Fall 1999: GeForce 256

May the Force be with you

In August 1999, NVIDIA made its most significant announcement with the launch of the GeForce 256 at the Intel Developer Forum in Palm Springs. Originally codenamed NV10, the GeForce 256 brought a number of revolution features to PC 3D graphics.

Just as the RIVA 128 had been one of the first graphics chips to offer a full triangle setup engine, the GeForce 256 was the first shipping implementation of on-board transformation and lighting. Incorporating a new processor with four-pixel pipelines at 120MHz, the chip offered a theoretical maximum of 480Mpixels/sec fill rate.

The chip also supported DDR and SDR SDRAM. With DDR memory offering twice the bandwidth of conventional solutions, the GeForce 256 was the hands-down performance leader. The GeForce 256 also brought a wealth of new image quality features such as cube-environment mapping and Dot-3 bump mapping.

GeForce 256

Furthermore, the GF256 was NVIDIA's first graphics chip to implement advanced video acceleration features such as HDTV compliant motion compensation and hardware subpicture alpha-blending. In November, NVIDIA released the Quadro, a high-end professional workstation chip based upon the GeForce 256.

Six months later...

When GDC 2000 came around in March, no one was expecting NVIDIA to announce a new product. The S3 Savage 2000 and ATI Rage Fury MAXX were only able to compete with the SDR version of the GeForce 256, and the card's main competitor, the 3dfx Voodoo5 was still nowhere to be seen.

The Xbox

NVIDIA Xbox demo at Comdex

NVIDIA took this opportunity to announce that its technology would be used in the Microsoft Xbox. For NVIDIA, the Xbox project is a return to the company's original console roots. In fact, NVIDIA not only has the design win for the console's 3D graphics technology, but also the multimedia subsystem. While the success of Xbox still hasn't been determined, NVIDIA can already be declared a winner. For the development of the Xbox GPU, NVIDIA has received a $200 million advance from Microsoft along with the marketing clout that comes with such a high profile project. As long as NVIDIA can come through with the chip design, it's a win or win-big situation.

One month after the Xbox announcement, NVIDIA launched the NV15, better known as the GeForce2 GTS.

GeForce2 GTS

NVIDIA's six-month product cycle has been described as Fall Dad's, and Spring Grad's. New platforms are to be released in the fall, with a new chip revision offering minor improvements and a higher clock speed made available in the spring. The GeForce2 GTS was more than just a simple "spring refresh." The GTS almost doubled the pixel fill rate of the GF256 and nearly quadrupled the texel fill rate by dramatically increasing the clock speed and adding multi-texturing to each pixel pipeline. The GeForce2 GTS also added S3TC, FSAA and improved MPEG-2 motion compensation.

GeForce2 GTS

This may sound like a lot, but the improvements in the GeForce 2 GTS did not seem like overkill. ATI engineers up north in Canada had just recently announced their own wonder-chip, the ATI Radeon. Although the Rage 128 Pro, ATI's previous chip, barely reached the performance level of the vanilla TNT2, the Radeon was supposed to be an equal to the GeForce2 GTS, offering similar performance and additional quality features such as environment mapped bump mapping, DirectX8 features, and a more impressive Transformation, Clipping, and Lighting unit.