When Intel looks to the next few years of computing, the dollar signs look a lot better in three dimensions, full motion and high definition.
The rise of what Intel calls the 3-D Web will transform the way people use the internet, offering a far richer experience, exemplified today by Second Life but turning more realistic and being deployed in medicine and business, Intel chief technology officer Justin Rattner said in a last-day keynote address at the Intel Developer Forum in San Francisco last month. The best news for the chip giant, however, is how much computing power all this will take.
"This may be the killer app of killer apps," Rattner told reporters after they keynote. "As people demand a more immersive, a more realistic experience, they're just going to push the computing demands to unprecedented levels."
The pressure on central and graphics processors has already soared as internet use expands from traditional web pages to applications such as Google Maps, Google Earth, multiplayer online role-playing games and Second Life. Going from surfing the traditional web to traversing Second Life can increase CPU load by three times and graphics processing load by 20 times, he said.
Servers need to work harder, too. A server that could support 2,500 concurrent players of the online game World of Warcraft, for example, could only support 160 Second Life participants, according to Rattner.
His keynote was colourful and entertaining for an Intel Developer Forum presentation, full of animation and simulations, but he outlined how much further this richer experience has to go. For example, the kind of realism people expect from movies requires shadows and reflections that move with objects in a scene, he said. Daniel Pohl, a research scientist at Intel, demonstrated high-definition environments in the game Quake that included those elements and others. They demanded nearly all the processing power of a server with dual quad-core Intel x5365 chips.
In medicine, doctors are aiming to allow surgery on virtual patients. With current technology, they can simulate cutting a rectangular patch of skin and opening it up in real time, but it would take far more processing power to do the same with a curved piece of skin or something like a cleft palate, according to Joseph Teran, an assistant professor at University of California, Los Angeles, who participated in the keynote.
Other challenges for making the 3-D Web a reality include user interface and standardisation, Rattner said. Tasks that take computer artists hours, such as creating shadows or reflections, need to be automated so that the average person can create user-generated content in an elaborate online world. Humans also need a better interface with the online world, Rattner said. He demonstrated the SpaceNavigator from Logitech's 3DConnexion division, which can be pushed and pulled to interact with a virtual environment, and the Novint Technologies Falcon, a three-dimensional game controller with force feedback. In addition, the mostly proprietary world of virtual worlds needs to embrace common standards so that, for example, avatars from one world can enter another, Rattner said.