It seems like yesterday that I was at the ATI press conference in Tunis, Tunisia where Mike Houston gave the presentation on how F@H was coming to ATI graphics cards. The presentation slides showed much promise for GPU folding, but the following months proved to be hit or miss with the drivers and many folders abandoned folding on their graphics cards. Other than the drivers, many enthusiastic folders expressed frustration about how the points were being calculated for GPU folding compared to the other ways one can fold.
It has been over a year since ATI got into the Folding@home project and it seems that the time is right for NVIDIA to join in! Mr. Vijay Pande himself was in town giving the presentation on behalf of Stanford University and the project team that he leads.
Vijay gave a quick background on what folding was and how a petaflop of computing can and does make a difference in research. He used the direct simulation of Alzheimer’s Disease as an example of just one area the team is working on.
He also pointed out that the common flu causes ~40,000 deaths a year here in the United States and that folding also does the direct simulation of viral infections as well. Vijay just touched on what folding can do, but if you don’t know about folding please check out their website to find out more information.
When it comes to processing power, you want to have PetaFLOPS worth of processing power and folding@home has an impressive 1.9 PetaFLOPS! If you look at the list of the fastest super computers in the world today, it shows how much power distributed computing really has.
The highlight of the presentation was when folding@home was announced to run on NVIDIA graphics cards. It seems that the folding team was able to program a folding client in C thanks to the NVIDIA CUDA software environment. NVIDIA has 70 million graphics cards on the market today that can use CUDA programming, which means that the folding project could get a huge boost in performance if more end users would join the project. As we mentioned this morning in the article on CUDA and Ray Tracing, it seems that more and more programs are using CUDA and this is a good sign for both NVIDIA and those that have one of the 70 million cards on the market that can already make use of the code.
To conclude, the presentation Vijay Pande showed the new NVIDIA GPU client running a live demo on the next generation GeForce graphics card. We had to blur out the product name and are unable to show you performance numbers since there is an embargo on information. We can tell you that Vijay seemed impressed with the outcome of the project and that it is the fastest thing to ever run folding. If you are interested in joining folding, be sure to check our Legit Folding Team as we have great bunch of people that can help you get started!