Distributed Computing Volunteer Projects
I wanted to make a quick post about some ways in which we can volunteer the idle processing power of our personal computers to help out on some pretty hefty research projects going on around the world. Most of these are based on running millions of large-scale calculations and simulations, which have traditionally required a massive and expensive supercomputer to be purchased and maintained on campus. Over the last decade or two, more and more projects have been popping up based on distributed computing, spreading the load of calculation over many smaller, inexpensive computers connected over the internet, which is where we come in.
To participate, users volunteer by downloading and installing a piece of software created by the project developers. Once installed, the software runs quietly in the background, similar to an antivirus program. Because the program is designed to use few resources and take advantage only of processing capability you are not using, you should not notice a performance impact on your machine while doing things like catching up on email or watching a video, at least not on a relatively modern computer.
Because these processes are run on normal, every-day computers owned by normal, every-day people, many refer to this kind of computing a supercluster, rather than a supercomputer.
As far as what these programs are trying to accomplish, it really depends on the project. One of the more popular distributed computing projects is Folding@home, sponsored by Stanford University, which simulates the formation of proteins. Through millions and millions of simulations, the ultimate goal is to reach a better understanding of why proteins sometimes form incorrectly, and how these malformations relate to disease. Folding@home has statistical tracking for individual users, so volunteers get to see how many simulations they’ve completed and contributed. This is a great way for students to track their impact on project, or to plead their case for extra credit.
Oxford University’s Climateprediction.net runs simulations of climate models in an effort to predict climate change over the next century. The idea is that the more accurately we can predict climate change, the faster we’ll be able to prevent it. Here’s a video snippet that relates the need for more simulation and more processing power for the project to produce usable results.
The Great Internet Mersenne Prime Search project works to identify larger and larger Mersenne prime numbers. If you have a few minutes, listen to an NPR story about the GIMPS project here. There’s a cash reward for any user who discovers a new prime while running the program, if you need help incentivizing volunteerism.
For more distributed computing projects, check out the list on Wikipedia.
~originally posted by Carson Waites.
1 Comment »
Filed under: Distributed Computing


