climateprediction.net home page
Posts by Arnaud

Posts by Arnaud

21) Questions and Answers : Unix/Linux : error: meaning ? (Message 25206)
Posted 20 Nov 2006 by Arnaud
Post:
Thanks Jim.

Reading more carefully the logs, I noticed that these messages appeared just after restarting the computer at the date 01/12/2056 and 01/12/2063.
Perhaps it is related with the yearly trickle files, but I can\'t tell exactly what\'s happened: perhaps the files were not properly closed when shutting down the computer.

And yes, I have a backup :o)
(I have read so much messages of BBC crashed models that I make a backup twice a week, now)

Cheers :o)
22) Questions and Answers : Unix/Linux : error: meaning ? (Message 25199)
Posted 20 Nov 2006 by Arnaud
Post:
hadcm3lbm_0187_05057730 - PH 1 TS 3706561 A - 01/12/2063 00:30 - H:M:S=3204:01:49 AVG= 3.11 DLT= 0.00
Error file dataout/0187fo.pjq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fo.piq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fo.pfq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fa.phq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fa.pgq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fa.peq3c10 has unknown file type
Error in getting file type
Error file dataout/0187fa.pdq3c10 has unknown file type
Error in getting file type
Trickling yearly means for 2063
Building file list for 2063 in dataout -- 7 entries
hadcm3lbm_0187_05057730 - PH 1 TS 3706993 A - 07/12/2063 00:30 - H:M:S=3204:18:42 AVG= 3.11 DLT= 1.23

Hi,
I started to monitor closely the log of my model as it\'s close to the end (I don\'t want it to crash just at the end :o))and I noticed this kind of error twice (2056 and 2063).
Does anybody know what this message mean ?
23) Questions and Answers : Unix/Linux : Linux Newbie Question: Best distribution for CPDN (Message 25180)
Posted 19 Nov 2006 by Arnaud
Post:
Ubuntu is OK (boinc is in the Debian repositaries, so you can install it via apt-get, AFAIK)
I\'m running CPDN on Mandriva and Suse 10 and it\'s working OK, too.
The only distro to avoid is Slackware: it seems there are a lot of problems with it.
24) Questions and Answers : Windows : Boinc manager is not able to connect to a BOINC client (Message 25167)
Posted 18 Nov 2006 by Arnaud
Post:
I have also tried the password located in the gui_rpc_auth.cfg file, again no success.

I must be doing something wrong somewhere. Maybe my configuration is incorrect

That should work, as the password required is in that file.
For remote access, you have create a file called remote_host.cfg containing the IP or name of the remote machines authorized to access BOINC.
Better explanations in the following threads:
http://www.boincsynergy.com/forums/viewtopic.php?p=44153&sid=a932072cebb8f1db81e46f61c72a404e#44153
http://setiathome.berkeley.edu/forum_thread.php?id=18507
25) Message boards : Number crunching : Preferences not transferring from here to my m/c (Message 25079)
Posted 11 Nov 2006 by Arnaud
Post:
It works on a 100 second cycle. So, for example you\'ve set 50% CPU. It runs at 100% usage for 50 seconds, then suspends itself for 50 second, then repeat...

I would guess the reason they\'ve done it with such a long interval is because there\'s too much overhead in the process (i.e., if it takes a second to get into suspend mode, then making the cycle quicker would increase the overhead). I\'d have preferred to see something measured in milliseconds...

--- Edit

I\'m not sure now, the above was based on the description I\'d read, but someone has now posted a conflicting description.


I\'ve tested 5.7.2 this afternoon and when setting the CPU at 50%, the CPU is suspended every second, so I don\'t think this version works on a 100 second cycle (perhaps the next ones...)
26) Message boards : Number crunching : Preferences not transferring from here to my m/c (Message 25050)
Posted 9 Nov 2006 by Arnaud
Post:
AstroWX said:
Martin, FYI: Seems I bucked before I was spurred. Waited a minute or two for \"Suspend network activity\" to take effect on each of the two boxes (long beyond reason in my opinion) before making my last post.

Wasn\'t long enough. Minutes ago, when I went back to check the Runs, thinking to re-boot to set things right, the Messages logs showed that both machines\' network activity suspended seconds short of five minutes after I entered the command. (That suggests some sort of time-out to me.)

I have also this problem after I update manually a project with Boinc 5.4.9
I have already reported this problem on the bug site but never had an answer, and I don\'t know if the problem had been corrected in the following versions (don\'t think so as I can\'t suspend network activity on my 5.4.9 linux machine with the 5.7.2 Manager)
27) Message boards : Number crunching : Limiting WU - can\'t finish what i\'m getting (Message 24999)
Posted 6 Nov 2006 by Arnaud
Post:
BBC and CPDN are giving the same model.
THe deadline is just an indication.
You can send the Wu after the deadline: It will be accepted by the server, credited and used by the science team.
Happy crunching :o)
28) Questions and Answers : Unix/Linux : computation error - cannot open input file dataout/atmos_restart.day (Message 24962)
Posted 1 Nov 2006 by Arnaud
Post:
Hi,
Perhaps, it\'s caused by a problem of permissions (can\'t open the file to write in it). So you could try to check that Boinc is not installed as root user and chmod -Rf 0755 the BOINC folder.
29) Questions and Answers : Windows : i have a pb with connection i\'m french and i don\'t understood alls writting (Message 24892)
Posted 27 Oct 2006 by Arnaud
Post:
Salut Cindy

Tu as deux problèmes:
1°) 63.46 MO de Ram ce n\'est pas assez pour faire tourner le modèle: il faut au moins 512 MO de Ram sous Windows.
2°) Les modèles CPDN actuels ne tournent pas sous Windows 98 (les modèles ne fonctionnent que sous Win 2000, XP ou Linux). Donc tu ne pourrais pas avoir de modèle même si tu avais suffisamment de Ram.

