Intel Corp. is a company with a wide—and growing—range of products, from chips and chip sets for PCs, servers and mobile devices to software, communications offerings and now even blade servers. At the Intel Developer Forum 2003 this week in San Jose, Calif., executives talked about common technologies that reach across the various platforms. But one key constant is Pat Gelsinger, Intels chief technology officer, who admits that an important part of the job is “connecting the dots” and ensuring that the products do tie together. Gelsinger spoke with eWEEK Senior Editor Jeffrey Burt at IDF about Intels plans for moving their offerings forward within the overarching message of the convergence of communications and computing.
During his keynote speech, [Intel President and Chief Operating Officer] Paul Otellini spoke of a number of different products—from Itanium and Xeon chips to Centrino, and Hyper-Threading and multithreading—and brought them all under the umbrella of the convergence of communications and computing. Can you talk about how all these products fit under that umbrella?
Its really hard to give a simple answer to that. Its a pretty broad set of things, and some things are relevant to each other, and some things just arent. Let me try to give a few examples of how were trying to hold these things together.
First, theres a handful of key standards, industrywide standards, that were trying to make sure we deal with across a very broad set of our platforms. And 802.11 is a very good example of that. Its a key thing for mobile, its a key thing for communications, were going to be taking it in through our handsets as well. So were trying to take some of the key standards and say, “Intel has it across the board.” … Were trying to take individual technologies and stretch them across multiple platforms for the different applications.
Another good example is, you take MMX [multimedia] technology. We did that [in] 95, for the desktop when we first introduced that, and eight years later youre seeing it show up on the handheld platform. Some of the security stuff—LaGrande technology—that was first being targeted for desktops, but were taking it up to servers and down to handhelds as well, so were trying to tie these things together. … Thats one thing we do to make these technologies fit under a common umbrella.
Another thing we do is a lot of the software and tools. If you have an Intel Compiler version 7.0, that same compiler—if you load a few different modules—is exactly the same development environment whether its XScale, IA-32 or Itanium.
The final thing is … when we get to the next level, the user perspective, much of it has to do with delivering usage models across the multiple platforms. … Like the location service that was mentioned in [Vice President and General Manager of Intels Mobile Platform Group] Anand [Chandrasekers keynote]. … How do you do location services if you just have a compute platform? Good luck. But when you combine compute and communications, from both WAN and LAN, youre going to look at a few sets of location services no matter whether youre indoors, outdoors, whether youre on a small handheld or laptop or desktop computer being able to pull that together for enabling the converged usage models.
You can barely fathom the amount of energy that goes behind tying each one of these things together. How I might do a security implementation for a laptop ends up being quite different than for a handheld. However, I still want to be able to deliver a common set of services and capabilities so that an application vendor could say, “I really dont care. I just want to know that Ive authenticated that service for whatever communications device might be at the end of that network.”
How far are you along in reaching this goal?
Its kind of like climbing Mt. Fuji. How far do you have to go? Its a long way up that mountain yet. But Im very encouraged because we have some very, very defensible proof points now where weve actually gotten there. … I think were past the first stage because we have real, concrete milestones, but theres a long way to go on this path.
PCI Express appears to be one of those standards Intel is determined to run across multiple platforms. Can you talk about that?
PCI … is running out of gas. From a technologists perspective, you cant push it a lot further. PCI Express is this new serial interface technology that says that Im communicating to my nearest neighbor in more of a daisy-chain fashion and that allows us to go much, much faster. We expect that over the life of PCI Express, were going to exceed 10 giga-transfers per second from the initial 2½-giga-transfers-per-second rate. And we expect that this will be as profound, if not more so, than PCI was when we did that in the early 90s.
