Nvidia has unveiled plans for a new line of graphics processor units (GPUs) which use the company's new Fermi architecture to improve graphics performance, and expand the role of the GPU into new processing tasks.

Nvidia said that it will use the Fermi architecture across its business, including the GeForce, Quadro and Tesla lines.

"It is completely clear that GPUs are now general purpose parallel computing processors with amazing graphics, and not just graphics chips anymore," said Nvidia chief executive Jen-Hsun Huang.

"We will look back in the coming years and see that Fermi started the new GPU industry."

The expansion of graphics chips into general computing tasks, known as General Processing over GPU (GPGPU), has been an area of interest for both CPU and GPU vendors in recent years.

GPUs are designed to process multiple operations simultaneously in order to render complex images, and are considered well suited to multi-threaded tasks. GPGPU is of particular interest for activities such as climate simulations and medical imaging.

Nvidia is not alone in its work on GPGPU systems. AMD has entered the arena with its ATI Stream platform, while Intel is integrating GPUs into general processing with its upcoming Larrabee chips.