climateprediction.net home page
Posts by mo.v

Posts by mo.v

21) Message boards : Number crunching : Credit updates? (Message 48615)
Posted 29 Mar 2014 by Profile mo.v
Post:
Could you please tell us whether you have seen your credits updated on any of the external stats sites?

Jonathan has been reconfiguring the entire credits collection process(es) but as far as I know the job isn't yet finished. He's been increasing the security of the process and redesigning it. It's a very big job because CPDN awards credits on a different basis from other projects.
22) Message boards : Number crunching : Credit updates? (Message 48605)
Posted 28 Mar 2014 by Profile mo.v
Post:
JS, I've just replied to your credit query in your thread in the Getting Started section.

In your CPDN account you can look at your computer and your models to check on the trickle progress of each task. The trickles show your computer's in contact with the server.
23) Questions and Answers : Getting started : Unusual forum options (Message 48604)
Posted 28 Mar 2014 by Profile mo.v
Post:
I've just posted about the credits update matter in another thread.

JS, on this forum keep an eye on the News thread which is right at the top of the Number Crunching section. That's where the moderators post news items. You can subscribe to that (and any other) thread by pressing the Subscribe button at the top to receive an email alert every time a fresh post is made there.
24) Message boards : Number crunching : Trickles not updating credits (Message 48603)
Posted 28 Mar 2014 by Profile mo.v
Post:
The credits haven't been updated on BoincStats for more than two weeks but we don't need to worry because the CPDN programmers always sort out this sort of thing eventually. We do, however, need to be patient!
25) Questions and Answers : Windows : Tasks are running but not trickling (Message 48595)
Posted 28 Mar 2014 by Profile mo.v
Post:
At the moment (well, for the last month) our credits haven't been notified to BoincStats. Don't worry about that because when there's a credit delay we always get the credits eventually. As long as your models are crunching and trickling and uploading their files you're OK.

On a laptop the most important thing to think about is whether the machine can crunch the number of models you want it to without overheating. If necessary download CoreTemp or another similar gadget to check the CPU temp.
26) Questions and Answers : Windows : Graphics display error (Message 48592)
Posted 28 Mar 2014 by Profile mo.v
Post:
Here's an active link to Shankasoarus's defective graphics:

http://imgur.com/YJvoEts

If you have graphics resembling these for any model type please post.

At first I thought these incomplete graphics could be related to some defective graphics we've recently had for Beta models (which the programmers know about) but the defects are not the same. I don't think it's the same phenomenon because on Beta all the crunchers AFAIK see the defects.
27) Questions and Answers : Windows : Graphics display error (Message 48587)
Posted 28 Mar 2014 by Profile mo.v
Post:
My computers, both Win 7 64-bit, have normal graphics for ANZ models downloaded in the last few days.

Has anyone else got text missing from ANZ or any other type of model graphics?

28) Message boards : Number crunching : New tasks, Charts (Message 48440)
Posted 18 Mar 2014 by Profile mo.v
Post:
http://ob.cakebox.net/cpdn_status/server_status.html

I like this. Usually we can't see how long a batch of models takes to be sent out. I think we should sticky it but not yet because it's quite easy not to see new stickies up at the top.

Thank you, Tullus!
29) Message boards : Number crunching : Model crashed: REPLANCA: Current time precedes start time of data (Message 48439)
Posted 18 Mar 2014 by Profile mo.v
Post:
At least these models aren't consuming much time or electricity.

When we see crash reasons for climate models spelled out in capital letters - REPLANCA, not Replanca or replanca - I now assume this follows some Met Office convention (they wrote the code) and means the cause is probably intrinsic to the model.

Not that all model defects produce an explanation in upper case. They don't.

One of my models waiting to be crunched from a completely different batch crashed on another computer with the stderr line:

Model crashed: READHIST: End of file in READ from history file for namelist NLIHISTO

I've no idea what this means but I'm assuming that the capital letters indicate a model with a different intrinsic defect.
30) Message boards : climateprediction.net Science : First Results in from Weather@Home 2014 UK Flooding Experiment (Message 48438)
Posted 18 Mar 2014 by Profile mo.v
Post:
Hello Hannah

I haven't yet fully understood what the light green dots represent. Is each one of these dots the sea surface temperature value for one model? As the models only run for one year, I presume the SST values do not rise in response to global warming even in the winter as observed? Or is there an upward temperature creep even during the single year?

Is each light green dot a single value or a pattern? I consider that a pattern should include a range of values.

How many light green dots are there for each model? Two perhaps - one for the global SST and one for the regional SST?

Sorry if these questions are way off base.

I haven't understood either how the attribution (or non-attribution) of the heavy rainfall to GW can depend on the position of the blue rainfall dots with regard to the light green temperature range. Isn't this a bit like saying it depends on whether the rainfall lies within a particular temperature range?

Surely the light green dots represent a different graph altogether, except that we have no vertical scale for it. It seems to me that here we have two graphs, one for rainfall and one for temperature, superimposed on one another; where the two graphs lie in relation to each other on the vertical axis looks to me arbitrary. Or are these positions constrained and determined by the results of lots of other climate models and real-world observations?
31) Message boards : Number crunching : Credit updates? (Message 48436)
Posted 18 Mar 2014 by Profile mo.v
Post:
free-dc gave me quite a few credits yesterday, which must mean 17th March, but on BoincStats I/we are still waiting. This must be because of the different times of day that each stats site imports its updates from the projects.

