Background: The History That Shaped Tomorrow
Without going through in detail the boring archives of yesterday’s activities between 3Dfx and NVIDIA, we should at least recall the bitter rivalry and the paths taken by two of the most prominent 3D chip makers in the industry. Here’s the flurry of three rapid years of development summarized in a page:
Before the Riva TNT from NVIDIA, snazzy 3D graphics only belonged to adopters of 3Dfx’s acclaimed Voodoo and Voodoo 2 lines of cards. Still it took a beefed-up Riva TNT2 to bring NVIDIA to center stage alongside 3Dfx. It was then when 3Dfx lauded the superior performance of the lethal Voodoo2 SLI (Scan Line Interleave) combination while NVIDIA insisted 32-bit quality to be of prime importance.
The next duel became a major turning point. NVIDIA’s engineers worked like paranoid ants, keeping to their philosophy of a 6-month product cycle, and worrying the living daylights out of the competition forced to play catch up. This rigid and well-executed strategy worked very well for NVIDIA and helped catapult the firm to the forefront.
While the world waited patiently for 3dfx and NVIDIA to release the barrages of attacks on each other, verbal conflicts ensued in chat rooms between die-hard followers. However, amidst great anticipation, the Voodoo 3 just did not bring enough to the table and it was easily matched by NVIDIA’s TNT2 Ultra.
We heard 3dfx preach the same technology, albeit enhanced, to boast SLI-like performance on a single 2D/3D accelerator, banging on the existing base of GLIDE specific game titles as the last straw. Whew! NVIDIA must have Microsoft to thank for the increasing usability of DirectX, which made GLIDE somewhat irrelevant.
Very quickly and before 3dfx could even counter the TNT2 Ultra threat, NVIDIA missed no opportunity to bring out their GeForce card, declaring it a revolutionary card never before seen in the world of PC gaming. The card featured an integrated Transform and Lighting engine and was attached with the marketing savvy ‘Graphics Processing Unit’ GPU acronym. All these went well with the audiences and the ‘Voodoo Magic’ dwindled ever since. The GeForce outperformed the fastest Voodoo 3, and ‘outnumbered’ the Voodoo 3 in phenomenal fill-rates as well.
Then about exactly a year back, having missed a product cycle, we saw 3dfx launch the behemoth Voodoo 4/5 cards, in a final attempt to recover its lost glories. Armed with a chip designed to perform full-scene anti-aliasing (FSAA) inherently, the Voodoo 5 5500 represented the last fizzle from the once great technological company. The card was huge, sported two chips working in tandem, performed reasonably in most tests, and provided ‘okay’ frame rates doing FSAA at certain resolutions. This was when the company decided to sell the pitch that “quality is king!”.
Just about the same time, NVIDIA dealt 3dfx the critical blow with the release of the polished and very very quick GeForce 2 GTS, where GTS stands for GigaTexel Shader. Though some of the diligent work NVIDIA’s team put in could not be immediately appreciated, the card delivered more than double the fill rate of the earlier GeForce. There, the battle looked somewhat lopsided with the GeForce 2 GTS demonstrating faster frame rates in the most demanding games than the Voodoo 5, and the jazzy technology demos that utilize the 2nd generation T&L engine, stole the show from the FSAA effects.
Game playability takes precedence over all fancy add-ons and depending on the genre of games, individuals will look for different features as priority. A flight-sim buff will not require 60 fps, but anti-aliased enemy fighters would probably make him happier. But isn’t that a hard sell to a niche group? Sports and racing types of games benefit somewhat from the AA filters, but playing such games at higher resolutions does help minimize noticeable polygon jaggedness. The associated performance hit makes people think twice about wanting AA filtered games, and mass acceptance will only occur when it becomes a ‘free feature’.
After the Voodoo 5 5500 product, the great 3dfx was no more. NVIDIA overran it. That was how the epic duel came to an end, leaving only three other dominant players such as ATI, Matrox and STMicroelectronics in the field of consumer 3D graphics. It is going to take them a lot of innovation and a well-thought out strategy to hold back a very aggressive NVIDIA. The firm not only proved to be able to clinch on to the performance crown, but having flooded the market with diverse product lines, it is also steadily eroding the competition’s marketshare in the OEM, dual-monitor, budget, mobile and even the console markets. There is no stopping ambitious NVIDIA from competing intently in every area they are capable of.
So fast forwarding to today, after releasing a bout of incremental chips and dominating the high-end with the very speedy GeForce 2 ULTRA, NVIDIA, again, shifts into high gear with the announcement of the GeForce 3 product. In a self-imposed obsolescence roadmap and with virtually no competition in sight, the company has recently been its only enemy.
To some, it is spearheading the direction of technology adoption amongst game developers. To others, it is bringing out too much ahead of its time, with products that commanded a steep price and loaded with features so advanced, that virtually no current game titles take advantage of.