At MPF today, Nvidia presented it's CUDA (Compute Unified Device Architecture) programming model that let programs written in C language to take advantage of the power of Nvidia's latest Series 8 graphics chip family. The goal is to use the power of the graphics chip (GPU) to assist the main processor (CPU) to do parallel computing and accelerate memory and floating-point intensive applications.

"GPU performance has been scaling much faster than Moore's law for the last few years and we expect to continue for the years to come. Our latest chip has 128 threads processors, handling 12,000 concurrent threads. Now, the challenge is how the heck you control as many as 12,000 threads at the same time. And that's what CUDA does... 64-bits support coming in 6 to 8 months", said John Nickolls, director of architecture at Nvidia (picture).

Some of the computing applications Nvidia is focused on:
- computational geo-science like oil and gas exploration;
- computational chemistry like molecular modeling, big in-body problems, etc;
- computational medicine like medical imaging;
- finance like the Monte-Carlo simulation.