If you've been watching diaries carefully, you've seen a few in the last couple of weeks promoting the Folding@Home project, and the Kos team within that project. This is a very worthwhile distributed computing project, where you assist the project by donating CPU power from your own machine(s). I've got a PS3 working hard on protein folding for the guys at Stanford (and the Kos team).
However, I've got another project that I contribute to with my desktop: ClimatePrediction.net
ClimatePrediction.net is run out of Oxford University in the UK, and it's goal is to build and refine computer models of the Earth's climate. Computer climate models are an important tool in understanding and solving global warming, and I'm very proud to be donating CPU cycles to this cause.
Climate models are extremely computationally expensive. A good model will divide the Earth's surface into millions of tiny regions, and then model multiple layers of atmosphere interacting with the earth and multiple layers of ocean, with complex models of atmospheric gasses and pollution and weather effects, for each and every region. And then each region will affect the condition in it's neighbors. And then this extremely complex system needs to be run across decades if not centuries of target space. And then you need to repeat the whole process with a different set of initial conditions and climate sensitivity parameters. And then you have to repeat the whole process again with different types of human influences added in to the mix, to see how much effect there is as humans change their behavior over time.
About ClimatePrediction.net
The aim of climateprediction.net is to investigate the approximations that have to be made in state-of-the-art climate models (read more about this). By running the model thousands of times (a 'large ensemble') we hope to find out how the model responds to slight tweaks to these approximations - slight enough to not make the approximations any less realistic. This will allow us to improve our understanding of how sensitive our models are to small changes and also to things like changes in carbon dioxide and the sulphur cycle. This will allow us to explore how climate may change in the next century under a wide range of different scenarios. In the past estimates of climate change have had to be made using one or, at best, a very small ensemble (tens rather than thousands!) of model runs. By using your computers, we will be able to improve our understanding of, and confidence in, climate change predictions more than would ever be possible using the supercomputers currently available to scientists.
Experiment Strategy (basic)
The climateprediction.net project comprises three separate experiments - one to explore the model we are using, the second to see how well the models replicate past climate and the third to finally produce a forecast for 21st century climate. Each model that we distribute will be used for all three experiments. Each model distributed is unique, and differs from all the others in three ways: the initial conditions it is started from, the attributes which force it to be in one particular climate state and the parameters which make up the actual model.
Yea, computationally, it's a massive problem. And you can help.
The Oxford guys have broken this massive problem into little chunks, and the chunks are small enough for you to run one on your home PC. And by small enough to run, I mean a single model run may take 3 months to complete! However, with enough people helping, it's amazing how much computation can be completed. As of this diary, roughly 54 thousand computers are active in the project, and a total of 42 million models years have been processed.
So, how do you assist? First, download and install BOINC, the Berkeley Open Infrastructure for Network Computing. This is a wrapper program for dozens of distributed computing projects, it's the software that manages the download of tasks, the execution of those tasks, and the upload of results. Second, tell BOINC that you want to attach to a new project, and select ClimatePrediction.net from the list. You'll have to create an account to identify yourself, but that's quick and easy. The process is mostly automatic after that, just let it run on it's own. The computation is run at idle priority, so it won't slow down your machine while you are using it (though it can hold onto some memory even when it's waiting for more important tasks to complete). You'll need network connectivity now and then to fetch models and return results, but the overall bandwidth is pretty small.
Getting Started
1. Please read the following before participating in the experiment:
* Rules and Policies
* License Agreement
* Technical Requirements
* Frequently Asked Questions
2. Download BOINC for your type of computer
3. Install BOINC on your computer
4. Run BOINC on your computer
5. You will be asked to enter the project's URL, which is climateprediction.net. The BOINC client should then connect to our server.
6. You can then create your account, or enter your current account details
7. The server should send you your first climate model experiment.
8. Watch your model using the visualisation
ClimatePrediction.net isn't nearly as competitive as Folding@Home. There is a DailyKos team to join, if you want to go that route. I've been number crunching on this project since 2004, and was without a team for most of that time. There are some tools to keep track of BOINC projects, but I've never paid them much attention.
::
Ok, here's a big issue to think about: Instead of turning your machine off at night, and letting it run in a lower power mode during the day, you may now ask your PC to run full throttle 24/7. While this doesn't really put much extra wear on the pc, you are consuming more electricity, and thus making global warming worse. Oops. :(
Normally, I'm all in favor of reducing my electrical consumption, but this is a special case. I think this is one of those places were consuming extra electricity is actually a good thing, just like switching from a gasoline powered vehicle to an electric one is also a good thing. Consuming a little extra power now will directly contribute to global warming understanding and solutions down the road, so the benefits outweigh the costs. If you don't agree, then you probably shouldn't sign up until you build your own private solar farm.
I will say one thing: Don't let BOINC run as a screensaver. Let it run in the background, but let the power management on your machine turn your monitor OFF, and with a very short timeout. You may be running the CPU at full throttle, but there is no reason to let your monitor consume power when you aren't looking at it. ClimatePrediction.net does have a pretty graphic showing the current state of your model, a representation of the globe with temperatures, clouds, etc, but don't waste too much power looking at the pretty pictures. :)