Nvidia unveils new GeForce 4 graphics

[J.A.F._Doorhof]

At an invitation only event held last night at the Palace of Fine Arts in San Francisco, Calif., NVIDIA Corp. finally unveiled their long-awaited line of GeForce4 graphics processing units (GPUs). The product line consists of three main components, the GeForce4 Titanium, GeForce4 MX and GeForce4 Go. MacCentral was there and has the details.

  
Mac users have already had their consciousness raised about the new processors thanks to recent announcements from Apple. Last week Apple announced the mainstream GeForce4 MX component would be on cards included with the new midrange and high-end dual processor Power Mac G4 models; yesterday the company briefly announced the GeForce4 Ti would also be available as a build-to-order option and later as a standalone upgrade products for existing "QuickSilver" Power Mac G4 systems. (Apple today re-confirmed the news.)

NVIDIA president and CEO Jen-Hsun Huang kicked off the presentat! ion by recapping some of his company's recent news. NVIDIA recently shipped its 100 millionth GPU -- a feat it accomplished in less than five years. He boldly suggested that NVIDIA would ship its billionth GPU in an even shorter time. He also indicated that NVIDIA is now the largest fabless semiconductor company in the world.

In introducing the new GeForce4 technology, Huang explained that NVIDIA's mission is to narrow the gap between what is possible with computer-generated graphics used in cinema -- movies like Pixar's Monsters Inc. and Square's Final Fantasy: The Spirits Within -- and interactive graphics that users of home computers and workstations can experience.

In a question and answer session following the presentation, Huang noted Apple's early adoption of the technology. Apple engineers are working very fast on implementing the features of the new technology in their own ha! rdware and operating system software, he said. Huang singled ou! t Jon Rubinstein, Apple's senior vice president of hardware for special credit. "He beat everyone to market, including us," joked Huang. "It speaks to the excellent engineering talent at Apple."

The GeForce4 line is actually comprised of five separate chips designed for desktop applications and three different chips intended for mobile configurations. Two models of GeForce4 Ti chips -- the 4600 and the 4400 -- provide different degrees of high-end performance suited for enthusiasts and professionals respectively. Three different models of GeForce4 MX chips have been developed to cater to the mainstream computer user and computer manufacturer; each offers a different range of performance and capabilities. Finally, the mobile GeForce4 Go is offered in three different configurations as well.

All of the GeForce4 desktop systems sport key features emphasized by NVIDIA such as Lightspeed Memory Architecture II, which NVIDIA senior director of GPU Business Tony ! Tamasi said delivers double the effective memory bandwidth -- and therefore much higher performance -- as before. The technology works by using an efficient compression technology and second-generation "Z-occlusion culling," which works by not rendering pixels that are hidden by other 3D objects in a scene.

Accuview Antialiasing Technology is another key feature. Tamasi said user polls showed that 60 percent of users demanded sharper, clearer pictures without jagged lines. With GeForce4, Tamasi said NVIDIA was now able to deliver that without creating any major performance penalties.

The GeForce4 Ti series specifically sports the nfiniteFX II engine, a programmable engine designed to let game makers create new levels of detail in their character animation and environments. The technology is up to three times faster as the previous highwater mark achieved by GeForce3. Tamasi demonstrated the technology's support of "z-correct bump mapping," which enables g! ame developers and 3D application makers to display much more r! ealistic levels of detail and natural texturing than has been available previously.

Of specific relevance to PC users but not Mac users is the GeForce4's support of "nView," a new flexible and controllable display technology that makes it possible for PC users to manage multiple monitors from a single card. Similar technology has existed on the Mac for a while now, but it's nice to see Windows users catch up.

The new chip family also features NVIDIA's Video Processing Engine (VPE), which provides integrated hardware MPEG-2 decoding and a variety of other features suited for DVD playback and video signal decoding.

A variety of OEM manufacturers attended the event, including Apple. Apple showed off NVIDIA's new "Wolfman" demo, which demonstrates some of the advanced features of the nfiniteFX II engine. The Apple station used a dual-processor 1GHz G4 equipped system attached to an Apple 22-inch Cinema Display, using a GeForce4 Ti-based card.

!
Apple expects to deliver the GeForce4 Ti next month.