I\'ve read several times that a model which hasn\'t trickled for six weeks or so will be considered \"lost\". However, I wonder if this policy is still being enforced. If so, is there a way to tell from the model\'s status page if the model is still considered active?
I\'m wondering because I had no problems running this model although I had to leave it untouched for some nine months with not a single trickle (for reasons discussed here). It eventually errored out because of my running Sophos, but that\'s unrelated.
I\'m asking because I\'ve yet again downloaded one model too many; I\'m running a Dual Core now, and the scheduler found it wise to give me TWO models (here and here). I don\'t want to run them simultaneosly, but I intend to run the second one when the first one has finished. A possible workaround would be to start the second model every once in a while just for trickling, but I\'d like to avoid that as well as it would increase the size of the folder when backing up.
Thanks for your help!

The 6 weeks \'policy\' is still there. At least, it\'s [about that time.
It doesn\'t mean that you can\'t continue with it, just that another person will also be running it IF IT\'S NEEDED. Which it may not be.

The sulphur model that you mention may have been one of a faulty batch that was issued in December, that would have crashed at the start of phase 2 anyway.
But all of the important info in sulphur models was in the upload at the end of phase 1, so in that respect, your model succeeded. Those sulphur models were a \"limited edition\", that were run late last year/early this year. Their extra (sulphates), data was needed for the current models, which started in March. They are no longer required.

If you only intend to run one model at a time, then I would suggest that you go to the Projects tab, click the cpdn project name, then click the No new tasks button.
This will prevent the download of extra models.

Then go to the Tasks tab, click the model that has gotten the least amount of processing done, and click the Abort button.
This will \"return\" it to the project for someone else to work on. There will be plenty more for you when you finish the one that you\'re working on.

It should be possible for you to go to the preferences page on your Account page, and set the number of processors to 1, but there\'s something wrong at present which is preventing this prefs page from updating the Manager. (Only a problem on cpdn.)
If you\'re still running SETI as well, you can set it there, and it will propagate to this project soon after BOINC has contacted both sites.

Thanks for your prompt and helpful reply! One question remains, though.

Then go to the Tasks tab, click the model that has gotten the least amount of processing done, and click the Abort button.
This will \"return\" it to the project for someone else to work on.

Err ... are you really sure? The model in question was already re-issued to me after a first result of the same work unit had errored out. Aborting it would generate another client error, and with the total number of two errors the WU\'s status should change to \"Too many error results Too many total results\" (I think), cf. this WU nearby. Correct me if I\'m wrong!

It should be possible for you to go to the preferences page on your Account page, and set the number of processors to 1

Unfortunately, this is a global BOINC setting, not a per-project one, and I run work from other projects on the second core. But that might be helpful temporarily when it comes to getting ONE new CPDN model after the present one(s) have finished.

Err ... are you really sure?

Ignore my advice if you want to keep the extra model sitting there.
But there is a possibility that it WILL be run by someone else soon after the model has been flagged as not being worked on by you.

\"Too many error results Too many total results\"

This is one of many items that only apply to other projects, which use quorums to determine the validity of work returned.
\"Too many total results\", for instance, is because the server software is expecting the \"quorum\" number of \"results\" (1 for this project), whereas it actually receives the \"total number of trickles\" (160), instead. This being far in excess of the quorum number, it flags it as an error.

You should also ignore my advice to set the number of processors to 1.
I thought that you were saying that you received too many cpdn models, and that you only wanted to have one.

OK, I have aborted the second model just like you said. I just had the impression that it would NOT be re-sent because it already had been re-issued once, but obviously, I was wrong on that.

You should also ignore my advice to set the number of processors to 1.
I thought that you were saying that you received too many cpdn models, and that you only wanted to have one.

You have got me perfectly right there. I would like _BOINC_ to use both cores, but CPDN to use only _one_ of them (with one single model) so that other projects can be run at the same time ... call it instant gratification. As BOINC doesn\'t provide for that kind of setting (yet?), I\'ll have to use some manual workaround when a new CPDN model (one!) is to be downloaded in a few months; that\'s what I was trying to say.

Thanks!

Hi,

You have got me perfectly right there. I would like _BOINC_ to use both cores, but CPDN to use only _one_ of them (with one single model) so that other projects can be run at the same time ...

The simplest way to do this would be once you have a CPDN WU working on 1 core set climateprediction to \'No New Tasks\'

The simplest way to do this would be once you have a CPDN WU working on 1 core set climateprediction to \'No New Tasks\'

Yes, I did that, and I suspended the now aborted model when it was still downloading; it never started that way, and everything was fine ... apart from there being two models instead of one. That part of the problem is now solved by simply aborting the idle model, but it might arise again when the next CPDN download is due. Maybe by then, a version of BOINC is available that can handle project-specific CPU number settings (*); if not, I\'ll change to \"use at most one processor\" before the download and back to \"two\" (along with \"No new work\") after that.

(*) [edit] see this interesting discussion (item 2b) on the BOINC message boards [/edit]