NVIDIA Company frequently criticized the limitedness of application sphere for CUDA platform, since all advantages are possible only with products on base of their own video chip NVIDIA not older than the generation GeForce 8xxx.
On GTC 2010 NVIDIA representatives together with PGI declared about the intentions to realize CUDA support for processors with x86- compatible architecture. It is assumed that CUDA-x86 will allow developers to roll the necessary applications first on the central processors and then transfer their fulfillment to the graphic chips. Applications speed with the use of systems with a small quantity of processor core will lag behind the version with the use of computational video chip resources . Compiler for the transfer of CUDA applications to the x86- compatible platform will be extended BY PGI .