into very, very low-power states of a fraction of a watt. Then we can move up to an active level and back and so forth.

But what I think this really highlights, you can see, is how the Intel Mobile Pentium III Processor-M takes advantage of the new architectural features of that in a wireless communications environment to reduce the power. In this case, the average power is reduced by about 40 to 50 percent over the Pentium III processor. Thank you, Jeff.

JEFF: Thank you, Frank.

(Applause.)

FRANK SPINDLER: And so we want to architect not only our processors and our systems to be able to take advantage and be cognizant of the new environments, the wireless communications capability that's going to be built in. We've done that with the Intel Mobile Pentium III Processor-M. We're going to continue that with our future products.

But as the previous chart showed, it's not just the CPU and the chipset. There are incredible opportunities in other parts of the system to reduce power consumption. And, in fact, the focus of the mobile tracks or one of the major focuses of our mobile tracks here at IDF is how to take power out of other parts of the subsystem.

We see, for example, that DVDs can take advantage of DMA to reduce power consumption. In the LCD panel, the single biggest power consumer inside of a notebook system today, the power voltage conversion efficiencies can be improved, especially in the inverter circuitry of an LCD, we're looking at technologies, for instance, that say can I dynamically adjust the intensity level of the screen based on the ambient light inside of the room so that as it gets darker, I can reduce the screen intensity and hence save power.

Things like battery. Battery technology is probably going to improve gradually, but there's technologies inside the platform, such as supercapacitors that can help reduce the voltage transients that you see on a power supply.

Did you know, for instance, that in many cases, you may have a healthy amount of power left in your battery, but because of a voltage transient that occurs inside the system, the voltage drops on the battery because of some sort of a current surge and your sensory circuitry tells you you're out of battery. Technologies like more accurate sensing or supercapacitors can help remove that inhibitor.

Driver softwares are a very, very great opportunity for all of us. We've found as we've done analyses on notebook systems, for instance, that many different types of drivers for other types of devices -- in certain cases, wireless LAN cards -- aren't taking advantage of the low power states available inside of a system or inside of the microprocessor.

So there's a whole laundry list here that I won't explain, but I invite you, please go to the track and get a feeling for some of the very specific actions we feel can be taken at the platform level to reduce power consumption. We believe with these types of techniques we can see a 30 percent reduction in the platform power of these other components inside of a notebook and help to extend battery life accordingly.

Another area, of course, is small innovative form factors. We want to have the ability to carry something that's as small as I want it to be and usable.

We are committed, and we remain committed, to segmenting our mobile product family in a way that will allow us to deliver not only products optimized for full size and thin-and-light notebooks, but products and road maps that are optimized for the smaller systems, the three to four pound mini notebook systems, the sub three pound systems, and any new products or technologies that may emerge from the innovation that you're bringing to mobile PCs.

We're introducing our low voltage and ultra low voltage Intel Mobile Pentium III Processor-M processor this year. We'll see that technology then move into these types of categories of notebooks.

We'll continue to have mobile optimized packaging, very small, light, surface mountable BGA types of components, that means you can get it into systems that go well below an inch in thickness.

And the 830 graphics product we announced with the Intel Mobile Pentium III Processor-M is going to be a big boon to these types of systems, because the integration of the graphics and the chipset functionality will provide for smaller space, and it will also lower the power consumption we're seeing. This is another case where we're extremely pleased and our product has actually exceeded our expectations in terms of the low power levels that are being provided on the 830 integrated graphics product.

And as I think it was demonstrated by Jim Allchin yesterday, you can see we have a prototype based on Intel ultra low voltage mobile processor technology of a tablet PC that is being championed in conjunction with Microsoft. And we'll see a healthy array of Intel Mobile Pentium III Processor-M -based systems that are tablet form factor as those systems emerge next year.

And now a look ahead. Paul yesterday talked about a longer-term direction in mobile, which is to develop an architecture that's specifically targeted for mobile.

In case you were wondering, the name Banias comes from a river in Israel, and there's an archaeological site on the river, and it was actually originally a place where it was honoring the mythological God Pan, and so Pan kind of evolved to Ban and Banias became the name of that river or particular site. It is being developed by our team in Israel.

