 |
|
|
41)
Message boards :
Number crunching :
Server Status
(Message 48033)
Posted 23 Jan 2014 by  mo.v
Post: Hello Dave I noticed last night that it's stuck at 13:59:00 and cannot be updated. After this had been confirmed by another member I emailed the programmers. It means there's an updating script that isn't running; normally it should update every minute. The script that should export our credits to the external stats sites still isn't working either. |
|
42)
Questions and Answers :
Getting started :
Security of the BOINC software and supporting infrastructure
(Message 48027)
Posted 22 Jan 2014 by mo.v
Post: I don't know whether you are thinking about a real or hypothetical state government, but it doesn't matter because no governments are involved in writing the software. Several pieces of software are involved. All the projects (climateprediction is a project) run on a software platform called BOINC. Read about it here: http://boinc.ssl.berkeley.edu/ I expect you've already downloaded it as otherwise I don't think you'd be able to post here. The chief programmer, Dr David Anderson, is at the University of California in Berkeley. He has a full-time and a part-time paid programmers to help. You can read more about the BOINC software here: http://boinc.ssl.berkeley.edu/wiki/User_manual Part of the security of BOINC stems from the fact that there are email lists on which volunteer helpers can post and collaborate with the development process. The openness of this process would in my view make it extremely difficult to subvert. The climateprediction climate models are based on climate models devised and used by the UK Met Office (called the Unified Model). The climateprediction (CPDN) programmers adapt these models for the specific research being undertaken and configure them for the three platforms: Windows, Linux and Mac. The core research staff are at the University of Oxford, but they collaborate with researchers at a number of universities, notably Southampton in the UK but also others in South Africa, Australia and the USA. The collaborative research is being extended to other countries. I suggest that in the blue menu on the left of this page you explore the Main page link. On the About page you will find a link to some of the staff involved in CPDN, both programmers and researchers. If you have more specific concerns please tell us what they are. |
|
43)
Message boards :
Number crunching :
VANISHING WU'S
(Message 48023)
Posted 22 Jan 2014 by mo.v
Post: Every night at midnight your time you'll still have your work quota put back to one model per core per day. While waiting for more good models we just have to get on with work from other projects, of which we should all be able to choose several we think are worth spending computer time on. The WCG projects are all highly respected. |
|
44)
Message boards :
climateprediction.net Science :
User defined region?
(Message 48022)
Posted 22 Jan 2014 by mo.v
Post: Of the two beginning with 7, one crashed and the other appears to belong to another batch as its workunit isn't marked No resubmission. Don't worry about the model whose number begins with 8 showing unrealistic temperatures. The starting values have to reflect the full plausible range. That's a wide range. Some of these will necessarily become unrealistic as they develop, but we crunchers are not supposed to decide where the cutoff point should be. The researchers have to calculate the likelihood of the different final values. In any case, if the temperatures or air pressure or anything else become so abnormal as to be physically impossible the models are designed to crash with recognisable messages in the stderr output. I've had a couple that have crashed in this way, but also a number of hadcm3 that have developed very high levels of sulphur particles without crashing. They probably show what the atmosphere was like after a volcanic explosion like Krakatoa ie very unusual but physically possible. So let your remaining models continue. |
|
45)
Message boards :
Number crunching :
Credit updates?
(Message 48019)
Posted 21 Jan 2014 by mo.v
Post: As we still haven't received a credit update on the external sites today, I've emailed the programmers to ask whether the program that exports our credits is inadvertently turned off. |
|
46)
Message boards :
Number crunching :
Credit updates?
(Message 48010)
Posted 19 Jan 2014 by mo.v
Post: Yes, I'd noticed that too. If we don't get our credits on the external sites tomorrow evening I'll email the programmers. Thanks for the reminder. |
|
47)
Message boards :
Number crunching :
ANOTHER UPLOAD PROBLEM
(Message 48008)
Posted 19 Jan 2014 by mo.v
Post: It isn't always obvious how to get into an AV to adjust the settings. I had one where the Help instructions were all on the web so if you had no web access you couldn't investigate what to do apart from trying all the buttons. The zip files won't have come to any harm through having to wait. |
|
48)
Message boards :
Number crunching :
models always for 10 years crash
(Message 48003)
Posted 18 Jan 2014 by mo.v
Post: Hi Steve It looks as if you have two nice computers. Here they are: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/hosts_user.php?userid=18823 But as you say, they've been crashing quite a few tasks and #1 in that list has been having more of a struggle. Here are its tasks: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/results.php?hostid=1307835 We can click on any of the task numbers and then see the details of the task by clicking on Stderr+. The messages that then appear and (sometimes) the error code (or exit status) often provide useful clues as to what happened. * The tasks (models) that say Download error are almost certainly due to some problem inherent in the models eg a missing file. Don't worry about them. * A small number of the crash messages show 5 or 6 times that there was an INITTIME error. This is something wrong with the model which tries to restart 5 times and is programmed to crash on the 6th attempt. Don't worry about these. * A few of your models crashed with code 25. This can be caused by a bluescreen crash. Is this ever a problem on this computer? * Sometimes two or three models seem to have crashed at the same time. I'm guessing that because they reported to the Oxford server at the same time. Something the models didn't like must have happened. One possible cause is that you turn off the computer without completely exiting from BOINC first. Before shutting down the computer you should: - open BOINC Manager - in the Activity tab stop computation. I also stop network access. - in the File tab click on Exit Don't just close BOINC Manager before shutting down the computer; that doesn't work because the models will still be crunching. * It's safer not to let your antivirus program scan BOINC and the models while they're running. Either exit from BOINC before scans or exclude BOINC from scans. Other members may well have extra suggestions. |
