Nvidia has suffered heavily in recent years, both with technology and economy, which has its main competitor AMD a significant lead in graphics processors.
Now, Nvidia ready to roll out the first consumer products company Fermi series. In the first instance we are talking about the new top model GeForce GTX 480 and GeForce GTX 470 Nvidia partners will provide video based on these processors from mid-April.
In a few months the series will expand with simpler and cheaper models based on the same technology.
GeForce GTX 400 series is the first from Nvidia that supports Microsoft's DirectX 11 technology, introduced with Windows 7 in October last year. AMD launched its first graphics processor with DirectX 11 support already in September last year.
The new top model is probably the fastest graphics processor on the market, but the tests - including this [H] ard | OCP - shows that the GTX 480-based graphics card uses a lot of power and runs very hot. It requires strong cooling, which in many cases will create a lot of noise. The deemed price of around 4000 million too high, compared with cards based on AMD's top models, such as Radeon HD 5970.
A summary of the impressions is here.
Cards based on Nvidia GeForce GTX 480 (Photo: Nvidia)
Nvidia claims that the GTX 480 is on average have 27 percent better performance in newer DirectX 11 software than any other GPU's, and up to 89 percent better performance at certain resolutions and settings. These figures are enough, but critics seem not so terribly impressed with the performance lead Nvidia has now been given.
GTX 480 can deliver more than 746 million pixels per second, enough to support a resolution of 5760 x 1080 pixels.
It has also received support for real-time, interactive ray tracing, which will provide more realistic scenes, and for the Nvidia 3D Vision Surround, which makes it possible to distribute the display on three screens in a kind of stereoscopic 3D. This requires, however, the use of special glasses with active shutter.
PhyX performance will be twice as high as in previous generations, and a new GigaThread benefits will provide up to ten times faster switching between the graphical and physics processing.
The new chips also have a number of programming interfaces available to applications via CUDA including C / C + +, DirectCompute, OpenCL, Java, Python, and Fortran.
More details about the new graphics chips found on this page.
