 |
|
|
41)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19853)
Posted 1 Feb 2006 by crandles Post: Most importantly, I do not know if the stuff I said is even close to being right. But, it is more or less what *I* gather. Sorry Paul but I don\'t think it is close. What you have written is more appropriate for what the Hadley Centre that developed the model does. My take on it is: CPDN is using the model (which is rather different from developing the model) to try to be able to produce a probability distribution function of future climate. Have you read the <a href=\"http://boinc-doc.net/boinc-wiki/index.php?title=Ensemble\">Ensemble</a> page recently? If that is too technical perhaps the CPDN home page. |
|
42)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19827)
Posted 31 Jan 2006 by crandles Post: OU course ... my guess is that I would need to come to Oxford to attend? A big problem for those of us that have, ahem, issues with airplanes ... :) Er no. The OU course is all distance learning - no centre to attend, no exam other than a home completed assessment. (I have forgotten whether the US is covered.) Edit M Dzbor wrote:
|
|
43)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19826)
Posted 31 Jan 2006 by crandles Post: Not sure I want to drag this thread back on topic but.... Rather than plucking figures like 10% out of nowhere, I probably should have used wikipedia as an example. They have 800,000 members and the last 3 week fund drive over Christmas raised 350,000 USD. I think they had 3 fund drives in the past year. They have quite a nice slogan: \'Give the gift of knowledge\' which you do whether you spend some time creating or editting articles to improve them or give money. Anyway, the lack of pledges here has convinced me it was a bad idea. You have to persuade people to pledge, then there is the possibility that the needed amount is not reached and then there is the possibility that the person never realises it is time to pay up on the pledge. That is 3 possibilities of failing to get the money. It is better to just persuade people to give so that there is only one chance of failing to get the money. I intend to pay my £100 despite the £10,000 in pledges not having been reached. I will now probably write a letter enclosing the cheque and hope to do so in such a way that the letter could be attached to a grant claim to show something like that there are lots of participants who are eager for the science outreach work and if a funder want to fund science outreach it is sensible for them to want to push on an open door rather than try to force science into closed minds. (Not sure if that would be useful.) My own inclination is that participants already contribute mightily via their volunteered flops - hitting them for money as well would feel slightly extortionate. [Though I\'m willing to be persuaded regarding the merits of the idea. :-)] That is a nice gesture. (But I am not sure you are in a position to be able to afford such a nice gesture.) |
|
44)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19824)
Posted 31 Jan 2006 by crandles Post: There are scientists looking at the \"stuff\", but, what directly is the result of *MY* contribution is invisible. I don\'t know if I have begun to answer this question with the ensemble answer, but I\'ll have another go: The ability to compare a particular model to lots of others with the same parameter except for X is a bit intangible and difficult to express. It is also a problem in how do you assess the benefit of having an ensemble size of 100,000 compared to ensemble size of 1000. Is there a strong diminishing return with increased sample size? (My brief answer would be nowhere near as strong as I am used to as an auditor because of the multidimensional nature of the data and the nonlinearities in the climate response.) I have raised queries on this sort of issue with the team and been trying for many months to get Dave Stainforth to come to the board and post something on sampling strategy and related issues that might cover this. It was annoying to be told I couldn\'t share the discussion with other participants. In fairness to Dave Stainforth, he has been to the PHP board to make a post and promised more in the next few weeks (that was nearly 4 weeks ago). |
|
45)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19822)
Posted 31 Jan 2006 by crandles Post: Some things, like the meaning of various parameters, was explained by (Dave Frame)? in the other \'message\' forum, (formerly the Community forum), and should perhaps, be added as part of the documentation somewhere. Umm, you mean like <a href=\"http://www.climateprediction.net/science/parameters.php\">http://www.climateprediction.net/science/parameters.php</a> (which is linked from <a href=\"http://www.climateprediction.net/science/strategy.php\">http://www.climateprediction.net/science/strategy.php</a>) |
|
46)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19816)
Posted 31 Jan 2006 by crandles Post: Perhaps some of \"Where does my small part fit in?\" is explained in the OU course: Modelling the climate. I don\'t think it is really. Perhaps we need an explanation of how an <a href=\"http://boinc-doc.net/boinc-wiki/index.php?title=Ensemble\">Climate Ensemble</a> helps. The trouble is I think it is obvious that an ensemble does help to provide a handle on the level of uncertainty. If this isn\'t obvious to someone else, then I think I would have trouble explaining. Perhaps my above linked attempt at explaining what a Climate Ensemble is a start. From there we can go on towards explaining how the different ensembles can be used to address different uncertainties. eg initial condition ensemble to get a handle on natural variability and the butterfly chaotic sensitivity. I think it is a big job to try to \"tie the two ends together\". I also think an expert could do a better job of it than I can. All I can do is try to make a start and hope that it slowly improves. Ideas on how to explain things better would be welcomed :) |