They're working full force on the Banias product, which is targeted for the first half of '03. We're already staffing people to look at the successor to Banias.

So this mobile optimized architecture is absolutely our long-term direction, and we'll see many implementations and many successors in this particular family.

It's going to be utilizing different techniques, as Mooly Eden, our design manager, talked about yesterday. Micro ops fusion, which is essentially taking two micro operations that would normally occur inside of a microprocessor, fusing it into one, and hence allowing it to execute quickly but also with less power.

We're using special sizing techniques. Every single circuit designer on the Banias project has targets and budgets not only for a performance capability but for a power level inside their portion of the chip design.

And we will optimize down to the transistor level for power. We'll actually do special sizing on transistors that will allow them to consume less power. And any one individual block or function will look at what is the performance I need to require, provide that, but then size transistors to optimize the power reduction while delivering that particular performance capability.

And then also, we're going to use aggressive clock gating, which is essentially shutting off the clocks to different elements of the microprocessor. So, for instance, a portion of the cache that may not be used. We will only power -- or only clock, the parts of the processor that are actively used to execute a particular instruction. And this will all happen dynamically.

So we're very excited about Banias, and we are very much architecting it around these four vectors of high performance, long battery life, seamless wireless connectivity, and to support small innovative form factors. We are looking at each and every one of these as we architect Banias for the future.

So my call to action, my request to the development community. Optimize along all four of the value vectors. Don't deliver one at the expense of the other. Users want all four of these value vectors.

Develop solution stacks that support seamless wireless connectivity. Make the users not care about what type of underlying connection mechanism they're using. We've got some really exciting things that we're showing in our wireless mobile showcase over at the convention center. I invite you to come look at it. Technologies that, for instance, move from a wide-area network phone that's connected to the notebook by Bluetooth, and then when it moves into the local-area networking environment, it immediately senses that and transfers control over to the local-area network connection. It's completely invisible to the user. All they see is that they have a connection still. They don't care about the type of connection, but they're getting the best available connection when it becomes available to them.

Get ready for the mobile Pentium 4. We're excited about this. We think it's going to have a very strong ramp next year in mobile; have those systems ready.

Build power efficiency into all components of the platform. And develop for the future usage models. Think in terms of you're going to have notebooks that are very small, light. You're going to have to support three to four pound type systems, but you want wireless communications capability, like local-area networking or Bluetooth personal connectivity inside of those systems. Think about those types of usage models that we're going to see in those systems.

In summary, we are very, very pleased with the technology that we are providing, the leadership technology we're providing today with the Intel Mobile Pentium III Processor-M; tomorrow, with the Pentium 4 processor; and in the future, with Banias and its successors. We're excited about the mobile space, we're focused on these four value vectors, and we're going to be relentless in our execution in bringing these new technologies into mobile and in working with you to help bring this ultimate best-mobile-computing experience to users. Thank you very much.

(Applause.)

FRANK SPINDLER: I'd now like to introduce Ron Smith, who runs our wireless communications and computing group, and he's going to talk about some exciting new developments in even smaller categories of mobile devices. Ron.

RON SMITH: Thanks, Frank.

(Applause.)

RON SMITH: Thank you. Good morning. Now, we've all heard about the wireless Internet. This is the next big thing that you hear people talking about; a lot of hype, some substance. But let's really see how we're doing and what it's really going to take to bring the Internet to the wireless world.

First of all, let's look at the first instantiation of the Internet on things like cell phones. The wireless application protocol has been the initial utilization of this kind of technology, and we can see that there's actually been some 12,000 WAP sites that have been developed. We probably all heard about the success of the iMode services in Japan, which has grown at a phenomenal clip, and if we really look at the content and the number of applications, there's some 42,000 to 43,000 iMode sites that have been developed. Not a bad start. That's a lot of content. That's a lot of applications.

But let's really put this in perspective. Let's look at it on a scale compared to the total number of Internet sites designed around the PC and the Internet as we know it.

