Intel Developer Forum, Spring 2001

Ron Smith
Tokyo, Japan
April 17, 2001

CRAIG BARRETT: Welcome to the Intel Developer Forum here in Tokyo. We've been working at Intel to make this conference your premier event to learn about technology for computing and communication.

You'll discover the latest technology trends. You'll receive graduate-level technical training, and you'll be able to discuss opportunities and challenges with peers and industry leaders.

Our industry is certainly one of the most exciting in the world. It has its ups and its downs. It's important to remember, though, that downturns always end, and that's what this conference is all about. It's about preparing for the next upswing, for the next upturn.

We know that technology always moves forward, and that's really why we're here today. As the world continues to turn digital, we all have a wonderful opportunity.

The Internet is the growth engine of the future. It's the center of innovation. It encompasses communication, information, and commerce.

What we're here to do is to develop great new technologies for this digital world of the future. We all have to invest in new technologies and the design and development of new products. In fact, if you look at Intel's actions this year, you'll find we're spending more capital than ever before, seven and a half billion dollars. We're spending more R & D dollars than ever before, 4.3 billion dollars. All this is to bring new products and new technology into the marketplace.

Throughout this conference you'll be hearing about our plans, you'll be hearing about our four architectures for the Internet.

The first is the PC platform. It's really characterized by our IA-32 bit family, and that's really embodied by the Pentium 4 processor and the extended PC concept.

The second architecture is the handheld client, something we call the Personal Intel Client Architecture or PCA. It's really for handheld computing and communication devices.

The third architecture is our networking family of products, the so-called get exchange architecture.

And the fourth architecture relates to servers. This is characterized by the Itanium processor family.

These architectures represent a significant opportunity going forward to span the Internet in all areas of computing and communication. We continue to evolve the PC architecture.

Japanese developers have contributed much to this evolution as witnessed by the ultra slim notebook and small form factor PCs.

Intel continues to lead the industry with low voltage, low power, high performance microprocessors that serve these important and rapidly expanding market segments. And we're excited about the global opportunities that now exist based on Japanese innovation in this area.

Intel is also very excited about the networking and communication marketplace. In this market segment, we continue to make substantial investments in research and development and bring new products into the market.

We've been very excited about the development of the DoKoMo i-Mode phone in this space in Japan.

The Intel Developer Forum is a place where we work together to accelerate innovative open technologies and solutions in computing and communications.

I want to thank you for helping us expand the power of the Internet, and have a great conference.

ANNOUNCER: Ladies and gentlemen, we would like to thank you for coming to the Intel Developer Forum Spring 2001 Japan. We would like to begin.

First of all, for the keynotes address, from Intel Corporation, we have senior vice president and general manager of the wireless communications and computing group, Ronald Smith, who will be talking about accelerating the Internet to wireless.

We're sorry to mention that there has been a change in the title, but we would like to call upon Mr. Ronald Smith to make the presentation.

(Applause.)

RON SMITH: Thank you.

Good morning. Welcome to the Intel Developer Forum for 2001. I'd like to talk to you today about the next-generation network and how we're going to accelerate the expansion of the Internet into the wireless domain. This is an area where Japan has taken a leadership role, and I'd like to talk about what Intel is doing to help facilitate this movement.

I'd like to, but the mouse isn't working. Do we have a backup? Okay. Let's try it. All right. There we go. Backup always works.

Yesterday's networks, as you can see, were largely different networks for voice, data, and wireless. They tended to be very proprietary, very vertically integrated.

Now, what you're going to see in the future is one universal network that's based on Internet protocol that's going to span from the conventional enterprise network through the voice networks to the wireless voice and data networks.

And key elements of this evolving network are the fixed and mobile clients, be they desktop PCs, mobile PCs, or cell phones and PDAs; the basic infrastructures, the hubs and routers; and the backbone, the basic servers that are in the back room that are basically running the network.

And as you heard Craig describe, Intel has four main architectures which are really designed to address these various elements of the total evolving network.