|
47)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19781)
Posted 30 Jan 2006 by crandles Post: Thanks Dave. :D Interesting numbers that put the situation in context: 170k/47k = 3.6 full time equivalent staff members. 45000/3.6 = 12,441 partipants questions per full time equivalent staff member to be dealt with in addition to their own work. So what is the reaction to this? Is it: I must reduce the number of times I ask questions and expect an answer? Or is it: to go off in a huff because our work is worth so little attention? Or is it: to think we must go ahead with trying to raise some more finance for them? Or is it: just suprise they haven\'t tried us for funding? |
|
48)
Message boards :
Number crunching :
Tropical Storms and Hurricanes
(Message 19776)
Posted 30 Jan 2006 by crandles Post: I just picked the first model that came to hand. Look at the difference in precipitation for India between summer and winter. So the models do show monsoon effects. < a href=\"http://www.boincforum.info/boincuser/Crandles/2fpaaa.pe15c10.x1.nc_7_06.JPG\">Pic 1 Summer</a> <a href=\"http://www.boincforum.info/boincuser/Crandles/2fpaaa.pe15c10.x1.nc_7_04.JPG\">Pic 2 Winter</a> |
|
49)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19775)
Posted 30 Jan 2006 by crandles Post: Hi Guys My plan only called for 45 donations of that size. There is room in the plan for 4050 people contributing just 10 GBP = 23.5 AUD and 40500 people who wouldn\'t be in a position to donate that much. Unfortunately as it stands I expect that far more than 40,500 people won\'t even see this thread. :( If there were lots of posts on this thread saying <quote>I pledge 10/50/100 GBP/USD/Eur/AUD provided only that other participipants pledge a total of £10,000 GBP</quote> then maybe it could make it into a news item or newsletter and get the coverage necessary. Unless those start pouring in, I think perhaps it will have to be relegated to an interesting hypothetical question of what should a participants\' representative try to negotiate if they were in a postition to negotiate something. |
|
50)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19755)
Posted 29 Jan 2006 by crandles Post: £100 GBP = $177 USD according to this <a href=\"http://www.xe.com/ucc/convert.cgi\">currency converter</a>. I guess we would have to get a paypal account up and running for collecting the money. (I would like to see gift aid being collected on donations from uk taxpayers.) |
|
51)
Message boards :
Number crunching :
Hypothetical? funding plan
(Message 19753)
Posted 29 Jan 2006 by crandles Post: I could write a cheque for £100 and send it off to Oxford. I haven\'t done this yet. Perhaps the cynic in me says it would just be gobbled up without ever knowing if it achieved anything. So I got to thinking how should participants fund the project in such a way that they would see some benefit and perhaps even have some say on how the funding was directed. So I decided to start this thread where instead on just giving £100, I pledge £100, conditional only on whether other participants pledge a total of £10,000. If the pledges reach that amount (or such other larger amount that other participants decide upon), then a participants representative would be elected to negotiate what should be funded with the money perhaps ascertaining a few options for the funding participants to vote on. If this allows more participants to see some tangible improvement to their crunching experience, perhaps more crunchers would be willing to donate. Does this have the potential to achieve significant funding. Lets see; if 10% of the 45000 participants pledge at least £10, and 10% of those pledge at least £50, and 10% of those pledge £100 then the total raised would be over £65,000. That should be enough to pay for a post or two for a while. I doubt that much will be raised but it is fun to think through the possibility and see what suggestions arise. I would like to see a participant going to interview some of the core team members once a month or so, record the interview and make it available over the internet for participants. That shouldn\'t cost much nor take much of the core team\'s time. I would also like to see some science being funded because that is what I am really crunching for. Anyway, ideas and pledges welcome ;) |
|
52)
Message boards :
Cafe CPDN :
Milestones Thread
(Message 19690)
Posted 27 Jan 2006 by crandles Post: Congratulations. I have also just gone through 100,000 too. :D If you see other people like me with odd large amounts today, it is because the credit for the spinups have been transferred. |
|
53)
Questions and Answers :
Windows :
newbie: confused by what \"to completion\" is meant to be
(Message 19550)
Posted 22 Jan 2006 by crandles Post: The to completeion time is very poorly calculated until you have done a few work units. Your 1519 total hours should be reasonably accurate. BOINC used to be good at this for CPDN ( and calculate 1519 same as you have) but poor for other projects. BOINC improved its method for other projects but now it acts in this manner of not being good until a few work units have been done. HTH |
|
54)
Questions and Answers :
Windows :
BOINC trying to download a second (and third) WU while still processing first...
