Graphics acceleration

Thought your fancy video card was only good for gaming? Think again. Its graphics processing unit (GPU) is really like a second, highly specialized CPU. When it comes to certain kinds of complex math, its performance puts your desktop CPU to shame.

Until recently, all that power went to waste when you weren't chalking up frags. But computer scientists are finding novel ways to use GPU acceleration to speed up applications off-screen, as well. For example, a Stanford University project -- which uses many PCs around the world acting together as a supercomputer to assist protein-folding-related disease research -- can offload calculations to the GPU to multiply its performance many times.

Because the kind of calculations used to draw 3-D graphics are also applicable to many other problems, GPU acceleration is potentially useful for a wide variety of applications, from math-intensive science and engineering to complex database queries. Newer, even more complex chips -- such as nVidia's Aegia physics engine -- can do even more. No wonder nVidia has begun working on chips for the workstation market.

Increasingly, your PC's performance won't depend on the speed of any single chip. As Advanced Micro Devices and Intel get into the game, expect future desktop CPUs to incorporate CPU and GPU capabilities into a single, multicore package, bringing the best of both worlds to gamers and nongamers alike.

High-speed Net access

Where would we be without fast Internet access? It's easy to forget that just 10 years ago, most of us were still using ordinary modems. The broadband revolution ushered in streaming video, MP3 downloads, Internet phone calls and multiplayer online gaming. And we owe it all to TV.

In the 1980s, cable companies were promising 500 channels of round-the-clock programming. Cable was poised to become the most important wire into the house, but the telephone companies had an ace up their sleeve. A new technology could push high-frequency signals over ordinary phone lines, which previously had been good only for low-bandwidth voice calls. The telephone companies saw this as an opportunity to offer video on demand and to compete with the cable companies at their own game.

Or so they thought. The plans of the telcos for video on demand dried up by the mid-1990s, but the technology remained. Now called Digital Subscriber Line, it had morphed into a high-speed household on-ramp to the Internet. The cable companies followed suit with a comparable technology, and the broadband speed race -- for both DSL and cable -- began in earnest.

Both cable and DSL still use traditional frequency signaling over copper wires, but new breakthroughs are poised to go mainstream. Fiber to the premises (FTTP) promises lightning-fast network speeds, and WiMax will push broadband into territories that wires can't reach today. As for what applications this next broadband revolution will bring -- well, we have only begun to imagine.