Introduction

Founded in 1993, NVIDIA has become the indisputable king manufacturer of 3D processors with best seller products like the acclaimed TNT & GeForce lines. Every hardcore gamer dreams to own a GeForce powered card because of their undeniable qualities that combine performance and reliability. Over the years NVIDIA has acquired unprecedented expertise in the 3D domain that brought our computer the ability to render, in real time, complex 3D scenes and characters with the most exhaustive realism. In fact, during the last six years, no other PC component has evolved as fast as the graphics subsystem. Today’s GPUs offer a performance increase of nearly 100x over the first 3D chips. All this innovation has been partly led by NVIDIA and its competitors. Economically speaking, NVIDIA was in 2001 the best stock value of the NASDAQ.  

Every six months or so, NVIDIA gratifies us with a brand new 3D graphics processor unit. Not only do new NVIDIA GPUs offer better performance but they usually bring unrivalled and innovative hardware features that are generally quickly adopted by game developers. According to Luciano Alibrandi, NVIDIA’s European Technical Marketing Manager, the main goal NVIDIA pursues is to make PC games as real as the blockbuster movie Jurassic Park. 

A few months after releasing the GeForce 3 Ti 500, NVIDIA officially unveiled their latest baby, the GeForce 4 on February 6^th. Introduced in brass band with all the bells and whistles you would expect for such a product, this new chip is simply the most powerful GPU NVIDIA has ever produced as well as the most powerful GPU available on the market as we’ll see in this review. Code-named NV25, the GeForce 4 is a totally revamped GPU, that is to say its architecture is radically different from the GeForce 3 despite it uses some of its predecessor’ specifications. 

With only one major competitor on the market it was time for NVIDIA to release a new GPU, since ATI has managed to catch up with the technology used by the GeForce 3 with their Radeon 8500 family. Not only is the new GeForce 4 Ti 4600 faster than ATI’s GPU, but it also promises to crank up the reality level in upcoming games with new impressive effects that will help to deliver lifelike motion and more realistic scenes.

First Look

For this test we received a NVIDIA engineering sample of the GeForce 4 Ti 4600. The first thing that jumps to the eyes is the uncommon size of this AGP card, a whopping 8.5 inches long. The second thing you’ll notice is the brand new stylish and streamlined NVIDIA branded heatsink that covers the CPU. This new heatsink covered by a fan has been designed to maximize the efficiency of the cool down process. Surprisingly there’s no heatsink on the memory chips of the card. NVIDIA chose to use FBGA packaged memory chips that have the incredible property to remain almost cold even after intensive use. I’ve checked this out and I can guarantee you that after 5 hours of intensive benchmarking the memory chips were almost cold. The card we got was hosting Samsung memory chips that are 2.8v powered and could support a maximum frequency of 350MHz giving some room for overclocking. 

In terms of connections the card offers three external connectors: one DVI-I plug to connect a digital LCD flat panel, one standard analogue VGA output and one TV output. The TV output cannot receive a Personal Cinema unit. However manufacturers can add on the GeForce 4 Ti 4600, the chip needed to support the Personal Cinema unit;  unfortunately it’s not natively supported. The DVI-I output is managed by a Silicon Image chip while the TV output is managed by the usual Conexant CX25871 chip. The key detail is that you can now output the video signal on two monitors at the same time, à la Matrox ‘DualHead’.

NVIDIA GeForce 4 Ti 4600