There's over 32 million unique Internet sites. So contemplating this for a minute, it should be obvious what has to happen. We really need to bring the wireless world to the Internet to take advantage of all that development and all that content that's occurring because there really is only one Internet. There's no uniquely wireless Internet anymore than there is a cable Internet or a DSL Internet. The wireless is the medium. The Internet is that ubiquitous entity we're all familiar with and we all know and love.

So what we really want to do is have a relatively seamless interaction between the Internet and this new class of devices, handhelds and phones, in the same way we already have that kind of connection from the desktop PC and the mobile PC.

So the need is that handheld clients and the applications that go on these handheld clients basically keep up with this robust pace at which new content has been applied to the Internet; otherwise, the experience, the Internet experience the end user is going to see, is going to be something significantly less than what they expect from an Internet access.

Now, in order to do that, Intel announced last year something we call the Intel(r) Personal Internet Client Architecture, or Intel PCA, and what this does is create a new development paradigm for these kinds of devices to enable a much faster development of new end user applications and content.

The key ingredients and the key methods in which this is done is there's a clean separation between the communications engine and the computing engine and the stacks that go on top of those. This is done at an abstract level, so independent of the level of integration of the actual hardware right down into the chip level, the developer still sees a separate communication stack and a separate computing stack.

This is important because the communications technology has to go through a network-type approval, and we don't want to have every application that's added to these systems have to go through that same arduous path.

We want them to be developed independently at the rate at which new applications, new content, and new capabilities can be developed for the end user.

The key to making this happen, in addition to this clean separation, is that the applications are written to a general-purpose processor that has sufficient performance headroom and scalability for the future. Because as we all know, we want to re-use as much software as possible as we move to greater and greater sophistication and new applications, and we want to be able to roll that forward. That requires scalability and performance headroom.

So let's look at some of the new kinds of clients that can be developed according to this paradigm that has been developed according to the PCA paradigm.

You can see on this picture, there's a wide variety of different handheld form factors, some with integrated cellular connectivity, some without, some that look more like a traditional PDA, some that look more like a phone.

All of these have been designed around the Intel(r) StrongARM* processor, which delivers that high-performance capability at a very low power to go into these kind of devices. In fact, the StrongARM processor still, today, even though it's been in production now for nearly two years, leads in terms of the figure of mere milliwatts per MIP, which is something you want as low as possible. Low power, high performance.

It's my pleasure to show you an example of a new client using the StrongARM processor and using our flash memory. This is one of the first so-called 3G phones. This is an NTT DoCoMo phone on what they call their FOMA network. And this is exactly the phone that is now in the initial trials and prototyping of the first 3G network, based on our StrongARM processor.

So this is a little bit of the glimpse into the future. As you can see, color display, sleek-looking form factor, new kind of device. Fits in your pocket.

Now, the other thing that's important in these devices is, we absolutely need nonvolatile, low-power code storage. We need this to store the communications protocol. And increasingly, as we add new data applications to these kind of devices, we also need it to store those applications as well as end user data.

Some time ago, Intel introduced something we call Intel StrataFlash(r) memory, which enables the storage of two bits of data in a single cell. We also know these devices, as the one I just showed you, are very space-constrained. So having that high-density storage in a small amount of space is also very important. And by putting two bits per cell, you can imagine what you get.

Well, Intel StrataFlash is now being designed into phones. It's designed into a large number of the PDAs. It's really become a very ubiquitous type of storage for these kind of devices. In fact, I'm very, very proud to say that as of this moment, we have now shipped over two billion megabits of this StrataFlash memory.

Now, I talked about the new range of clients and how we need to get these new clients out there and how in order to attract the applications and develop the applications at Internet speed, we need to have these applications available on all kinds of devices. Not that every application would be available on every device. But we need for a preponderance of these to have some level of seamless application from one device to another.

In order to achieve that, we really need to enable the technology suppliers, like people in this audience, to bring those basic core capabilities in terms of things like media players, audio players, middleware, content, to these new class of devices and have, for the end user's benefit, a relatively seamless portability of those content and applications from one device to another, such that a PDA would give you a subset of what you'd see on a desktop or a mobile PC, and a cell phone would give you a good subset of that.

