Soon on the market will appear the first motherboard which uses lucid hydra 200 chip, which allow not only to effectively scale the productivity of similar video card, but also to unite in one system video card from different producers.
NVIDIA chapter , Mr. Jen-Hsun huang, for the first time expressed its relation to lucid hydra 200 chip. According to him, this is a terrible idea. Drivers, applications and their optimization under SLI and CrossFire - without this multi-chip graphical solutions hardly can work, as considers NVIDIA leader . Let us recall that Lucid hydra rests on the hardware realization of video card acceleration , no drivers for the work of the corresponding chip will be required. Mr. Juan adds, that SLI and CrossFire is widely accessible on the market - in this on what their competitive advantage consists.
In the same interview NVIDIA chapter reported, that its company deals with the development of program solutions, but earns from the sale of hardware solutions. Here we deal with the platforms CUDA and PhysX, which NVIDIA moves without the special material benefit for itself, but compensate the expenditures with video card sales, which support CUDA technology.
With respect to plans in the chipset sector for Intel platform NVIDIA again state that to release products with DMI support the company does not have the possibility, but it will continue the release of chipset with FSB support for long. For atom platform NVIDIA company prepares something special and uncommon.
Mr.Juan also reported about wrong rumors about the low speed Of fermi in games. NVIDIA specially made an accent on the computational abilities Of fermi, in order to disguise the mediocre speed of this architecture in games. In reality, NVIDIA greatly reported about the GPGPU- abilities Of fermi only because of the best correspondence to this theme to the profile of (GTC) measure, on which the architecture was presented. Fermi possibilities in the graphic acceleration sphere impress, as asserts NVIDIA chapter , simply about them, it will be described later.
In ten years according to Mr. Juan, all processors will have a different structure, combining the computational, graphic and other specialized core.