About a year ago at WWDC ’08 (Apple’s annual World Wide Developers Conference) they announced something special; they were working on the next version of their operating system, called “Snow Leopard”. Without being specific, they briefly mentioned some of the new technologies they were developing which would tap into the power of the GPU (graphics processing unit), greatly enhancing Snow Leopard’s performance. Apple’s CEO Steve Jobs said it would be out in about a year (*wink* the upcoming WWDC ’09), and the company’s been tight lipped about it ever since.

Now I know what you’re thinking – just about every technology company out there says they’re going to deliver increases in performance. But I’ll get to what makes this announcement particularly important in just a moment.

Many computers contain both a CPU (central processing unit) running the vast majority of their operations, as well as a GPU inside a graphics card which accelerates anything graphics related. The CPU contains 1, 2, 4, and sometimes 8 cores capable of running several applications at once. Your GPU on the other hand, has dozens, if not hundreds of cores! The problem is, unless you’re doing something like playing a video game, all those GPU cores sit idle in your computer. Completely untapped processing power.

Granted, each GPU core isn’t as powerful as a main CPU core, but they don’t need to be for a lot your everyday computing tasks. Tasks that the CPU finds intensive, like editing/ rendering HD video for instance (which can take your computer hours to complete) are really just a lot of the same simple operations (algorithms) being performed over and over again, until the video is finished. This is where harnessing the potential of the GPU and its parallel processing abilities comes in handy. Instead of having 1 or 2 CPU cores running the same simple algorithm over and over again for hours (nearly dragging your computer to a halt), why not divide the task up and hand it over to the hundreds of GPU cores! The Mythbusters demonstrate an interesting way of showing how the CPU & GPU differ from one another in the embedded video below… its quite entertaining!

Nvidia, a GPU manufacturing company partnered with Apple, has already demonstrated the parallel computing capabilities of its graphics cards using an architecture called CUDA. Inexpensive Nvidia graphics cards coupled together, combining several hundreds, even thousands of cores, are able to reach near supercomputer performance in a single desktop unit. Researchers using Nvidia’s CUDA technology have reported being able to do things that used to take them several days, in just a matter of hours!

You wouldn’t normally find more than one or two graphics cards in today’s average computer, but the added performance of a few hundred cores from an off-the-shelf card should still offer substantial performance gains. Apple’s Snow Leopard is touting two new built in technologies, “Grand Central” and “OpenCL” which will utilize the Nvidia GPU’s inside Macs for general level computing, much the same way CUDA does. Who knew the parallel processing technology driving video games would some day surpass giants like Intel.

For Mac users, it’s going to offer a huge leap in performance. As someone who edits a lot of HD video, I can’t wait to be able to render videos in minutes rather than hours. When Snow Leopard finally hits the store shelves (which could still be a couple of months away), I think we’ll see Macs drawing in Windows PC users at an increased rate, or anyone for that matter who’s looking for more computing performance.

The keynote address kicks off WWDC ’09 next week on Monday, June 8th, where many expect to hear specific announcements about Snow Leopard, as well as the next iPhone & iPod Touch.