So I'm very happy now to introduce an example of this. Joining me on stage is going to be Peter Meehan, vice president of Macromedia. We're going to give you a little demonstration of what this looks like. Peter, welcome to IDF.

PETER: Good morning, Ron.

(Applause.)

RON SMITH: Okay. So, Peter, what can you tell us about the relationship between Intel and Macromedia?

PETER: Ron, as you know, Macromedia and Intel have been working closely together for the past number of years trying to optimize Macromedia software specifically on the Intel architecture. There's a number of examples, one of which might be Macromedia 8.5 Shock Wave Studio*, which includes the Intel Internet 3D graphics engine that now enables people to create compelling 3D graphics, distribute them in realtime across the Internet. So that's big news.

We also continue to work in Macromedia Flash software. It's the most widely distributed piece of software on the Internet. It covers about 97 percent of all desktops now with the Macromedia Flash Player. And we look forward to working together with you as we go forward.

RON SMITH: I'm sure you're going to be able to show us an example of that cooperative work here, right?

PETER: Yeah. We have some pretty neat stuff we'd like to show you today.

RON SMITH: Let's take a look.

PETER: We share this vision that there's one and only one Internet. You're going to want to access common data, common data and applications across a wide variety of devices. It has to be a scalable Internet experience.

So what we'll show you is the same content running on a PC, a PDA, as well as a mobile phone. So here's an application that the people at a company called FunMail created which takes simple text-based messages, wraps them in a rich media experience, and adds animation, text, graphics, as well as sound.

RON SMITH: Wow. And that's really a powerful message.

(Laughter.)

PETER: Yes. Certainly it's been optimized for the PCA architecture. And an example of this is the Compaq iPAQ*, which has an Intel StrongARM processor running inside of it.

PETER: So what we're going to do is go and show the same application. It has a resolution that's a little bit less than the PC. It has a limited amount of colors relative to the PC. But you see it still has the same type of rich media, graphics interactivity, and the performance is just about the same as a PC almost.

RON SMITH: It sure is powerful. And it continues to give you that powerful message.

PETER: So the next thing we'd like to show you is the NTT DoCoMo phone. This is the phone actually being utilized today in the marketplace. We'll show you the same content running on this phone today. So you can see that, you know, it's a much smaller screen than the previous devices it's not as colorful. It's not as interactive, it's not as high performance, but still you can take the same content and transition across these different devices. So you have a scalable Internet experience, depending upon the capability of the device.

RON SMITH: It's great. So the users can move from one device to another and, you know, when they go to the smaller device, they have some subset of what they know and love on their PC as well.

PETER: You bet.

RON SMITH: That's great. So, you know, this is really neat. We all love cartoon characters and so on for consumer apps. But what about businesses? I mean, can businesses take advantage of this kind of technology?

PETER: That's a good point. There's actually a really compelling application in service that was developed by a company called Ice Works in collaboration with the National Basketball Association and the Staples Center down in Los Angeles, where people that went into the finals in June would get a Compaq iPAQ if they were in the slip section and, of course, had a wireless connection. It was fully loaded with the service I'd like to show you right now.

So, again, this is native Macromedia flash running directly on the device. And what you'll see is the front end of the application allows for the consumer to go and look at information on today's event, namely, the basketball game itself. People would be able to sit in their seats and order food or drinks and have that directly brought to their seats

RON SMITH: So we have different icons for food and merchandise.

PETER: Fun and games. And also some information about the Staples Center itself, which we'll show you in a second, that allows you to preview visually where your seat might be at a future event and then purchase that online.

RON SMITH: Now, I understand that, actually, in the last game of the playoffs at Staples Center, they actually passed these out to people and were actually using some of them.

PETER: Yes. So this was, you know, real product that was in trial and actually got tremendous feedback. Because it's an engaging, compelling, rich media experience that kind of melds not only the Macromedia flash side of things, but also takes full advantage of the capability of the device itself.