To put that in context, when we did PCI, we thought about it as a chip-to-chip interface technology, and then it grew up into a bus. In this case [with PCI Express], we see it as a bus as well as how we connect our chips together. But its not just for computing—for servers and desktops and mobile [devices]—but were making that a fundamental element of our communications platform as well. So we expect the breadth of PCI Express to have an even greater impact on a broader set of industries than even PCI was. And youll see every chip set, every CPU that we build for everything down from handset devices all the way up to big, honking Itanium servers, all utilizing a commonly architected interconnect technology in PCI Express.
You expect to see it take off by the end of next year?
You see the first products here [at IDF]. Theres a real chip set that will be going into production with it. … Those are early beta versions of the real products that are going into production, so youll start to see some of them showing up late this year. But next year is when youll really see the first wave of PCI Express products.
I expect that … in 05 theyll be the avalanche of PCI Express. Next year therell be a lot of early products, but there are still a lot of PCI-X things going on. But in 05, therell just be an avalanche of products across all segments.
Throughout the event this week, Intel has talked about Hyper-Threading, multithreading, dual-core processing in its chips. Why is this important to enterprises?
Theres two ways to consider that. … We started back in the 386 and 486 with what we called instruction-level parallelism, having the processor do more than one instruction at a time. Thats what pipelining and super-pipelining, super-scaler, out-of-order super-scaler pipelining, all of that kind of stuff was about. But youre still doing instruction-level parallelism—taking a thing and [asking] how can I make two, three, four, five instructions operate at once? Threading is the next abstraction.
What we see is were running out of gas with instruction-level parallelism. Theres still more there, but its becoming diminishingly cost-effective for me to get to the next level of parallelism at the instruction level. Thread-level parallelism is jumping up to the next paradigm. Rather than looking at single instructions, Im looking at basic blocks and how can I execute those basic blocks in parallel. So I start executing multiple threads. Microsoft put threading into the operating system starting back with NT, so thats gotten fairly mature. In the server supercomputer space, theyve been doing threading and threading tools for a couple of decades, so that software structure is fairly mature.
Now what we have is this great opportunity where I now have a big enough transistor budget—Ive got a 90- and 65-nanometer—I can really start putting not just little hooks for threading, which is what Hyper- Threading was. It took a very sophisticated, out-of-order core and said, “If I parallelize just a few resources … you fake the software into thinking that Ive got two threads of execution, and in lots of cases I got benefit from that, 20, 30 percent.”
What multicore says [is] not only can I do that, but I now have enough transistors [so] that I can start putting two cores in place. … Tulsa [a future dual-core Xeon chip] has two Hyper-Threaded cores, so from the software perspective, it looks like four threads of execution. From a software perspective, you start seeing a lot of threads being able to execute in parallel. Paul [Otellini] also said that were going to have that in both desktop and mobile products as well. … Just like weve had this rapid progression over the last decade in instruction-level parallelism, youre going to see the same thing with threads and multiple cores: Were going to have Hyper-Threading, were going to have multiple cores, were going to have speculative multiple cores, were going to have out-of-order, speculative multiple cores, just a very, very rich set of innovations happening for thread-level parallelism. This year we touched on it with Tanglewood [a future Itanium chip with more than two cores]. Two cores isnt the limit. The numbers going to be a lot bigger than that as we look forward.
This is the foundation of how were going to be building computers and microprocessors for the next decade, decade-and-a-half.
Intel has said that the growth of markets in Asia and Eastern Europe will impact what it builds, what it sells and at what price. Can you explain this?
The influence is occurring pretty much at every level, as you can imagine. What it is is that I just have more people there. I have more salespeople, I have more marketing people. Theyre in more factories, just being close to the local marketplace. With the products, were increasingly taking into account the issues of Asia. Some of those might be the standard issues there—802.11, is it the same in Asia as its going to be in other parts of the world? … Spectrum allocation might be different in other parts of the world, so you just have to start comprehending a much more global set of considerations as you go into communications and as you focus on Asia.