First of all, there's the Intel architecture which is a well-known backbone of the desktop PC and the mobile device. It's also used in some of our servers.

We have a new architecture for the server family, the Itanium processor family, which is our 64-bit architecture, and we have in the basic infrastructure, the Intel Internet Exchange Architecture.

Now, you'll hear a lot of speakers talking about these different elements through the course of this conference. What I'd like to focus on today is the Intel Personal Internet Client Architecture, which is our main architecture for the new emerging wireless handheld devices.

So let's talk about how this new architecture fits into the new wireless Internet networks.

First of all, if we look at a worldwide basis today and we look at what's happening in the wireless communications market, basically the average revenue per user for voice services is actually on the decline. Even though the number of subscribers is increasing, the amount of revenue that that mobile service provider gets for voice services per user is actually declining.

And in the meantime, we're going through a whole new infrastructure deployment. Here in Japan, it's just starting next month with the first deployment of the so-called 3 G network or wide band CDMA. In other parts of the world they're in the midst of the two and a half G deployment, and ultimately also the 3 G deployment.

Now, these capital costs in many parts of the world also involve very high cost for spectrum, for the so-called 3G spectrum. It's very difficult for the mobile service provider to really make a profit just based on voice with this kind of market. So it's absolutely imperative that new sources of data revenue actually come about. And you've seen this in Japan with the packet data services that are now available, and there's a large number of users for these services in Japan.

Now, to overcome this and to really provide rich new capabilities for this mobile service provider applications, to develop these sources of new revenues, means much faster to delivery of new applications and services than have occurred up until now with the conventional paradigm for cell phone and wireless infrastructure development.

The other thing that's going to be required is if you really want to do data, you need a much richer client to take advantage of these new services. This is what Intel's technologies are all about addressing, are these two key areas.

And the key architecture that addresses these is the Personal Internet Client Architecture, which we introduced last fall here in Japan. And what this client architecture is about, if you look at the basic device, it has three major subsystems: A computing subsystem that today, in a cell phone, really only does real-time control, but increasingly there will be actual computing done in these data devices. We have, of course, the communications subsystem, which interacts with the air interface. And of course you have a memory subsystem for storing not only communications code but, increasingly, the end-user application and data.

And what the Personal Internet Client Architecture does is it basically abstracts the software stack and the hardware from these basic subsystems. So even though the actual silicon may be integrated together in various forms, it looks like three independent subsystems. And they each can evolve separately based on the rate at which that technology evolves.

The other important thing is that the applications can be written to a general-purpose, high-performance microprocessor. This is really important, something we learned from the PC space, to get a whole thread of applications going, that you want a general-purpose microprocessor.

The benefits are many. For the hardware OEM, it enables them to deliver more products off of a basic investment and deliver those to the market faster. So in other words, one could change the computing subsystem while keeping a communication subsystem constant and not have to go through type approval before they introduce a new device. And off of the same basic platform, you can actually develop a wide range of devices.

For the operating system and content vendor, it provides performance headroom for developing new applications, and a commonality that's going to span multiple products and, in fact, multiple generations of technology.

And for the service provider or carrier, it enables more new applications and more new services to get into the market more quickly, to help them get those sources of that data revenue I talked about earlier.

Now, I could keep talking about this, but I think it would be much better if we actually showed you a demonstration of some of these kinds of capabilities. So joining me today on the stage is Timothy Childs who is the vice president of strategic relations for Eyematic Corporation.

Timothy. Hi, welcome.

TIMOTHY CHILDS: It's a pleasure to be here.

Let me show you some of the amazing things we're doing over here at Eyematic.

RON SMITH: Okay.

TIMOTHY CHILDS: Well, what you're going to see today, some of the 3D technology you're going to see today as part of our rich media authoring and publishing solutions that we have. You know, with so many new devices that are having the SA-1110 chip in it, it enables us to play just about anywhere these days.

And, in fact, we have actually ported over some of our 3D players onto the SA-1110 architecture, and let me show you some stuff. Do you recognize this fella?

RON SMITH: Yeah. Hey, that's my boss, Craig Barrett.