|
49)
Message boards :
Number crunching :
VANISHING WU'S
(Message 47988)
Posted 15 Jan 2014 by mo.v
Post: If you mean will one of us ask the programmers to minus your computer(s) so that you can't receive new tasks until you've posted on the forum and sorted out the problem, then no, of course not. When we come across computers that appear to be serial model crashers the first thing we do is look in detail at a selection of the models to see what the problem seems to consist of. We also look at whether the problem of crashed models has been going on for quite a while without the owner doing anything about it and without posting on the forum for advice. The last thing we want to do is penalise people for one-off or short-term problems which any of us could have to deal with from time to time. You and your computers don't fall into these serial-crasher categories. In fact it's rare for members who post on the forum to be serial model crashers. If instead you were wondering whether the server will reduce the number of tasks you're allowed per core per day, then yes, I'd expect the server to treat each aborted model like a crashed task. This is the BOINC rule. However, even if you abort so many models that your daily quota is reduced to zero new models per core, after your midnight the quota will be increased to one per core per day again. I've aborted quite a few of these defective models myself and advise everybody else to do the same. I believe that these models are likely to crash anyway at some stage so there's nothing to be gained through wasting electricity and computer power on crunching that cannot be of any help to the researchers. |
|
50)
Message boards :
climateprediction.net Science :
Prof Myles Allen looking for some funding
(Message 47967)
Posted 14 Jan 2014 by mo.v
Post: http://www.bbc.co.uk/news/science-environment-25684933 It's good to see CPDN and the Met Office working hand in hand on this. |
|
51)
Message boards :
Number crunching :
News and Announcements
(Message 47965)
Posted 13 Jan 2014 by mo.v
Post: You will see that things are up and running again apart from one upload server. Jonathan now needs to make an extra backup and will be closing down the databases again within an hour from now. He expects this extra backup to be completed within 24 hours. When this forum and consequently this News thread are down we always try to post updates on the boinc_dev forum in the Projects section where there's a thread for news of server outages from all projects here. |
|
52)
Message boards :
climateprediction.net Science :
Sherwood et al Climate Sensitivity at high end of range
(Message 47963)
Posted 9 Jan 2014 by mo.v
Post: Thanks for keeping us up-to-date with this. I missed a lot during the holidays. These ideas about atmospheric mixing look important. |
|
53)
Questions and Answers :
Unix/Linux :
"This project may not have work for your type of computer" on AMD processors and Linux x86_64
(Message 47954)
Posted 8 Jan 2014 by mo.v
Post: You mean on the following page in the Ubuntu section at the bottom? http://boinc.berkeley.edu/wiki/Installing_on_Linux |
|
54)
Message boards :
Number crunching :
News and Announcements
(Message 47951)
Posted 8 Jan 2014 by mo.v
Post: Jonathan says: Hi, I have to take the project database off-line for the first of (hopefully only) two upgrades at the end of this week. I will schedule the downtime from: 12 Noon GMT on 10 Jan until 12 Noon on 13 Jan 2014. There will be no database access during this downtime. For those who are interested, this the first step towards retiring our old database server, and moving her functions over to the virtualised infrastructure at the Oxford e-Research Centre. Jonathan Miller CPDN System-Administrator |
|
55)
Questions and Answers :
Unix/Linux :
"This project may not have work for your type of computer" on AMD processors and Linux x86_64
(Message 47946)
Posted 8 Jan 2014 by mo.v
Post: There is no problem with AMD computers, whether they use Windows or Linux. All the CPDN models are 32 bit. Could this be the problem? Does the following BOINC reference page help? http://boinc.berkeley.edu/wiki/Installing_on_Linux Please also see the following post by Geophi about 32 and 64 bit: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/forum_thread.php?id=6746#38077 |
|
56)
Questions and Answers :
Getting started :
climateprediction.net showing "No work available to process"
(Message 47936)
Posted 6 Jan 2014 by mo.v
Post: I think that's a good strategy. Especially as I forgot to mention a serious disadvantage to that method. On both computers you'd have to keep suspended the models not being crunched on it, and BOINC projects refuse to send work to computers where any task from those projects is suspended. This could create a very annoying situation where you might see work available but not be allowed to download any of it. There are some tips about how to maximise the number of models a computer gets when they become available in the last few posts here. |
|
57)
Message boards :
Number crunching :
Project has no tasks available
(Message 47935)
Posted 6 Jan 2014 by mo.v
Post: It's also important to remember that CPDN has reduced the number of times a computer can request extra work from the server to once per hour. We cannot change this setting which was decided to limit the load on the server. The countdown to the next hourly attempt is shown in the Projects tab of BOINC Manager. Do not try to ask for work now by clicking the Update button as this will reset the time to 60 minutes. Patience rules! And of course to see what models are available go to the Server Status link in the blue menu to the left. |
|
58)
Questions and Answers :
Unix/Linux :
How can I interrupt a task for new O/S install?
