Los Angeles (CA) - Up until this year, the enthusiast gamer has been, by definition, a PC user - or, in deference to Apple, a "computer user." This year, thanks in large part to Sony, that's changing. From a marketing perspective, the enthusiast is the person who is more willing to spend money and adopt early. The money an early adopter might end up investing in a high-end, PlayStation 3-based gaming console is essentially what he might spend on at least a high-end computer, if not necessarily the top of the line.
For that matter, the PlayStation 3 is essentially a computer, albeit without a keyboard and mouse. But it does, after all, run Linux, as do a growing number of Intel-based PCs. So the comparison is inevitable: Of the two platforms - the PC and the PS3 - which is the better investment from purely a gaming perspective? Intel, of course, has a tremendously biased opinion on this subject. But in the light of more substantive competition than it has ever faced before, will Intel's perennial arguments continue to hold up? Or will subtle changes to those arguments shed some light as to where Intel wants to go with its push into gaming?
The company rented out a conference room to itself at this year's E3 Expo, where it showed off Core 2 Extreme-based desktop and mobile PCs that were clearly cranking out pixels at faster speeds than their predecessors - according to Intel, about 40% faster. Humphrey Cheung and Scott Fulton of TG Daily spoke with David Tuhy, Intel's general manager for its desktop products division; and Jodi Geniesse, the company's communications manager for mobile marketing, about some of the underlying issues governing Intel's new performance message - including lower power - and how it plays out to a new breed of customer who may be moving up from a generation not of PCs, but of consoles.
TG Daily: There are two top-of-the-line processors that are being tested here at E3 for performance. One of them is the Cell. The other one is in this room. Brass-tacks-wise, from the gamer's perspective, why is the Core 2 Extreme better?
David Tuhy: You almost have to look at it from a system level...Cell, which is an IBM and Sony collaboration, is a very, very dedicated processor specifically to their console frame, highly optimized for the pixel processing. Obviously, we're putting our thing into a PC, which is an open environment, and we're trying to advance on standard operating systems, to run games on top of this. So we have to worry about a lot more...
Intel's general manager for desktop platforms, David Tuhy, showing off one example of a portable desktop Viiv PC.
In the Cell environment, it's a Cell processing engine plus a third-party graphics vendor. It's highly optimized to that workload for giving the gaming experience to the console. I think they've done a great job; they've actually got a three-operand flow through their architecture, we actually have two. They can do multiply/accumulates based on their architecture, where we do our multiply through SSE [Streaming SIMD Extensions], and on our Core 2 Extreme, we double the SSE performance. It's actually got a true 128-bit SSE engine, which we market as "Advanced Media Processing."
The PC version of that is a CPU plus a graphics card, sometimes two separate, sometimes more. We have a system out here which has our top-of-the-line microprocessor, the Core 2 (the combined brand for the next-gen CPUs formerly code-named "Conroe" and "Merom"), with a quad Nvidia system for graphics. It's a different price point: It's going after the enthusiast gamers, but it can draw a heck of a lot of polygons. [I don't know off the top of my head] the number of polygons it can draw versus a Cell, but I think it's going to be higher, because there's a lot more bandwidth on the quad system than on the Cell system.
TG Daily: Up to this point, for the last 15 years, the argument in favor of PC gaming has been, "Hey, we have the performance lead. We can put gamers into the experience." Surely, Intel is reliant upon second parties like Nvidia and ATI to help you out getting [polygons drawn], but even that's still part of the Intel architecture. How does that improve and evolve with Core 2 Extreme?
DT: Specifically, we have taken a very different approach with Core than we have taken with any other architecture before on the desktop. We're leveraging what we've learned with the mobile architecture, that we learned back with Centrino, and we improved that in the desktop and the server architecture with 64 bits, virtualization technology, floating-point engines - we bill them together as "Core Engine." So for us, Core 2 is the biggest move forward in terms of an architecture advance that we've had in the last five years...in terms of net gain, both in terms of its performance - which is averaging 40% more in desktop - and also, at the same time, lower power. Usually, that's not what happens for us. Generally, when our performance goes up, so goes the power. We went a different way with a very power-optimized architecture.
The second thing we did is, we didn't require new instructions. We chose to optimize the current instruction set and give it more resources, so the SSE engine is a true 128-bit SSE engine, the cache now has advanced dynamics, it's a four-wide, out-of-order [instruction queue] machine. So several things we've done in the architecture where you can basically take an existing game and it just runs faster. Which is great, because it can plug and play with existing stuff.
