Better climate models are key to understanding how best to protect the environment and food production, but they require massive computing resources. You can help. Climateprediction.net allows researchers to parcel out simulations to computers that are online but not being used to full capacity. More than 96,000 idle computers in 138 countries have already been tapped to run thousands of simulations. Each participant receives a unique task, which could take days or months to complete. Current projects include models of atmospheric and oceanic forces acting from 1920 through 2080. Interested? Log on to www.climateprediction.net and download BOINC, a software platform that allows you to volunteer computing power for a range of scientific ventures.
Note: This article was originally printed with the title, "Wanted: Spare Computer Power".
See what we're tweeting about
11 Comments
Add CommentInteresting Mindset on Display....It's not "volunteer your unused processor cycles for "Climate Research"....It's "help us measure global warming"..After all the "science" is settled; isn't it. Any bias at SA?
Reply | Report Abuse | Link to thisBut the Earth hasn't warmed in 10 years despite rising greenhouse gases. If anything the global temperature has begun to fall in the last 2-3 years. So why would I allow my computer to burn fossil fuel derived electricity to feed a climate model which cannot account for this hiatus?
Reply | Report Abuse | Link to thisYou Idle Computer Could Help Contribute to Global Warming, the Energy Crisis, [insert catastrophe here].
Reply | Report Abuse | Link to thisJoebob, the science is by no means completely settled, which is why the ClimatePrediction.net models are used by the researchers not only to predict a wide range of possible future climate scenarios and the likelihood of each, but also to improve the techniques of climate modelling itself.
Reply | Report Abuse | Link to thisJohnA, global temperature rises are affected not only by levels of greenhouse gases but also by El Ni�o and La Ni�a events in the Pacific. These are factored into ClimatePrediction.net models and also, I believe, into other types of climate model.
The models are in fact very sophisticated and different types of model are used for different research projects.
TacoTRUCK, the question of whether the extra electricity used to process climate models is worthwhile has been discussed at length by members of ClimatePrediction.net on their forums. Some members attempt to offset the extra consumption of electricity by reducing their energy expenditure on other things, or let the heat from their computers replace some of their domestic space heating. The same arguments apply of course to all research projects run on home computers. Some such projects are listed here:
http://boinc.ssl.berkeley.edu/projects.php
If volunteers did not run tasks from such projects on their home computers, the research would in some cases be run on supercomputers. In other cases sufficient funding might not be available and the research would not be done, or would not be done so thoroughly.
There's ALWAYS somebody who wants to take over ur computer! No way you gonna goof mine up!
Reply | Report Abuse | Link to thisGreat idea to use the idle computers for this model, but besides narrowing down the data to measure, am assuming a fraction of these computers be used for measuring how proposed solutions to climate change would adjust these same models?
Reply | Report Abuse | Link to thisZenaV, in no way do the ClimatePrediction.net models 'take over' anyone's computer. If the models took away our control of our own computers I don't think many people would want to run them. But the fact is that more than 50000 home computers are currently running these models. See the statistics here on the home page of the project's web site:
Reply | Report Abuse | Link to this<url>http://www.climateprediction.net/index.php</url>
The models run on the BOINC platform (Berkeley Open Infrastructure for Network Computing) which is described and well documented from here:
<url>http://boinc.ssl.berkeley.edu/</url>
The BOINC program can be completely uninstalled at any time, the times at which the research tasks run are set by the user, the amount of processing time devoted to them is also determined by the user, and all the well-known research projects run a forum where members can get advice on making the research tasks run exactly when and how they want.
Jonathan, one of the more recent ClimatePrediction.net research projects is being run jointly by Oxford and Carnegie-Mellon Universities. It's called the Geoengineering Experiment and uses 80-year simulations to study the effect of a variety of possible mitigation scenarios on future climate. The project is described in more detail here:
<url>http://www.climateprediction.net/science/geoeng.php</url>
(I hope my html tags will now produce active links!)
Let me repeat for mo.v: Your headline is your premis...let's all volunteer our computers to help "calculate global warming" (anthropogenic is assumed in all pc climate discussion)
Reply | Report Abuse | Link to thisThus, for the purposes of your proposed study, global warming (anthropogenic) is given...its existence is NOT the purpose of the study, only the magnitude and rate of the assumed man-made warming is at issue
Sounds more like a funding proposal then science to me.
The headline of the article isn't mine; it was written by or for Barbara Juncosa of Scientific American.
Reply | Report Abuse | Link to thisI am not advocating that we should all volunteer our computers to run ClimatePrediction models. There are plenty of other research projects of equal scientific worth that need the contribution of volunteers.
It is by no means assumed in all discussion of global warming that it is caused solely by anthropogenic greenhouse gas emissions. It is common knowledge that many other factors are also involved, for example volcanic emissions and changes in the sun's energy output. Some of these factors can be roughly predicted due to their cyclical nature but some can not. Some anthropogenic emissions eg soot and other aerosols tend to lower global temperatures. There is therefore a wide range of possible future scenarios, and hitherto unkown factors may yet emerge, confounding all the predictions. It cannot therefore be said that 'only the magnitude and rate of the assumed man-made warming [are] at issue'. A small proportion of climate models that have been processed on more than one stable computer show a fall in global temperatures up to 2080.
But what is most likely and less likely to happen to future temperatures and precipitation, given the inputs from current knowledge, is emerging fairly consistently from all types of climate model.
I am not one of the researchers involved in the ClimatePrediction studies, nor am I in a position to propose any particular research project as I am not a specialist in this field. I have, however, been running ClimatePrediction models on my home computer for several years. I am interested in how the models develop and in how the model design can be varied for different types of study.
It should be noted that idle computers consume considerably less power than active ones. Long gone are the days of computers needlessly spinning in circles executing NOPs waiting for something to do. So by installing this software (like SETI@HOME etc.) you're using more power and thus spending more money, ultimately contributing to global warming. That's not to say it's not worth the trouble/money (I don't know) but it's certainly not free.
Reply | Report Abuse | Link to thisThere's no free lunch.
To verify this for yourself, install the software and note if the fans in your laptop keep running rather than spin down as they usually would when your laptop is idle. Note how hot the exhaust gets. Your laptop is consuming more power. The same works for modern desktops.
You're right, there's no free lunch, except that if during the winter months the extra computer power replaces space heating provided by electricity, there's probably little or no extra expenditure. There would still be extra expenditure compared with gas space heating.
Reply | Report Abuse | Link to thisI must say though that I'm surprised at how little extra heat my fairly new dual core produces when running two climate models side-by-side, compared with the heat that an older and less powerful machine used to produce running a single model. The newest computers are even more energy-efficient.
Potential volunteers need to look critically at the projects' websites and judge whether the research is worthwhile. I feel that a lot of the research eg into malaria and proteins really needs to be done and contributes to the common good.
The ClimatePrediction research team have made the volunteers' model results available for free download by climate research teams worldwide.