So here you see two logos of the two different teams. You could go and look at information on today's lineups, statistics of individual players, the team. You could get live video clips and those types of services. You could also look at the individual players on the teams.

If you're sitting in your seat and you see the players out on the floor, you're not quite sure who they are, you get a graphic of the teams on the floor.

RON SMITH: Yeah, who is this big guy in the center? He's kind of relatively unknown there.

PETER: Yeah, nobody knows who this guy is. Look at his profile, and we'll call up a photograph that has when Shaquille O'Neill was born, how tall he is and other personal information.

RON SMITH: That's a big dude.

PETER: Very large man.

RON SMITH: Wow! I'm really getting familiar with this sport now.

PETER: So there's all kinds of great things that we can do. If we wanted to not just look at today's event, but also go in and look at a future event, we could go in and get some more information on the Staples Center itself. You can also go and look at all kinds of different things.

Let's go and grab some information on a future event so that we can see how I might be able to go and look at the actual seat I mentioned to you before.

RON SMITH: There you go.

PETER: So here's the graphic of the Staples Center itself. Again, Macromedia Flash is a compact, lightweight, broadband experience even through a wireless device here. If I want to look at a seat I might want to purchase in the future, this is what the view might be.

RON SMITH: I recognize that. That's the nosebleed section. That's all I can usually afford.

PETER: And then there might be, you know, another seat, if I wanted to pay $100 or so to get a view like this, you could look at it online and say, gee, maybe it's worth it, and you'd want to purchase that ticket.

RON SMITH: Where is Jack Nicholson?

PETER: I think he'd be sitting right about there. But anyway, this is a business application. It's real. It's been launched in a trial. And there's also other examples that utilize Macromedia Flash on Intel PCA devices that are quite interesting. We're getting a lot of great feedback.

RON SMITH: This is awesome. Looks like Intel and Macromedia working together on PDA is a slam dunk.

PETER: Absolutely.

RON SMITH: Thank you for coming. Great demo.

(Applause.)

RON SMITH: Okay. So that was just one example. But today, we're announcing something that's going to enable many, many more developers to participate in this kind of application development and portability from one device to another around the Intel PCA architecture.

So today, we're announcing something we call the Intel PCA Developer Network, and this is a network of tool suppliers, operating system and independent software vendors, client manufacturers, even service providers with an offering of capabilities to help new solutions get developed and brought to the marketplace for these new kinds of clients.

Now, I could spend all day up here talking about this, I'm so excited about it, but I'd much rather show you. So I'd like to introduce to you now Ken Edwards, who is in charge of our PCA Developer Network marketing. Ken.

KEN: Hi, Ron.

RON SMITH: Thanks for coming.

(Applause.)

RON SMITH: So Ken, what do you have to show us here today?

KEN: Well, Ron, I'm going to show you what we're delivering in the Intel PCA Developer Network. We're providing development support, technical support, and marketing support for companies that are speeding the delivery of the wireless Internet in the form of applications, devices, and services.

We're providing that support through an easy-to-use Web site, and I'd like to walk you through the features of the Web site.

RON SMITH: Now, this Web site is actually live as of this moment. This isn't just a mockup. This is live, and these developers could go out there, click on the URL and off they go.

KEN: That's right. You can go by the Intel booth and you can actually register in the Intel booth and become a member of the Developer Network.

RON SMITH: Okay, great. Excellent. So what's on here?

KEN: Okay, this is the home page of the PCA Developer Network. If we look at something simple like development support. Here you can see we provide information, technical information, about the network, white papers, application notes. So if you're developing a system, you can get application development support.

Let's look at something like the Integrated Performance Primitives. Last year at IDF, you showed a demo of a company that used these primitives to shorten their time to port an application from IA to PCA.

RON SMITH: Yeah, that was the blinded with science guy and Britney Spears. You're not going to sing and dance, are you?

KEN: No, I don't sing and dance. But it took them only a few days instead of the months it would normally take because with IPP we bring primitives for MPEG, JPEG that companies can use. Companies like Microsoft are seeing 10 to 15 percent performance improvement by using IPPs.

