14 DAY TRIAL // There are two very important things about NVIDIA's newly launched Fermi-based GeForce GTX 470 and GTX 480 graphics cards with DirectX 11 support. One is the fact that they recovered NVIDIA's crown as owner of world's fastest graphics. The other is that, for all intents and purposes, they are still completely unavailable. Now, however, thoughts have turned not towards this duo's impending availability (April 12) but towards the Santa Clara, California-based company's future plans, including its roadmap for bringing Fermi to the mainstream.
It is already known that the GF100 graphics processing unit (GPU) is too large, hot, strong and expensive to be used in video controllers meant for the mainstream and entry-level markets. As such, NVIDIA will naturally be making a toned-down chip, currently known as the GF104.
So far, however, very few things have been found out, which is not surprising, considering how long it took the web to land leaks about the GTX 470/480, even until a couple of weeks ahead of the release. The most recent report published by Fudzilla is, regrettably, no more revelatory than all others before it, though it does confirm the launch schedule.
The first GF104-based graphics cards will come out in summer, probably in June. The report's exact details are that, “The new card is codenamed GF104 and it is currently taped out, works and performance looks good.”
This means that, instead of being late for ages and failing to make an appearance even after being called on stage (as it was, and is, the case with the existing GTX 400), the mainstream adapter might actually stick to the schedule. GF104 is much smaller than the GF100 and has 64 TMUs, 256 shader cores, 32 ROPs, a 256-bit memory bus and will supposedly take the place of the GeForce 8800GT, later known as the 9800GT and, afterwards, GTS250.
End-users might remember the 8800GT as one, if not the best-selling NVIDIA adapter priced at under $200. If the GF104 should manage to stick to this price point and come out in the allotted time, it might actually prove quite successful. Consumers can only hope.