Cranking up the heat

It has been quite a while for me, but I have finally got my computer back to running Folding@Home. Folding@Home (F@H or Folding for short), is a distributed computing project maintained by Stanford University, which benefits disease research. The program uses you CPU or graphics card to simulate protein folding, hence the name. Anyways, I recently purchased an Nvidia 8600 GTS graphics card from a friend. While I current graphics card is faster than the 8600, I intended to use them in tandem for different purposes. My faster ATI card drives my displays that I uses for games and work, and the 8600 runs Folding non-stop. This gives me the advantage of using a graphics card for Folding, which is faster than regular CPU folding, but at the same time be able to do all of my normal work without lagging. Although I could settle for the 1800 points per day that my graphics card can manage alone, I don't want to. In a quest for more folding power, I have created two virtual machines running Ubuntu Linux to run on my ultra-fast CPU. You may ask, why go with virtual machines? Isn't that slower than running a program natively? Yes and no. In this case, I choose to run the Folding client in Linux because the Linux version can make use of more efficient F@H (a2) cores which are much faster. However running the a2 cores in a virtual machine is slower than running them on an actual Linux machine. Now when I fire up both virtual machines and the GPU core, I can fold about 6000 points per day. With all those processors running, I end up with a lot of heat. My 8600 GTS card averages 70C folding, and my Core i7 CPU averages about 58C.