Ga'ash Soffer, ooit werkzaam bij AnandTech (deed daar een aantal CPU artikel maar werd blijkbaar weggeschopt) heeft een artikeltje over de GeForce 256 in elkaar gedraaid. Hier heb je z'n conclusie:
Perhaps the biggest gripe I have about the announced GeForce is its failure to offer anything new over the TNT2 Ultra. The image quality improvements and geometry processing power which will probably never be fully used do not warrant a $250-$350 dollar price tag. When GeForce is released, it will drive TNT2 Ultra prices down even further, making the TNT2 Ultra a much more appealing buy. Furthermore, if S3 can provide decent drivers and a stable Savage 2000 chipset, the GeForce will not have a market at all. Early indications point to a ~$150-$200 Savage 2000 board price point. Combine low price with better fill-rate than GeForce and geometry processing, (which may not be as fast as that of the GeForce, but still not likely to be the limiting factor in performance, making the difference completely negligible) the Savage 2000 looks to be a hands down winner over the GeForce. Although Savage 2000 will probably be released a month or two later than the GeForce boards; even the most hardware crazed will not want to shell out $250 for a board which will probably be obsolete in two months. As it stands now, the GeForce looks to be a dismal failure due to it's extremely unbalanced design and very high price. Look for the TNT2 Ultra to be the chipset of choice until the release of the Savage 2000 and/or 3dfx's T-buffer packing "Napalm/ Rampage / Voodoo4 / some other name floating around". There simply won't be enough room for the pricey GeForce.