(Message 47933)
Posted 5 Jan 2014 by mo.v
Post: Earlier versions of BOINC are available from here, though as Les says I don't know whether this is good for your flavour of Linux: http://boinc.berkeley.edu/dl/?C=M;O=Dclimate_change_screensaver.exe |
|
59)
Questions and Answers :
Getting started :
climateprediction.net showing "No work available to process"
(Message 47930)
Posted 5 Jan 2014 by mo.v
Post: Hi again Michael Here is a suggestion about something you could do to crunch some models from computer #1 on computer #3. #1 is the source computer which I've numbered according to its place in the CPDN list and #3 will be the destination computer if you decide to do this. * You need some sort of external storage eg an external drive to move stuff from one computer then onto the other. * Both computers are Windows so the CPDN server shouldn't mind. * SETI will not allow work to be downloaded onto one computer then the results uploaded from another. This could make this idea impossible. Bearing this in mind you may prefer just to keep #3 busy with SETI work using a short work buffer of 1 or 2 days until more CPDN work becomes available. Then in the Projects tab first stop SETI downloading new tasks and then increase the work buffer to 10 days to maximise what you get from CPDN. * If you do want to go ahead, first empty the Recycle bin on both computers. So if you accidentally delete the wrong thing you can immediately restore it. * Do not try to move individual models. There was once a method for this that worked but that was several years ago and probably for BOINC 5. It was very complicated and the person who designed the method no longer posts. So instead you must move or copy the entire Data folder contents. * Find where the BOINC Data folder / directory is on both computers. Bear in mind that BOINC gives the same name (BOINC) to both its Data folder and its Programs folder. You will not be doing anything with the Programs folder, which is the one that contains a subfolder called Location. The Data folder contains a subfolder called Projects. Here's a list of where the Data directory should be: http://boincfaq.mundayweb.com/index.php?language=1&view=592&sessionID=ad14436e0d6a6b9295985e7343791057 * You will lose all the tasks on computer #3 (SETI tasks) so complete them first and set SETI to No New Work. * Suspend all the projects and all the tasks on both computers. Suspend the tasks individually in the Tasks tab so they can be restarted individually. * On both computers exit completely from BOINC (File > Exit). * On #3 navigate to the BOINC Data folder, open it, select the entire contents, Copy, then Paste these entire contents somewhere safe. Could be somewhere on computer #3, could be elsewhere. Saving these contents means that later you can restore this folder if you want. Name what you've saved carefully: where it came from & the date. * Now delete the entire contents of the Data folder on #3. You now have an empty Data folder. * On #1 navigate to the BOINC Data folder, double-click on it to open up its contents. Select all the contents, then Copy. Paste these contents on an external drive. * Link #3 to the external drive. Copy the contents (that came from #1) and paste them into the Data folder of #3. To do this right-click on the Data folder and in the drop-down menu select Paste. * You should now have the same Data folder contents on both computers. You can reopen BOINC Manager and should see the same tasks on both computers. You will have to decide which models are to be completed on #1 and restart them there, and which are to be started or completed on #3 and unsuspend them there. * When models have completed and reported on one computer I think it should be possible to abort them on the other computer. * At any time you can revert to the original BOINC Data folder contents on computer #3 by exiting completely from BOINC, deleting the Data folder contents, then pasting into it the original contents that you saved. But if everything goes as planned you won't want to do this. * I think this method should work and it would be useful for completing models for example from a computer that is going to be scrapped or have its disk wiped. But I don't particularly recommend doing this in other circumstances. I think we have to get used to the fact that CPDN work is often now intermittent and we should keep our computers busy by joining more projects, of which there are plenty. |
|
60)
Message boards :
Number crunching :
MORE DOWNLOAD ERRORS
(Message 47927)
Posted 4 Jan 2014 by mo.v
Post: Since CPDN installed its current BOINC server version, which was rather a long time ago, I believe that something has been wrong with the number displayed for a computer's daily permitted task quota. This happened so long ago (when Milo installed this version) that I can't remember his explanation. I think the quota number we see on a computer page (the displayed quota number) is random. However, the server knows what the quota number should be and applies that. What we see is not necessarily what we get. For example, both my computers currently show a daily quota of 0 but they have done nothing to deserve this punishment and when tasks become available I expect both computers to be allowed some. The only quota number to display correctly is -1 which appears when Andy manually minusses a computer that's a serial model killer. When Andy and Jonathan install a new BOINC server version we may have an interesting week. At a meeting a while ago in London I told Andy that Carl and Tolu (CPDN's original programmers) kept a list of CPDN customisations for BOINC. He didn't know about this list. It's so long since CPDN upgraded its BOINC that I now think it might be better if the list is lost forever. |
©2024 climateprediction.net