Checking on How Your Climateprediction.Net Work Is Progressing
From Unofficial BOINC Wiki
Processing work for Climateprediction.net (CPDN) is a little different than for most of the BOINC Powered Projects because of the time it takes to complete one model. With most Projects the run time is sufficiently short that the Work Unit will be completed within hours, usually days, and certainly within a week or so, and therefore we can wait until the work is completely done. To compensate for the very long run times, the progress of the model is regularly reported back to the Climateprediction.net (CPDN) Project as a matter of course. You can check your model's progress by performing some of the processes outlined below.
Contents |
[edit] Climateprediction.net Graphics
The best way to see if the model is running properly on your computer is to use the show graphic button on the Work tab of BOINCMgr.
However this article is mainly about seeing what has been returned to the Climateprediction.net servers and made available to you. This checks that work is being trickled up to the server correctly.
[edit] Procedure
[edit] Your Account
To see how your work is progressing, go to the BOINC Project's website and select "Your Account".
This is the "Your Account" Page.
To see how your work is progressing, we can look at the "Results For User" Page which shows all of the Results in progress or completed for all your computers. An alternative that may make it easier to find a particular Result that you are interested in is to select "View computers", then select the computer by clicking on the Computer ID, then clicking on the number of "Results" to see the list of Results for that computer as shown on the "Results For Computer" Page.
[edit] Results for User/Computer
This computer appears to working OK. The Server State column shows "In progress" because the Work Unit is in progress. The Outcome is "Unknown" because the Work Unit hasn't been returned yet.
Click on the ResultID for more information.
[edit] Result Page
First there is an information section. This tells you more about the Work Unit. Exit Status and Validate State can tell you more about any errors that have occurred. 0 (0x0) and OK mean that everything appears OK.
There are links to the Work Unit Page, the computer ID and to the Trickles.
Next there is a perturbed Parameters section. This tells you how your model differs from the standard model by telling you which Parameters have been altered. See [1] for details of each Parameter.
Next there is a section for graphs of your run. You may want to compare your run to some of the First Results
Click the Work Unit link for more details of the type of model and whether any other Participants have run this Work Unit.
Or click for the graphs (detailed next).
Or click for the trickle information.
[edit] The Graphs
From the result page you can click to see various graphs. This is the all phase temperature graph.
[edit] What does this mean?
This is a 3 Phase 45 year model. First Phase measures fluxes necessary to keep the model stable. Second control Phase tests whether the measured fluxes do keep the model stable. The third Phase includes a forcing of doubled CO2 to see how the climate reacts to such a change. The climate sensitivity is the difference in temperature between Phase 2 and where Phase 3 would reach equilibrium temperature. This model hasn't quite reached equilibrium, however it is clear enough where it is heading. Processing more than 15 years in Phase 3 could be done to be sure about it but the error in estimating the equilibrium temperature from the shape of the graph so far is considered to be small. Therefore, it is considered better to process more models than continue processing for more than 15 years. An exponential extrapolation is used to calculate the equilibrium temperature. Details of this extrapolation are in the methods section of the Nature Paper.
[edit] Trickle Information Page
Click on the number of trickles on the Result page to see details of the trickles.
This page gives details of the trickles that have been sent up to the server. Details of time sent, the computer the result ID and the result name are obvious enough. The Time-step and Phase is showing where you have got to with the model. The CPU time is the amount of time that your computer processor has been working on the result. The average Sec/TS is seconds per timestep; so lower numbers represent higher speed systems. The average is over a whole model. Therefore, if you change your system to speed it up, it will take a long time to be reflected in this average sec/TS measure. You may want to calculate the sec/TS over just the last trickle in order to see what speed improvement you have got.
[edit] What if there are no trickles?
Your first trickle may take a while a come through. You may need to process a model year, trickle up the information and then wait for upto a day before it appears.
[edit] I have definitely waited long enough and still no trickle, what now?
- Check that your firewall is allowing BOINC to make contact with the server.
- Check that network access is allowed - see commands menu.
- With a connection to the internet on, go to the projects tab select the project and click update. Then look at the messages tab to see if that appears to be sucessfully making contact.
- If the server is not responding, it could be down and there may be information in the news or message boards. The server status page might indicate if there is a problem.
- If BOINC makes sucessfully makes contact but does not send a trickle, maybe there isn't one to send. Check you are not getting a succession of failed models.
[edit] Work Unit Page
There is a link to the following page from the results for user page, the results for computer page and the result page.
This tells you which application you have run. It should be clear whether it is a Slab Model or a Sulphur Cycle Model or a Transient Coupled Model in the application details.
Someone else may have run the same model as you. In this case the other person failed on download. However, sometime you are lucky and someone else is successfully running the model. If so you can compare your graphs to what the other persons model is doing. Don't be surprised if you can detect differences - see Why doesn't the exact same model work out the same?