TIMOTHY CHILDS: Indeed. You know, we bumped into Craig the other day at Intel, and with our Eye Publish software, all we needed was two photographs and we're able to create a realistic-looking 3D character running on a Pentium 4.

RON SMITH: Wow, that's great!

So it's clear that the SA-1110 actually gives you the performance to view 3D content --

TIMOTHY CHILDS: Mm-hmm.

RON SMITH: -- like this good-looking guy we have here, but, you know, static images don't do a whole lot for the end user.

TIMOTHY CHILDS: Right.

RON SMITH: What about adding the subtleties of facial expression and lip sync? I mean, these are pretty labor-intensive tasks to program into these devices, are they not? How do you get past that?

TIMOTHY CHILDS: That's a very salient point, because as you said, doing facial animation to point has been very arduous and very difficult to do to date.

What we try to do at Eyematic is radically simplify that process so we can easily capture facial expressions using our Eye Publish software and a simple Web cam. And actually even make him talk. Would you let me show a quick demo?

RON SMITH: Yeah, why not? If he can talk, why don't you see if he can give me a salary increase.

TIMOTHY CHILDS: Okay. Let me -- Hey, how about a raise? Ha Ha.

So, in fact, we used this software to create a little piece of content for you. Let me see here. Hit play.

Let me see here.

RON SMITH: Okay.

TIMOTHY CHILDS: Hit play.

CRAIG BARRETT: Hi, Ron. Glad to see you're having fun. I understand you're in Japan today spreading the good word of Intel PCA. That's great.

Okay. You better get back to work. Say hi to our friends over there for me.

RON SMITH: Sure thing, Craig.

Oh, you know this is great. So content creators can actually bring new characters to life about as fast as they can conceive of them.

But, you know, what are the real-world applications for this?

TIMOTHY CHILDS: That's a perfect question, Ron. And let me show you how Eyematic's technology with empower real-world applications that are available today in Japan over wireless services.

So you can imagine our content service provider friend Cybird has a piece of content application service called Robo Robo, and as you know, Robo Robo is a Java-based character and cartoon animation service that they provide.

Now, if you can imagine such services such as news and information and sports and things like that empowered with much more rich communication and facial (inaudible) and stuff like that, it's very, very exciting. In fact, I think we have a message made just for you on the iPAQ. Can we switch over to the iPAQ, please. Great. Can you hit, play, please.

CRAIG BARRETT: Hello, Ron. It's Craig. I'm in California right now, and I wanted to see how everything is going today. Are you getting all those StrongARM and XScale design wins you promised? I seem to remember something about lots and lots of software taking advantage of our IPP libraries. See you again tomorrow. Davomata (phonetic).

RON SMITH: You know, if only Craig to see for himself.

CRAIG BARRETT: Hello, Ron, it's Craig. I'm in California right now --

RON SMITH: But if only he could see for himself some of the cool things that are going on in the mobile market here in Japan.

So with these realistic graphic images, we can even sense emotion.

TIMOTHY CHILDS: Indeed.

RON SMITH: And, you know, that has its good points and its bad points. I'm not really sure about this idea of getting messages from your boss when you're on the road.

TIMOTHY CHILDS: Well, you know, sometimes that can be really good or really bad, depending on what the news is, obviously. So -- But what's even more fantastic than that is how the carriers are responding to how Eyematic's technology enables new types of applications that expand just beyond games into new revenue streams.

You know, Cybird and other content vendors can use this technology to deliver a wide range of information services such as what we said before, surf, news, and sports, and stuff like that.

RON SMITH: Right.

TIMOTHY CHILDS: But can you imagine your favorite movie star delivering, say, your stock information first thing in the morning to you?

RON SMITH: Well, I'm not sure I could stand that first thing in the morning these days.

TIMOTHY CHILDS: Well, you know, it's kind of been rough lately, but I think it's going to get better, and we sure hope so.

