Why powerful NVIDIA, which seemed to be heading in recent years safely, gave the performance crown to ATI? One theory says that at fault excess focuesd at CUDA technology by group from Santa Clara,

Recall that CUDA allows you to run applications on a specially prepared graphic processor unit (GPU), which has several, and sometimes even more increase in speed compared to the same applications running on the main processor unit (CPU). This technology is supported on cards GeForce 8 and above, and use it with programs such as computing, video Battery, or recently released Adobe CS4 package, addressing, inter alia, for professional graphics processing and assembly of non-linear films.

On the one hand, work on the development of CUDA is a good move, giving NVIDIA ability to enter into new and professional markets. On the other hand, optimization of the GPU API firmly complicated construction graphics processor, which resulted in enormous problems with the transition to 55 nm process technology. Secondly, dedicated to developing CUDA forces and measures have not been spent on essential functions from the point of view, yet critical audience NVIDIA products – players, enthusiasts and ordinary members are waiting for good performance relative to the price. In turn, ATI, preparing a new series of Radeon HD 4000 knows first and foremost a 3D fast low price, neglecting to optimize the speed of calculation in the programs that run on graphics card. As a result of a series of GeForce cards as soon count the structure of proteins in the Folding @ Home and accelerating action CS4, but it ATI is cheaper and more efficient in games of cards, serving in Appendix DirectX 10.1.

Whose strategy will prove effective in the long term? Experience over the next months. At the event, on the situation better out ATI and NVIDIA collect on the black clouds. Most likely, this configuration will remain the same at least over the next six months.