(Message 19548)
Posted 22 Jan 2006 by crandles Post: Version 4.x had a bug that seemed to cause this download extra unneeded WUs. 4.45 seemed to have cured it but I occasionally see reports of it happening with 4.45. The other thing is that if you suspend a WU and BOINC wants to do work on that project it will download another WU. So if you want to suspend a project do it from the projects tab not the work tab. There are 3 levels of suspend - WU level, project level and suspend activities from file menu (Commands menu in v5). HTH |
|
55)
Questions and Answers :
Windows :
Unable to contact server
(Message 19531)
Posted 22 Jan 2006 by crandles Post: It looks like your model crashed about 97.7% of the way though a model. This is extremely unfortunate timing! I think it almost certainly failed before trying to make contact with the server, so that server glitch did not have any effect on the outcome of your model. Your computer looks stable enough having previously completed 4 models without problems. With that level of stability it could be a set of parameters that are a bit weird and cause the model to fail. The scientists want to know about these. (It could also be a very rare glitch on your computer.) So what is the appropriate action? Is it to give up on CPDN? or Is it to complain about the size of the work units? (This is unlikely to be helpful as it has been thoroughly considered and deemed necessary to have these long WU.) or Is it to occasionally do a backup so if a model does fail at 98% it can be rerun without loosing all that work? It is, of course, your decision. |
|
56)
Message boards :
climateprediction.net Science :
transfers keep retrying
(Message 19401)
Posted 17 Jan 2006 by crandles Post: Yes probably too late. Can I suggest you email them to carlc at atm dot ox dot ac dot uk If you say you had a permanent upload failure and give the resultid, result name and attach the zip files that didn\'t upload (but you copied). If you can also send the result_name.xml file that might help identify which server it should be sent to. |
|
57)
Message boards :
climateprediction.net Science :
transfers keep retrying
(Message 19373)
Posted 16 Jan 2006 by crandles Post: Will sort out a method when/if you report a permanent failure. HTH |
|
58)
Questions and Answers :
Windows :
Visual Fortran error
(Message 19372)
Posted 16 Jan 2006 by crandles Post: Unfortunately W98 is not supported with the BOINC client. The classic client can I think be used, but you don\'t get boinc credits for it. To find that client, you need to look for the open university page from the home page. |
|
59)
Message boards :
climateprediction.net Science :
transfers keep retrying
(Message 19234)
Posted 12 Jan 2006 by crandles Post: I suggest you copy the 4 files so if it permanently fails instead of temporary failures after 2 weeks, you will still have the files and the science won\'t be lost. I don\'t know what the problem is but it occured to me that you may not have uploaded any large files before. Could there be a size limit on your firewall? |
|
60)
Message boards :
climateprediction.net Science :
How does this project work?
(Message 19233)
Posted 12 Jan 2006 by crandles Post: The sulphur model is broken into 5 phases. Each time your computer passes 1/24 of a phase it sends back a trickle to say where the model is up to. This way you don\'t have to wait ages for credit. The science is uploaded at the end of each phase. You can see details of the trickles sent back if you drill down into your results. You have sent back 2 so far. Your computer took less time to do the second - maybe you were using the computer less. HTH |
©2024 climateprediction.net