But the main thing I want to leave you with today is that using Intel technology, our content is truly scalable. Since we only need one kilobit per second upstream, we can animate these characters over mobile devices that you see here. If you can switch over to the iPAQ, to this phone here. As the performance of the chipsets and the devices increase, the image quality increases dramatically.

This is very significant for us, and we're very, very excited about what the StrongARM S-1110 empowers us to do. And the SA-1110 allows us to do some amazing things today, and it's going to be even more incredible later this year with XScale coming out. And we're very excited about that.

RON SMITH: You're right. It is going to be incredible, and XScale is going to be incredible. And I want to thank you very much.

TIMOTHY CHILDS: It's our pleasure, Ron.

RON SMITH: It's a great demo.

TIMOTHY CHILDS: Thank you so much. Take care.

RON SMITH: Good luck.

Okay. I'm having mouse difficulty again here. Obviously -- There we go. Whoop.

Okay. So as you can see, this is one of the many kind of new applications that one can get with the high performance microprocessor and this kind of architecture we've talked about with PCA.

Now, we're working with a whole host of different application developers. You can see from their logos that many of them are recognizable companies here in Japan as well as in other parts of the world in this mobile space.

One that just recently joined the fold was IBM Corporation, and they're going to work together with Intel to enable PCA-based Internet-ready wireless devices and applications. And IBM is going to port its new WebSphere Everyplace suite, which is the embedded edition of it, it's part of their pervasive computing initiative, to the Intel Personal Internet Client Architecture. And this will provide a set of embedded client middleware for handheld and wireless solutions.

And Intel will feature and distribute these with the middleware and tools to developers in our forthcoming PCA software development tool kit.

If you want to find out more about what's available, I suggest you go check at the developer.Intel.com, and it will talk more about some of the capabilities for developers.

Okay. So we talked about why the independent computing and communication stack is important for getting new devices to market and to open up the opportunity for developing these kinds of applications. So let me talk about the different building blocks Intel is going to provide in each of these areas.

First in the computing area, we already have our StrongARM processor, which is already in devices like the Compaq iPAQ that you saw demonstrated, which gives leading power performance capability today.

And we're going to be augmenting that later this year with our XScale Microarchitecture products for this space, which will deliver a quantum leap in the power performance; by 10x, the performance at the same power, or one-tenth the power at the same performance.

Now, another capability that we have that goes with this is something we call the Intel Integrated Performance Primitives. What these are are pre-optimized blocks of code that are already written to the media instruction set. So it really helps give a performance improvement for things like audio capability, video, and graphics. They're written independent of the OS, because it's written directly to the hardware, and the function calls are absolutely identical across all of our architectures.

And what this enables one to do is write something, say, on a PC using the IA-32 architecture, and that part of it that's optimized can be automatically ported over to the handheld platform using StrongARM or XScale.

And this is a really compelling kind of capability which I'd like to talk to you a lot more about, but I think it would be easier to just demonstrate it.

So joining me today on stage is John Eckstein who is the general manager for Beatnik Japan.

John, welcome.

JOHN ECKSTEIN: Nice to see you today.

RON SMITH: Good to see you, too.

JOHN ECKSTEIN: Didn't know you'd be here.

RON SMITH: Oh, I think that was a little bit of a put on, but, John, what do you have to show us here today?

JOHN ECKSTEIN: Well, thanks very much. I'm very happy to be here to show you some of the cool audio applications we're doing at Beatnik.

Actually, just last week, Compaq Japan announced the release of the Japanese iPAQ. The Japanese iPAQ will be bundled with the Beatnik player, which takes advantage of the Intel SA-1110 and and StrataFlash memory.

RON SMITH: Would you, SA-1110 and StrataFlash? That's really music to my ears.

JOHN ECKSTEIN: Come over here and let me show you what we're doing.

RON SMITH: Okay; great.

JOHN ECKSTEIN: Actually, I've got an iPAQ right here, and this is the Beatnik player. Let me see if I get it set up right there.

RON SMITH: Yeah, let's set it up to avoid reflections. Okay. That looks cool.