RON SMITH: Awesome.

KEN: In addition, on the development support page, we highlight information, and today we're highlighting the Intel Flash FDI 4. This is technology that allows six times performance improvement, works with Intel StrataFlash, allows the system now to be the lowest cost, highest performance.

RON SMITH: You got it.

KEN: And you can see a demo of this again at the Intel booth where Motorola worked with Jamdat to develop a system using this technology.

RON SMITH: We're announcing this system today, FDI really optimizes for both code and data storage in the single Flash memory, and we already had the support tools live on the Web site.

KEN: On the Web site.

RON SMITH: Awesome.

KEN: If you're a developer and developing a system, say you run into a real technical problem. We provide technical support. We have a page that we have frequently asked questions, a glossary of terms, training. But in addition, we provide a way that you can directly contact Intel. You've got a detailed technical question, it's confidential, you can e-mail it to an Intel engineer. Here we reply directly to you.

RON SMITH: Awesome. Just don't send it to me. It might take a longer time to debug.

KEN: Once you've got the product finished, you're looking for marketing support. You want to get your product into the market. We provide marketing support with case studies, product showcases, and the thing to always remember, you're now part of a Developer Network. As part of this network, you're part of a community of other systems and companies that are also accelerating the wireless Internet and working with Intel PCA. So you can take advantage of it.

RON SMITH: So they have a collaborative network online working together towards providing full solutions. That's awesome.

KEN: And one of the key features is our solutions catalogue. Here is where all the products that support the Intel PCA come together so a customer looking for support can come here, and he can look through all the companies that support Intel PCA. You've got people like Adobe.

RON SMITH: Alphabetical.

KEN: Alphabetical. Microsoft is down here.

RON SMITH: IBM.

KEN: We've got over 200 companies currently part of the network.

RON SMITH: Each has its content and how it relates to PCA and our product on there as well?

KEN: Absolutely. So if you're looking for a tool, a C++ compiler, you can come and search on it by product type and say show me the C compilers that support PCA, and you get a list of it, click here and get product information and data on those products.

RON SMITH: Excellent. Wow, this is pretty complete and pretty awesome. This ought to help our developers speed new applications to new clients, get collaborative efforts, and get the solutions to the marketplace much quicker.

KEN: Absolutely. I'd like to encourage everybody to come by the Intel pavilion, look at the Developer Network, join the Developer Network, and join Intel with the PCA.

RON SMITH: I'm sure they're all anxious to do so. Thanks a lot, Ken. This is great. Thank you very much.

(Applause.)

RON SMITH: So what you see here is our new Developer Network. It enables advance information on Intel products, what is coming, how you start your design and your effort. It has access to the suites of development tools and platforms. You saw some examples of that. Some of their software building blocks like the IPPs and the FDI which we featured here. Basic application and technical support, a live wire back to our technical support people at Intel. Cooperative marketing efforts between all the members of the network to go work on different solutions and how to market those. And networking and match working opportunities among the various developers.

So once you're in this network, you get access to all those tools and all those capabilities from all the companies that are members. And you can see, it's real easy to get to. Just go to www.intel.com\pca\developernetwork. Fairly straightforward, easy-to-remember URL.

As Ken mentioned we have very broad membership. You see a lot of familiar names here. There's over 200 members and more than 300 tools that are available today at introduction, all live on the Web site. And there's many, many more to come. I hope, as you all walk past our booth, sign up, and get your capability referenced in here as well.

Now, I'm also proud to announce that we have seven new companies that have endorsed the Personal Internet Client Architecture. These are people who have come out with public support because they understand the need that has to be satisfied here and how PCA is going to enable this new ecosystem of new kind of devices. And they include carriers, mobile service providers, hardware manufacturers, OS manufacturers, as well as independent software vendors giving some of the core tools.