We have [people] that were putting on the ground to understand the differences in how computers are utilized. We have this very, very naïve look from the U.S. that [computers sold in Asia] are just cheaper. And that is not at all the case. In many cases they want superior technology. This is an asset that they might be buying. A PC in China is a family asset that is an investment in the modernization, the education and the future success of that family. As such, youll see uncles and aunts and grandparents pooling their money to go buy a PC. They dont want low-end, they want a good product. As such, we have great Pentium and Pentium 4 sales in China.
Its very counterintuitive, but they do look at things very differently. When you look at a computer in China, they turn it off and cover it up when theyre done. … Thats how theyre culturally comfortable with keeping that asset safe and secure. There are language issues and social issues in these different marketplaces, so we have to consider those much more in how we think about and design our products. So its really affecting us in how we run our company to how we sell our products as well as how we design and look at them for the future.
How does that manifest itself in what customers in the U.S. see?
For the most part, weve been able to make those differences transparent to what somebody would see here. In some cases, well try to bring technology in the U.S. more quickly, such as cell phones. I was always bothered that I could go to Japan and get a 6-ounce cell phone with a beautiful LCD display, and Id come home and use my little dinky black-and-white display that was twice the weight.
In other cases its been accelerating some of the things that weve tried to do overseas so we have the best of both worlds. In the consumer electronic sense, much of that innovation has and continues to be done out of Asia, and its largely … bringing that back into the U.S. When we think about things like the digital home, we see that as really a bridge between the best of the PC and IT industry and the best of the consumer electronics industry.
Shrinking the Process
Intel will soon becoming out with Prescott and Dothan, both built on 90-nanometer technology, and you have an aggressive road map for continuing to shrink your manufacturing process, down to 22 nanometers in 2011. How will this impact products that customers will be using?
There are two perspectives on it. One is that we have been on this two-year heartbeat of process introduction for almost a decade and we do not see that changing. In fact, we see our ability to do that is becoming more precise and scientific as we look forward to 65, 45, 32, etc. When you think about that, you have the capital costs associated with that, the engineering investments to do so. We think that becomes one of our unique competitive advantages. We were clearly in front of the 130[-nanometer process], were well in front in the 90 nanometer, and we think that will be the case for the technologies in the future.
What we are working to do as a company is line up leading-edge process and manufacturing technology. That is the sum of what we do. Now theres a lot of innovations we do architecturally, with new interfaces, with new capabilities, etc., that are leading-edge products that utilize that new process of manufacturing technology. When we line up those two things together, were golden.
From the industrys perspective, what … has been the unwritten insurance policy for the entire technology industry has been the simplistic view of Moores Law—you get more for less. And that just continues forward at reasonably predictable rates. You get to decide how you utilize Moores Law. Its not like Intel has a proprietary lock on Moores Law. It works for everybody, but just says, “Hey, were going to be able to put more transistors in place at the same cost, at the same power budget, at the same size. I can make things smaller, I can make them faster, I can make them more integrated, I can make them lower cost, I get to play Moores Law whichever way my business or product interests work. But that becomes this underlying premise that the entire industry can build upon.
If we stood up tomorrow and said, “Moores Law is not going to hold, its slowing down to once every three or four years,” that would change the dynamics of every aspect of the industry. The foundry guys, that changes their model. The equipment providers, that changes their model. The software guys, that changes their model. But were saying with such bold confidence, “Were on track. The next decade, weve got it figured.” Its not going to be easy. Theres a lot of innovations and a lot of problems we have to overcome, but we put that in place so that the industry can say, “OK, great, I can operate a certain way, make certain predictions, build certain business models [and] product plans around that.”
Thats not just critical for our business, but its the critical underpinning for the entire IT industry, and increasingly becomes the underpinning for the industry more broadly. Its not just the IT industry anymore, but what are the interesting innovations that are happening in transportation? They utilize silicon. What are the interesting things happening in medicine? Genomics and all of those things [are] based on computing. The interesting things that are happening in the geological sciences. How are people doing better mining or better exploration? Its all based on computers. More and more, the other industries that are built around computing become dependent on that same heartbeat.