JOHN ECKSTEIN: This is the Beatnik player that's running on the Beatnik Audio Engine or BAE. We originally developed the BAE to run on the IA-32, but we wanted to deliver this application for handset applications as well.

Given the power and performance of the StrongARM, we decided to use the SA-1110, and that's just ideal for this type of a rich multimedia application.

Now, the Beatnik Audio Engine is a wave table synthesizer and sound sampler that we license to developers. We also work closely with artists and record labels in Japan. In fact, one of our partners is a popular group in Japan named Dreams Come True or Dorikam (phonetic).

Dreams Come True has been using Beatnik's technology on their Web site. They're using it to help promote their own music.

Oops. No, that's Britney Spears, actually.

Their Web site is dctgarden.com, and you can go there and hear some of the Beatnik technology there.

So today I'm going to show you a song called 24/7 by Dreams Come True using the Beatnik player, and this shows you the type of interactive Karaoke type application you could make with the StrongARM.

So, a one, a one, a one, two.

(Music playing.)

JOHN ECKSTEIN: So there are multiple channels of sound here, and I can turn them on and off. Can we have a little more audio, a little higher. So Yoshida Mila has a beautiful voice and I want to listen to her voice. So by clicking on this, I'm soloing her voice.

Now, I can turn on other channels of sound. And you can make your own song this way.

So that shows you the kind of application you can make, but it was importing from the IA-32 where the Intel IPP libraries really helped us.

The Intel IPP libraries or Integrated Performance Primitives we actually downloaded off the Intel developer site, and they helped us make the port from the IA-32 to the StrongARM in just three days.

Not only that, but because they're optimized, it helped us reduce the decode time for an MP3 string from 70 seconds to only 3 seconds. And the best thing for us was that they were very quick to use, and our time to market was much faster than it would have been otherwise.

RON SMITH: Wow, that's really great news, John. 22x performance and faster time to market? What a fun application.

You know, I can see that the Beatnik Audio Engine is going to be a big hit with all those people who are musically inclined like myself. I can't wait to get back to the office to do a little Karaoke myself. Maybe I can even have Craig the Avitar join me.

JOHN ECKSTEIN: Well, I'm sure they're all looking forward to hearing your Karaoke back at Intel.

Anyway, we're very, very pleased with the performance of the StrongARM. We're running 32 channels of audio on a StrongARM processor, and we're still only using 25 percent of the processor. That leaves a lot available for other applications.

We're also very excited with XScale, and we're planning to port the engine to the XScale as well. So let me know as soon as that's available.

RON SMITH: I sure will. That's even better news.

So tell me, how do the developers out there really find out about how they can use -- how Beatnik Audio Engine can help them?

JOHN ECKSTEIN: Sure. Well, great sound will go a long way for helping developers enhance their applications for wireless devices.

Traditionally, handheld devices have lacked the entire quality audio you would get on a PC. But the Beatnik Audio Engine has solved that problem. We have APIs for C, C++ and Java, and this will allow developers to create all kind of cool interactive games, Karaoke applications, interactive ring tones for mobile devices, for cell phones or other consumer appliances.

After today's presentation, developers can contact me for more information about Beatnik, or tomorrow afternoon I'll be making a presentation at IDF and would welcome to talk to anyone.

RON SMITH: Wow, that's great, John. Thanks a lot for coming out today. Take care.

JOHN ECKSTEIN: Ron, thanks very much, and good luck with your new music career.

RON SMITH: You bet.

Maybe you want to turn this off so I don't have to start my music career now. Thank you, John.

(Applause.)

JOHN ECKSTEIN: Thank you.

RON SMITH: Okay. So what you saw John demonstrate there was a basic decoding engine to enable this 32-bits of music. And originally they developed this on a PC, and the original non-optimized decode took 68 seconds.

By using the IPPs, they were able to get a 22x performance improvement. And not only that, they were able to port it directly to the iPAQ or the StrongARM-based platform in just three days.

So this gives you an idea of what the IPPs can do in terms of helping to improve developer productivity.