So we have, on the one hand, a service provider pretty well-known, AT&T Wireless. We have hardware manufacturers in Compaq and Hewlett-Packard. We have operating system vendors with Symbian who is doing the operating systems for some of the smart phones among some of the major manufacturers in Europe, and Palm who we're all familiar with, also a provider of an operating system, as well as our friends Macromedia who joined us on stage and Adobe who provide some of the key underlying tools capabilities for really developing these rich, new applications and deliver the applications themselves to the market.

So the ecosystem is growing. We've had tremendous momentum. It's been less than a year since we've introduced PCA and the marquee companies are lining up behind it.

Now, I've talked a lot about the concept, I've talked a lot about the architecture, we talked about the Developer Network and tools, but I do want to remind you that Intel does provide core building blocks to help deliver this architecture to the marketplace.

We talked about our baseband chipsets and demonstrated our Micro Signal Architecture earlier here in the year. This is for the communications portion of this. We, of course, have Flash memory. We've talked about over 2 billion megabits of StrataFlash have now shipped. We've shown you some of the key enabling software we have like the Flash Data Integrator and Persistent Storage Manager. And, of course, our flagship for the general-purpose processor that's key to this architecture has been the StrongARM processor. What you've seen is how designed into a whole myriad of new kind of clients.

Last year, we also gave you a peek at what an XScale microarchitecture looks like, this is the next generation of the StrongARM architecture and what I'd like to do now is give you a little bit more of a peek into the future of this capability around the XScale microarchitecture really brings to these new kinds of clients and what you're going to be seeing in the future.

So joining me on stage now is Sammy Tao. He's one of our application engineers from our handheld computing division. And you'd like to show you a little something here. Sammy, welcome to IDF.

SAMMY: Thanks a lot, Ron. Great to be here.

(Applause.)

RON SMITH: So, Sammy, what do you have to show us here?

SAMMY: Well, Ron, we have two reference systems here. They're built by this Ohio-based company, Axilent Systems, and they're configured almost identically. So they have both the general-purpose processor. What I like to propose is that we do this taste test. Instead of the Pepsi challenge, we'll do the PCA challenge, see if you can see the difference.

RON SMITH: There's only one major ingredient that's difference, I wonder what that is.

SAMMY: Yes. One was the StrongARM processor. The other has an Intel(r) Xscale(tm) microarchitecture based processor, the first silicon. So what's on these boards, they're configured the same. We can get a good side-by-side comparison. They have all the components that handheld and wireless and mobile developers need, including Intel StrataFlash.

RON SMITH: Now, what are you actually going to show us in this demo?

SAMMY: Both of these run Windows CE*. And there's a video decoder. And back in the lab, we're seeing performance nowadays of about 15 frames per second, 320 by 240 video. And in the next generation, up to about twice that, 30 frames per second.

RON SMITH: We're going to really show full video that might be used in an end user application.

SAMMY: Right.

RON SMITH: And we're going to be doing the decoding completely on the general-purpose processor?

SAMMY: Right. Completely on a general purpose processor.

RON SMITH: We don't need a DSP for this?

SAMMY: That's right. Let me show you.

RON SMITH: Great. Let's take a look.

SAMMY: So here we have the first example. You can see that the motion is pretty good. There's a little bit of artifacting, motion artifacts and a little jerkiness. But this probably represents what's best in class in today's technology.

RON SMITH: Okay. That looks pretty good.

SAMMY: Yeah.

RON SMITH: That's about what people would expect from video running on a handheld device. With all the motion and interaction in, there they'd probably be reasonably satisfied with that.

SAMMY: There's a lot of heavy-duty computation going on in there.

RON SMITH: What does this other thing do? I mean, how does it look?

SAMMY: Well, now you can see the difference here. I mean, look how smooth and fluid it is. There's very little motion artifacting, not as many dropped frames as before. Really smooth, nice, just the wave of the future.

RON SMITH: That truck looked like it was actually rolling down the street.

SAMMY: It wasn't jerking across like the first one.

RON SMITH: That's cool. Gee, now, I wonder what the difference is between these two platforms. Do you think we could unveil that here, have a real unveiling of what we really saw?

SAMMY: Sure. Sounds like a good idea.

