With Moore's law -- the law that says the number of transistors on a given size of silicon can double every 18-24 months (often while the price continues to drop) -- just having reached its 40th anniversary and with Intel saying that it's good for at least another 20 years (with prototypes to back that claim up) if not longer, the company has clearly hit its stride when it comes to chips and is now turning to other areas of innovation such as security, partitioning, and management in order to find its edge over company's like AMD that are paying attention to the very same issues.
Leading the charge on many of those alternative fronts is Intel's vice president of technology programs Frank Spindler. Last month, while visiting the company's Santa Clara, CA-based campus, I sat down with Spindler to discuss those initiatives -- everything from multicore chips for handhelds (yes, for handhelds) to what a chip company can do to help with desktop security to whether robots will ever be smarter than us to the point that the Will Smith movie I, Robot could be a reality. In the interview, which is available as both an MP3 download and as podcast that you can have downloaded to your system and/or MP3 player automatically (see ZDNet's podcasts: How to tune in), Spindler offers some of the details on these initiatives -- for example, when and why the company's LeGrande Technology is tied to the availability of the next Microsoft desktop operating system (Longhorn).
Here are some highlights from the interview:
Spindler on how future systems will scale: Instead of chips per computer, it's computers per chip. Intel has 10 programs going on today that involve multiple processors on a single piece of silicon -- spanning servers, desktops, and mobiles.
Why we need more peformance in handhelds: To go explore the world or Web and recognize certain types of information in that huge sea of data and then mine it. For example, to recognize certain patterns of face...maybe I'm looking for images of my daughter when she was 5 years old. Then there's synthesis of environments or creating scenarios or situations that might be of interest...The sort of things, like the Genome Project, that are done on massive systems... the ability to do those sorts of very complex operations will be at your fingertips.
Regarding the arrival of Intel's virtualization technology (Vanderpool): It will be here this year. There will be some solutions for it, but it may be dormant on some systems.
Spindler on whether Intel's LeGrande desktop security technology solves a problem that Microsoft can't solve on its own: We like to think we are solving problems for users and they're the ones concerned about security.
How low can the silicon go (in thickness): We have built prototype transistors that are 10 nanometers wide, which is around 10-20 atoms of thickness.
There's more to the blue screen of death than meets the eye: Intel's Active Management Technology will be able take a system's vital signs even after the system appears to be frozen and inaccesible.
Branching out is nothing new for Intel. It's been doing it for years in the areas of desktop management and wireless. But the renewed emphasis on how else the company can add value to its chips comes at a time when, at least in terms of buzz and excitement (perhaps not in terms of shipments), its biggest nemesis--AMD--appears to have something of an edge.