Je t\'invite donc à t\'attacher à d\'autres projets BOINC qui sont moins gourmands en Ram et qui supportent Windows 98.
Si tu tiens absolument a faire tourner CPDN, il faut rajouter de la Ram et mettre à jour ta version de Windows 98 vers XP ou éventuellement installer Linux.

(Thanks to Les Bayliss who found out the cause of your problems but said his French was too rusty to answer (but I doubt it...:o)))
30) Message boards : Number crunching : How to run CPDN on both cores of an Intel Core Duo T2300 ? (Message 24566)
Posted 5 Oct 2006 by Arnaud
Post:
Did you try to change the number of CPUs in your general preferences ?
(On multiprocessors, use at most: 2 processors)
31) Questions and Answers : Unix/Linux : Trickle 30 of 24 & 323.569 of 259.248 (Message 24276)
Posted 12 Sep 2006 by Arnaud
Post:
KBoincSpy is not designed for coupled models (just slab and sulphur) so the infos given by this program are wrong.
Sulphur and slab models had phases of 24 trickles and 259.248 timesteps, if I remember correctly
Coupled model has just 1 phase and trickles every 1 model-year, AFAIK
The timesteps are also differents because in the CM there are 72 timesteps/model-day whereas in the slab/sulphur models there were 48 timestep/model-day
I saw recently in the Boinc_dev mailing list that the creator of KBoincSpy asked for new xml tags in the xml files so that KBoincSpy can give good infos about CPDN models: perhaps the future version of KBoincSpy will be better with CPDN...
HTH
32) Message boards : Number crunching : Dumb Question? (Message 24023)
Posted 18 Aug 2006 by Arnaud
Post:
Hi, there are no dumb questions on this forum !!
First the model computes the atmosphere datas and then the ocean datas: the ocean datas are computed faster than the atmosphere datas that\'s why you think the increment is 1 hour.
But in fact, the increment (called timestep) is 1/2 hour for both the ocean and the atmosphere.
An extract from the Model introduction page:
\"The coupled model runs asynchronously, which means that the atmosphere model runs first for some time then the ocean model runs for some time, taking turns. In the case of the model used in the climateprediction.net experiment, the individual components run for one day at a time.\"

RAC is the average credit for one day. As it\'s a mean value, it\'s going up and down slowly, but it\'ll stabilize.
HTH
33) Message boards : Number crunching : Signature (Message 23933)
Posted 14 Aug 2006 by Arnaud
Post:

[tag]http://www.boincstats.com/signature/user_388548.gif[/tag] with tag=img
34) Questions and Answers : Getting started : Seasonal attribution project borked (Message 23859)
Posted 5 Aug 2006 by Arnaud
Post:
http://attribution.cpdn.org/create_account_form.php
Edit: it doesn\'t work...:o(, but I can create an account and attach via the BoincManager under Linux Mandriva:

sam 05 aoû 2006 19:55:44 CEST|http://attribution.cpdn.org/|Scheduler list download succeeded
sam 05 aoû 2006 19:55:44 CEST|http://attribution.cpdn.org/|Sending scheduler request to http://climateapps3.atm.ox.ac.uk/hadam3_cgi/cgi
sam 05 aoû 2006 19:55:44 CEST|http://attribution.cpdn.org/|Reason: Requested by user
sam 05 aoû 2006 19:55:44 CEST|http://attribution.cpdn.org/|Requesting 43200 seconds of new work
sam 05 aoû 2006 19:55:45 CEST|http://attribution.cpdn.org/|Scheduler request succeeded

If you want, you can contact me via PM (arnaud25 on the phpBB forum)and I give you the account I just created (you\'ll have to change the password, name and email address, of course)
35) Message boards : Number crunching : Signature (Message 23281)
Posted 21 Jun 2006 by Arnaud
Post:
Go in the forum preferences and in the \"Signature for message boards\" box:
Use [tag]http://www.boincstats.com/signature/user_177287.gif[/tag] by replacing tag by img
36) Questions and Answers : Windows : Changing resource share (Message 22985)
Posted 1 Jun 2006 by Arnaud
Post:
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/prefs.php?subset=project
37) Questions and Answers : Unix/Linux : Log file messages (Message 22953)
Posted 27 May 2006 by Arnaud
Post:
Well, the file account_www.climateprediction.net..xml shouldn\'t have 2 dots: it should be account_www.climateprediction.net.xml
Perhaps you can try to rename it and run the little program again.
38) Questions and Answers : Unix/Linux : Log file messages (Message 22950)
Posted 27 May 2006 by Arnaud
Post:
See here: http://www.climateprediction.net/board/viewtopic.php?t=4916
39) Questions and Answers : Unix/Linux : Another instance of BOINC is running (Message 22944)
Posted 27 May 2006 by Arnaud
Post:
Hi,
Usually, it\'s a problem of lockfile (delete it) or of permission/owner/group on the Boinc folder (chown and chmod it like your old Boinc folder with the recursive option)
40) Questions and Answers : Unix/Linux : No longer getting work units (Message 22848)
Posted 22 May 2006 by Arnaud
Post:
Hi,
It seems that you have Mac computers in your account.
Unfortunately, there is no work for Mac now: the coupled model only works on Windows and Linux and AFAIK, the slab and sulphur model are no longer distributed.


Previous 20 · Next 20

©2024 climateprediction.net