Specialised applications developed in the domains of fundamental or applied research, financial computing, communication and other demand ever increasing computing power.
With the power of PCs rapidly growing over the past few years, a solution has been available on the market for a while, clusters. This involves inter-connecting a number of PCs via a high debit network. This technique gives a lot of computing power but remains very expensive, cumbersome and consumes a lot of energy.
The arrival on the market of very powerful parallel processors (graphic cards) the past few years has led to the development of a new solution: GPU-Computing more commonly know as Stream computing.
In fact, as Dave Orton, president of ATI Technologies Inc. affirms, « graphiccardswere not onlyconceived to makegraphics. ATI processors have one of the most technologically advanced computing architectures in the world. Today, we are bearing the first fruits of our work with the apparition of Stream Computing initiatives. Large firms, key software developers and academic institutions are today demonstrating that Stream Computing is the key to enormous performance gains for several applications ».
It is now possible to carry your applications to the graphic card using our parallelisation technology. The GPU-Computing technology was awarded the ANVAR prize by the French Ministry of research for its novelty.
By harnessing the graphic card power, GPU-Computing is a faster and cheaper solution than the cluster. GPU-Computing is also the best way to reduce your energy and hardware maintenance costs.