RON SMITH: On the right was the first one. On the left was the second one, just so we don't get confused here as we change classes on this challenge.

SAMMY: Right.

RON SMITH: Let's go. Let's do an unveiling. Wow, what a surprise. The first one is our current SA 1110 StrongARM processor delivering state-of-the art capability today. And this little baby is our new application processor based on XScale.

SAMMY: It's a sneak peek on first silicon of XScale-based general-purpose processor. It's really going to enable the next generation of device, mobile, wireless, handheld clients based on PDA.

RON SMITH: Does it surprise you that XScale showed better performance?

SAMMY: No, definitely.

RON SMITH: Absolutely not.

SAMMY: Yeah, that's right.

RON SMITH: So thank you, Sammy. That was a great demo.

SAMMY: Thank you, Ron.

(Applause.)

RON SMITH: Now you have an idea of what kind of capability, what kind of full, even multimedia, capability the new XScale microarchitecture will unleash, enabling yet a new class of handheld devices.

So, let me summarize the key messages around our accelerating to the wireless world.

There's only one Internet. And what we want to achieve is the most ubiquitous access to that Internet on any kind of client, whether you're using it in your office, whether you're taking it with you on the road in a mobile PC, or whether it's something you have in your pocket or on your belt, or in your purse.

Intel PCA enables these wireless handheld developments both -- in a new paradigm to keep pace with the Internet, both in terms of the clients themselves and the applications. And the products and the tools to support those products are ramping in -- very, very fast today.

An example is we've launched the PCA Developer Network to show you the new kinds of capabilities that are out there to help establish this new paradigm and bring these new products to market at Internet speed. The momentum around PCA has been far in excess of what we expected, and it continues to build with new market-maker endorsers like the seven marquee names we talked about today.

And the XScale microarchitecture will deliver scalable performance based on compatible with the StrongARM for future applications.

Now, with that, I'd like to summarize what you've heard today from Louis and Frank and myself.

We're seeing a changing landscape out there in terms of new and richer kinds of clients. And, in addition, these new clients are also going to connect with our new server technologies, which you'll be hearing a little bit more about tomorrow. And Intel can provide the building blocks in all these key areas.

So every major junction in the Internet, Intel has the building blocks to help enable that, enable a new class.

The key is that we develop successful solutions. We've talked a lot about the enabling technologies to provide that in terms of I/O, in terms of a balanced mobile platform, in terms of development tools and capabilities for the PCA architecture.

We need to facilitate collaboration. That's why we have a developer forum. That's why we continue to work with all our industry partners to help deliver these new innovations and these new capabilities. And we need that among all types of devices. And that's why you see us continuing to invest in new I/O capability, continue to invest in new technologies to enable more capable notebook computers, and why you see us invest in new kinds of technologies for the handhelds and cell phones.

Louis talked to you about new I/O innovations that will advance the extended PC. The connectivity among the clients is also key. So having new capabilities like the third-generation I/O, the USB 2.0, and the serial ATA are key to developing that extended capability in and around the PC.

Frank talked to you about how the mobile PC is now measured on new value vectors. Yes, performance is still important, but battery life is important, connectivity, wired and wireless, connectivity everywhere is important, and so are new innovations and form factors. And the key is all four of these vectors need to continue to be optimized, and we showed some of the underlying technology that will enable that to happen.

And last, the Intel PCA really enables the movement of the Internet to the next frontier -- that is, available anytime, anywhere -- through wireless connectivity and through a wireless network. And we've introduced today a concept to help accelerate that, a network of developers which you all are welcome to join called the Intel PCA Developer Network.

Intel will continue to provide the innovative building blocks, will continue to work with our industry partners to develop these new class of clients, and to keep the enthusiasm and the innovation around these clients going.

So with that, I'd like to thank you very much for participating in our client-day keynote speeches.

Thank you.

About Intel
Intel (NASDAQ: INTC), the world leader in silicon innovation, develops technologies, products and initiatives to continually advance how people work and live. Additional information about Intel is available at www.intel.com/pressroom and blogs.intel.com.

* Other names and brands may be claimed as the property of others.