Our next building block is in the memory subsystem. Intel is a leading supplier of Flash memory technology. We're already shipping in volume at .18 micron technology.

We just introduced at the U.S. equivalent of the Developer Forum the Persistent Storage Manager for storage of large files, which is something that's going to be delivered with Microsoft's Windows CE and ultimately other operating systems as well.

We also have what we call the Flash Data Integrator which enables a very fast download of applications into the memory storage, and it also helps segregate between what part of the stored data is code -- for example, for the communications code in a wireless device -- and what part of it is for the end-user's data.

The next member of our family is in the communications subsystem. These are the baseband chipsets.

Intel is already a leading supplier of PDC baseband chipsets here in Japan for those that are purchased through independent suppliers. And we also announced a new signal processing architecture in December of the year 2000, and I'll get to that in a little bit.

But today I'm very happy to announce the newest member of our PDC family, the PDC Charm 2.0. This is based on a dual-rate baseband PDC standard. It handles beta applications and Java, and it supports the requirements and services of all the PDC carriers within the same single chip. So it's not segregated by a different carrier standard. It provides a capability for i-Mode and all the other packet data services in one chip. There is a press release on this for members of the press.

Now, the other new capability we've introduced is the Micro Signal Architecture. This is a very high performance and very low power wireless signal processing technology.

In addition, it provides very high efficiency in terms of code compilation from a high-level language. This was architected basically from scratch to enable this to be done.

So it has high performance, it has enhanced media instruction set similar to what you saw us demonstrate with XScale, it has dynamic power management which enables the power to be managed based on the load of the applications or load of the capability, and it's optimized for C and C++ programming.

And this architecture was jointly developed in a collaborative effort between Analog Devices and Intel.

So it's my pleasure today to introduce the leader of that effort, Dave Borland, who is our director for the Micro Signal Architecture engineering team in this collaborative effort. So Dave.

DAVID BORLAND: Good morning, Ron.

RON SMITH: Hi. Good to see you here.

DAVID BORLAND: Good to be here.

I wanted to give you an update on the new class of processor that ADI and Intel have been working on.

RON SMITH: Okay.

DAVID BORLAND: To establish a baseline for you, today in 2 G phones, DSPs run somewhere between 90 and 120 MHz.

RON SMITH: Right.

DAVID BORLAND: And with two and a half G coming out, people will be moving to 160 MHz as they go forward. So we wanted to design a processor that met that requirement. So let me have the demo introduce itself.

RON SMITH: See how you did.

>>: Hello from the ADI/Intel joint development in Austin, Texas. This voice is brought to you using a C compiled AMR decoder running in real time on the architectural test chip of the Micro Signal Architecture.

DAVID BORLAND: So what we have here is our first evaluation board of the Micro Signal Architecture running the VO coder {?} that's commonly used in cellular. And from the display you can see that we're running at 170 MHz.

RON SMITH: Wow, that's pretty impressive. 170 compared to a state of the art around 120 today. That's quite an achievement.

DAVID BORLAND: That is, that is. But we didn't want to stop there. We set our goals a little bit higher because we wanted to provide headroom for 3 G and beyond.

RON SMITH: Right.

DAVID BORLAND: Cellular operators are banking on a lot of data being delivered to these types of devices, and we wanted to make sure we increase the processing power to meet those demands.

So let me change some buttons here and see if we can get something a little bit higher.

>>: Hello from the ADI/Intel joint development in Austin, Texas. This voice is brought to you using a C compiled AMR decoder running in real time on the architectural test chip of the Micro Signal Architecture.

RON SMITH: Wow, 340 MHz. That's pretty impressive.

DAVID BORLAND: That is. We're seeing even better results than that. Let me go through some of the numbers with you.

RON SMITH: Okay. Why don't we take a look at the actual data you've actually measured on this device so far.

DAVID BORLAND: Okay. So if we could look at the foils. At 1.6 volts we're presently seeing greater than 400 MHz in performance. And through the dynamic power management that you mentioned, we can turn the voltage knob down to .7 and we're seeing excellent power. We're seeing .12 milliwatts per megaMAC.

