The passed half of the year 2004 established a new landmark in the development of consumer 3D graphics as the two graphics giants – NVIDIA Corporation and ATI Technologies – both unveiled their new generations of graphics processors. The birth of a new architecture is always an Outstanding Event; long before it really happens, the community starts boiling with variegated and contradicting rumors, grounded on the scraps of information from the companies themselves as well as from mass media.
That was roughly the beginning of 2004: the press was engaged into a hot discussion on the characteristics of the upcoming solutions from NVIDIA and ATI, although no one seemed to have absolutely trustworthy information. We knew that the new processors would both work with power-frugal and high-clocked GDDR3 memory; NVIDIA’s NV40 was also certain to have 16 pixel pipelines against 8 or 12 or 16 – who knows? – pipelines in ATI’s R420. The frequencies were indefinite, too, something in a range of 500-600MHz. As usual, the truth became apparent only after we saw the real silicon.
So, let’s try to evaluate the progress of the developers of graphics processors from our traditional point of view: are these GPUs good for us as end-users? As usually, we tend to bring you the most detailed picture of a massive amount of benchmarks, today we have 35 of them!