climateprediction.net home page
Posts by old_user2467

Posts by old_user2467

21) Questions and Answers : Wish list : Gebrauchsanleitung in deutscher Sprache (Message 17177)
Posted 14 Nov 2005 by old_user2467
Post:
Hallo,

ich habe keine genauen Quellen, wo und ob es solche \'Handbücher\' in Deutsch gibt. Aber für den BOINC Manager könnten Dir evtl. auch die Erklärungen auf BOINC.de helfen. Für BoincView weis ich nichts.

Wenn Dir das noch nicht weiterhilft, so stelle einfach spezifische Fragen. Ich werde versuchen sie zu beantworten. BoincView benutze ich selber nicht, so kann ich Dir dabei nicht helfen.

Aber noch ein Tip für Dich: Benutze einfach das offizielle phpBB Forum von cp.net. Da tummeln sich viele Leute mit sehr gutem Wissen um BOINC und den AddOns. Ein Versuch ist es sicherlich wert! Ich benutze selbst nahezu ausschließlich das phpBB Forum.

Gruß
22) Message boards : Number crunching : New app? (Message 15662)
Posted 4 Sep 2005 by old_user2467
Post:
> I've just a minor problem with BOINC 4.45. After started for some reason it
> detects that it would take 2300 hours to finish the CPDN WU and reports the
> system is overcomited and start to crunch other projects by earliest deadline.
> In that case I've to suspend all other WU's wait until corrects the remaining
> time and continue to crunch for other projects.
>

I observed the same behavior with sulphur_cycle. Starting BOINC and the sulphur experiment BOINC reports that my machine is overcomited. But running only cp.net it has no opportunity to run workunits with earlier deadlines so it has to crunch sulphur.

Ciao
23) Message boards : Number crunching : http://phkup21.unibe.ch/boinc/file_upload_handler is dead (Message 14342)
Posted 13 Jul 2005 by old_user2467
Post:
The upload handler is working now.

Thanks again to Carl for notifying the folks at Uni Bern.

Ciao
24) Message boards : Number crunching : http://phkup21.unibe.ch/boinc/file_upload_handler is dead (Message 14328)
Posted 13 Jul 2005 by old_user2467
Post:
Hi Carl! Thanks for your answer and all your effort!

Ciao
25) Message boards : Number crunching : http://phkup21.unibe.ch/boinc/file_upload_handler is dead (Message 14326)
Posted 13 Jul 2005 by old_user2467
Post:
Hi... can not upload completed result to http://phkup21.unibe.ch/boinc/file_upload_handler. Do they know that their server is down at least till yesterday noon?

Ciao
26) Message boards : Number crunching : Attach allocated but not downloaded WU to client_state.xml (Message 13965)
Posted 27 Jun 2005 by old_user2467
Post:
Thanks John!

But attaching the "lost" result to boinc manually is not possible? No chance?

Ciao
27) Message boards : Number crunching : Attach allocated but not downloaded WU to client_state.xml (Message 13871)
Posted 26 Jun 2005 by old_user2467
Post:
Hi,

Two days ago <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=86867">one of my boxes</a> finished its last run and requested new work. It tried to contact the sheduler and finnaly it downloaded one new result <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=957769">957769</a>. But the result list of this host shows another allocated result <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=957688">957688</a> assigned two minutes earlier but not downloaded by this host.

There are no vestiges of this earlier allocated result and the message output hasn't shown anything that would give any sign that my box tried or failed to download this result 957688. It seems that this result was allocated to my box without any knowledge of this itselfe.

Is there any way to attach these allocated but not downloaded result to the client_state.xml of this host so it is shown as a "ready to compute" workunit?

Ciao
28) Message boards : Number crunching : Main forum down? (Message 13864)
Posted 26 Jun 2005 by old_user2467
Post:
Well explained. Thanks!
29) Message boards : Number crunching : Server/Trickle Problems Being Worked On (Message 13832)
Posted 24 Jun 2005 by old_user2467
Post:
&gt; Now that it's the weekend in the UK....

Well, not only in the UK ;-)

&gt; ... I suspect this will not be fixed until next week.

Yes this would be realistic.

&gt; Oh well, I still crunch, and am patient.
Its the best way to keep cool at these high temperatures out there: Not to excite themselves and to retain a cool head.

Ciao and a nice weekend!
30) Message boards : Number crunching : Linux vs. Windows crunching speed (Message 13831)
Posted 24 Jun 2005 by old_user2467
Post:
geophi wrote:
&gt; The fastest AMD PCs are generally in the 1.6
&gt; sec/TS range, while the fastest P4s when running only one model, with HT off,
&gt; are closer to 1.3 sec/TS (in Linux).

I think your estimated fastest speed of the AMD section with 1.6 sec/TS at its best isn't very realistic. I use a moderate overclocked AMD Athlon64 3000+ Winchester currewntly at 2208MhZ with DDR400 Infineon Memory and I got about ~1.9 sec/Ts on average. So I would say already AMD Athlon64 3800+ or 4000+ with its 2.4Ghz should reach not overclocked an processing speed of approx ~1.7 sec/TS. Since one can also ocerclock these processors very well their expected speed should reach 1.5 maybe 1.4 sec/TS. I think its not so unrealistically.

Fourthermore one should keep in mind the much lower power consumption and therefore the much lower temperatures of the AMD section. In my opinion a very important issue towards the efficiency in office use.

Ciao
31) Message boards : Number crunching : This good or bad? (Message 13799)
Posted 24 Jun 2005 by old_user2467
Post:

&gt; Can you tell us what error messages you are getting? Is it always the same
&gt; ones?