RON SMITH: Wow, one architecture that extends from .12 milliwatts per megaMAC all the way up to 400 MHz?

DAVID BORLAND: That's right.

RON SMITH: That is impressive.

DAVID BORLAND: We believe this is a new class of processor. Not only is the speed and power excellent, but it's very easy to use, and you can truly program in C.

RON SMITH: Well, let's take a look.

DAVID BORLAND: So if we look at the AMR VO coder as an example, historically, that's had to be hand coded with assembly language because the compilers have been so inefficient.

A classical DSP, for example, if you directly compiled this application in C, you could see MIPS anywhere from 100 to over 600 MIPS of requirements versus hand optimized of 15.

With our architecture and our new tool set, in an afternoon you can compile this down to 42 MIPS. And with only one percent of the code converted to assembly, you can get 32 MIPS.

RON SMITH: Wow, that's at this impressive. And of course we have MIPS to spare with our new signal architecture -- or Micro Signal Architecture.

DAVID BORLAND: Absolutely. And with DSPs, they've been historically very efficient at signal processing, but also very hard to use.

We've added a lot of microcontroller type features to improve the ease of use, things like cache memory, 32-bit addressing, memory protection, user and supervisor mode, a number of features that really improve the usability of the device.

RON SMITH: So how would this compare in terms of actual time that it might take today to do the code development versus what it's taken with the traditional DSP approach.

DAVID BORLAND: Let's go back to our AMR example because the answer is it's dramatically different.

In our AMR example, on the old model it took 13 and a half months to hand code the assembly language for the AMR VO coder. In our new model, literally in an afternoon you compile it, and ten days later you can convert one percent of it to assembly to achieve very good performance.

RON SMITH: Wow, that's impressive. So what has taken over a year you can do in a matter of days.

DAVID BORLAND: Absolutely.

RON SMITH: And not only that, but it opens it up to a wide range of programmers that are C programmers as opposed to just the assembly language programmer.

DAVID BORLAND: That's right. New people can help now.

RON SMITH: Wow, this is impressive. I'm sure there's going to be a lot more new devices coming on the market very quickly when we bring this product to market.

DAVID BORLAND: Absolutely.

RON SMITH: Okay. Great job, Dave.

DAVID BORLAND: Thank you.

RON SMITH: Congratulations to you and your team. Excellent achievement.

DAVID BORLAND: Thanks, Ron.

RON SMITH: Thank you.

Okay. So you see what's in store in terms of the new signal processing kind of architecture in terms of speeding time to market and providing a lot of performance capability.

So if we put these all together and we look at what it requires in a basic wireless handheld device, we have a compute engine in our StrongARM processor, soon to be XScale Microarchitecture. We have our baseband chipsets which we have the capability today as we evidence with the PDCharm2, and we have coming this new signal processor architecture which will not only give higher performance at very low power capabilities but also enable very quick programming, very quick time to market, the thing which has really the limited the ability to get new hand sets out to market heretofore.

And last but not least, of course we continue to put a huge investment in our Flash memory technology, including software capability to also speed things like program download and storage.

So in summary, Craig told you when we started with the video that we're delivering four new architectures to enable the build-out of the evolving network, the IA-32, the Itanium processor family, the Internet Exchange Architecture, and the Personal Internet Client Architecture.

And the Intel PCA or Personal Internet Client Architecture will accelerate applications and services for these new classes of wireless handheld devices for the Internet which will help solve this dilemma of getting these new data services going very rapidly in the marketplace.

Today we demonstrated how our Integrated Performance Primitives helped speed application development. And in addition, we also showed how Intel is delivering the building blocks to help realize this vision of a ubiquitous wireless Internet, and we demonstrated that not only with the launch of our PDCharm2 baseband chipset, but we showed you yet a new innovation for the communications piece with our Micro Signal Architecture.

With that, I want to thank you very much for your time and attention, and I hope you enjoy the rest of the conference.

Thank you very much.

About Intel
Intel, 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.