Qax, I hope your students have been thinking a little about the value we attribute to intrinsic and extrinsic rewards. I also hope they've all got successfully running climate models.
32) Message boards : Number crunching : Model crashed: REPLANCA: Current time precedes start time of data (Message 48433)
Posted 18 Mar 2014 by Profile mo.v
Post:
For anyone reading this with 1997 EUR models created on 17 March 2014 it would be worth doing the same as Dave: force them to run immediately by suspending other tasks. As they will almost certainly crash after a few seconds you will then be more likely to receive work when the next good batch of models appears.
33) Message boards : Number crunching : Model crashed: REPLANCA: Current time precedes start time of data (Message 48430)
Posted 18 Mar 2014 by Profile mo.v
Post:
I had four 1997 EUR models downloaded on 17 March. All crashed at exactly the same moment (1m21s) with REPLANCA.

As soon as they'd started I tried to open the graphics globe of one of them to see whether there was anything of interest. The attempt to open the graphics window failed (the window was just an outline filled with black) but it had a terrible effect on BOINC Manager. All visible progress by other normal tasks in the Tasks pane froze and then the whole Tasks pane (or maybe the whole of BM, I can't remember) greyed out. After the four tasks had crashed, BM returned to normal and the graphics window showed a blue globe with zero crunching recorded.

REPLANCA gives BOINC a bit of a fright.
34) Questions and Answers : Windows : screen saver blanking repeatedly while climateprediction runs (Message 48415)
Posted 16 Mar 2014 by Profile mo.v
Post:
Years ago with a computer that had a much slower CPU, far less RAM and a rather antique graphics card I found that activating the screensaver frequently caused the whole computer to freeze. I also found that when I activated a model's graphics in BOINC Manager it was fine as long as I kept the model graphics window small, but if I maximised the graphics window I immediately got the same problems as with a full-size screensaver. I concluded that on my computer at least the problem lay in the size of the climate model's graphics.

The graphics for these models do seem very intensive when displayed full-screen; I'm not sure whether with RAM or the graphics card requirements or both. Could you compromise by doing the following: deactivate the screensaver, then open the graphics window in the Tasks pane of BOINC Manager. Don't maximise it but leave it open. Whenever you want you could leave it displaying on your desktop, though of course at reduced size.

The screensavers for other projects usually don't display so much real-time data (sometimes almost none) and are therefore easier to run full-size.
35) Message boards : Cafe CPDN : Scotland team (Message 48130)
Posted 9 Feb 2014 by Profile mo.v
Post:
I'm glad I have to log in afresh to my bank every time. Otherwise if a cookie logged you in and the computer was stolen the thief would be able to get straight into your bank account.

But I find that with Windows cookies keep me logged into all my project and forum accounts. Except yesterday when I used Windows Disk Cleanup to have a total clean sweep; I now find myself logged out of a lot of sites, though not all.
36) Message boards : Number crunching : VANISHING WU'S (Message 48063)
Posted 27 Jan 2014 by Profile mo.v
Post:
You are right; it's midnight servertime, not usertime. Sorry for the mistake.

I believe that after the server's midnight, computers are allowed to request new work at a random number of minutes during the next hour. I'm not sure how this affects CPDN as our computers can only request work once per hour anyway. I imagine it means the work quota is reset at the computer's first server contact after server midnight.
37) Questions and Answers : Getting started : "Communications deferred" ... continuously (Message 48062)
Posted 27 Jan 2014 by Profile mo.v
Post:
I agree fully with Jim. If you look at the Tools menu in Boinc Manager and click on 'Add project' you can mouse over projects in the list, see what they do, and choose the research you think is most worthwhile. There are more Boinc projects than the ones that appear there, but these are selected by the Boinc programmers in Berkeley as being safe and secure.

If you set a fairly short work buffer eg 1 day for extra projects it means that when CPDN jobs appear again you can set the other projects to No new tasks and then increase the work buffer. That should maximise your chances of getting what you want.
38) Questions and Answers : Getting started : "Communications deferred" ... continuously (Message 48049)
Posted 25 Jan 2014 by Profile mo.v
Post:
Hello Alan and welcome to the forum.

At the moment there are no tasks available for download. Keep an eye on the Server status page (to which there's a link in the menu on the left side of this page) to see what there is.

A few isolated tasks are being made available if they've crashed on another computer. Don't spend time on these if the workunit they belong to is marked No resubmission and the workunit number is hadcm3n7.... as they're defective. They have to be aborted.

It's also a good idea to keep on the forum News thread which is at the top of the Number Crunching section. You can subscribe to it and receive an email to alert you to new posts made there. That way you'll know what's going on.
39) Message boards : Number crunching : Server Status (Message 48041)
Posted 24 Jan 2014 by Profile mo.v
Post:
The SS page is now updating.
40) Message boards : Number crunching : Server Status (Message 48039)
Posted 24 Jan 2014 by Profile mo.v
Post:
Dave said:

'I did notice it seemed stuck yesterday, I should have known someone would have been on to it without my posting anything!'

That's not a reason not to post! You may well be the first or the only person to notice something. If problems are posted on the forum it often reassures other members who worried or puzzled about what's going on.


Previous 20 · Next 20

©2019 climateprediction.net