Hi Andrew,

its rather difficult for the moment because this box is located elsewhere. On tuesday I can have a look at the stderr.txt. But I don't know how long informations about crashes are keept in this logfile. But to have a look at <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/results.php?hostid=22924">its resultpage</a> one can find that they seem to have reported the exit code -5.

But I have a self written program to monitor my working boxes and their modelstates. Maybe I have the time to add an selfdiognostic eventlog which is reported to my server application in such cases. But this needs some time.

Ciao
32) Message boards : Number crunching : Server/Trickle Problems Being Worked On (Message 13794)
Posted 24 Jun 2005 by old_user2467
Post:
&gt; I have a model that is 7 hrs away from completion but it has only 23
&gt; trickles as far as I can see. Is this normal? Am I misunderstanding this.

Hi,

you are not 7 hrs away from completing your whole model but only 7 hrs away from completing phase1 of your model. Each phase has 24 trickles. I read elsewhere that there is a bug in a one of the newer boinc version (maybe 4.43?) that is showing modelprogress of one cp result only to completion of each phase. If you get through phase 1 and entering phase2 Boinc will show that you are at 0% through the model.

As I read there periodically it is recommended to update boinc to a more stable version. But I don't know what version this would be.

Ciao
33) Message boards : Number crunching : This good or bad? (Message 13792)
Posted 24 Jun 2005 by old_user2467
Post:
One of my boxes behaves not more normal as it did before. After downloading a Hadsm_4.12 modell it continuously chrashes hadsm_4.12 and after doing this some times the the whole result crashes. This box completed already one run successfully with a lower version of Hadsm (I believe to remember that it was 4.04).

All the system equipment is the same as on my other hosts (WinXP pro SP2, Norton AV 2004, BOINC 4.19, no other AV, malware, spyware, or something else). I will watch at this attentively and if things doesn't getting normal, unfortunately I will have to detach this box from cp.net.

Ciao
34) Message boards : Number crunching : Server/Trickle Problems Being Worked On (Message 13732)
Posted 22 Jun 2005 by old_user2467
Post:
The first missed trickles are shown recently on my account but still total credit is wrong.

Ciao
35) Message boards : Number crunching : 7 days no trickles (Message 13717)
Posted 22 Jun 2005 by old_user2467
Post:
Maybee it was replaced in a newer version of hadsm. I don't know whether it really was and when in what version it was done.
36) Message boards : Number crunching : Has someone contacted Oxford about the credit outage? (Message 13646)
Posted 21 Jun 2005 by old_user2467
Post:
&gt; Where is that Community forum ...?

<a href="http://www.climateprediction.net/board/index.php">look here</a>
37) Message boards : Number crunching : BOINC cp.net Teams pages aren't working anymore (Message 13599)
Posted 20 Jun 2005 by old_user2467
Post:
Its not true that team pages aren't working. Try to watch at <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=21">this teampage</a> and it will work!

Ciao
38) Message boards : Number crunching : How long to get 1 trickle on 500 MHz Celeron? (Message 13516)
Posted 17 Jun 2005 by old_user2467
Post:

&gt; I believe CPDN gives a full year to return a result.

I have also one very slow approx 10 years old Pentium I 200Mhz box. Its to slow to work for cp.net. But I had running it for Seti and now for Einstein. Crunching one workunit for Einstein needs approx 160 hours or nearly 7 days which lies a bit under the report deadline of 7 days. But as this box is serving also as an internet gateway and as HTTP server and is serving many other tasks it happens that the report deadline of some Einstein result is exeeded. But as so long there have at most two other boxes finisheded their result for the corresponding workunit nevertheless I got the credit for my work and the result is recorded as normal. But if there have already 3 boxes reported their results my result is seen as over and no credit is given.

But as cp.net in the normal case alocates each workunit only to one host there will be no other hosts that activate the report deadline. So I asume that also after report deadline the reported and finished result will be recognized as finished and credit will be given. But cp behaves in many things not as normal as one might think. So perhaps it will react in a completely different way. So one has to try out to say more reliable.

Ciao
39) Message boards : Number crunching : Announcement: Database residual problem - misallocated WUs (Message 13172)
Posted 6 Jun 2005 by old_user2467
Post:
&gt; With the exception of result 677468 (which I can't explain either) your
&gt; credits now look to be right
&gt;
&gt; Result 872328: 29 trickles, 2741.01 credits
&gt; Result 780535: 35 trickles, 3308.11 credits
&gt; Result 698345: 61 trickles, 5765.57 credits


As you can see all hosts have trickled one times since my first post. Fell free to have a look at these results this evening or tomorrow, when some host have trickled another time but credit is not granted till one further trickle gets in.

It is not so much problem to me as it seems that things are corrected when further trickles arrive and credits are not much worth then the science behind. But if one result would be briefly before finish I would feel better if all things run normal as they should. You may feel different but I have the strange feeling that something isn't running correct and we will hear more complaints.

Ciao
40) Message boards : Number crunching : Announcement: Database residual problem - misallocated WUs (Message 13163)
Posted 6 Jun 2005 by old_user2467
Post:
&gt; 780535 trickled last 5 June 21:37 hmm - time difference? doubt it.

No time difference... utc+2 hours... so after at least 4 hrs it had should been fixed. But it will be fixed (hopefully) if another trickle by this host is send in. But this behaviour is not normal in the meaning that this wasn't normal before the server software was updated.

And an update to the last post:
Result 872328: 28 Trickles and 2551.97 of ~2646,56 credits

Host <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/show_host_detail.php?hostid=86867">86867</a> has trickled this morning (in europe) and now the trickle that was not credited till yet is now credited in this way.

So I don't know. Is it possible that the validator or sheduler hangs some time back or that some reported trickles are lost and if the next incremental trickle is handed in the validator gives credit to the highest known trickle?

Ciao


Previous 20 · Next 20

©2024 climateprediction.net