I have read that it\'s normal to get work units that take 700+ hours to complete. I currently have 5 work units on my system, so with 2 CPUs, only two are running. Why would BOINC download so many work units to just sit on my hard drive? Shouldn\'t it wait until the first two are done, or at least only download one extra, instead of three extra? It seems a waste, like somebody else should be processing the work units that I won\'t even be starting on until weeks from now when the first two are completed.

There was a version of BOINC that would download extra models at startup sometimes. Try upgrading to the 4.45b version found here: http://arnaudboinc.free.fr/ or the 5.1.4 development version from berkeley.

---

BOINC WIKI

BOINCing since 2002/12/8

There was a version of BOINC that would download extra models at startup sometimes. Try upgrading to the 4.45b version found here: http://arnaudboinc.free.fr/ or the 5.1.4 development version from berkeley.

I joined ClimatePrediction.net a day ago, and received a sulphurn download.
After 24 hours, it appears it will take 3000 hours, or 125 days (4 months) to complete.

This morning I found a 2nd Sulphur download, which is also running.
Together they will take 8 months to compute. That\'s OK, deadline is 11/21.
I just want to ensure that I won\'t get a 3rd download, which will make all three miss the deadline.

However, I would really prefer to have just one project at a time, and see it through to completion before getting the next, since they are so large.

I\'m also running SETI (50/50). I have 10 short completed projects waiting to upload for days now -- lost one through error. I expect Sulphur will have glitches/ problems too, and fear losing weeks of computing time -- especially if my PC goes down. I believe you\'re \"trickling\" results, so not everything would be lost. But it\'s just safer to run one large project at a time.
REL

---

Trickles are just to let the server know that the model is still alive. There is no science data in them.

To prevent a catastrophic failure after months of processing, make a backup of the entire BOINC folder every so often. AFTER first suspending BOINC and waiting until it does stop. This will prevent the many data files in cpdn from getting out of sync during the fairly long copy process.

Restoring when running multiple projects is tricky, but is documented in the BOINC Wiki.
To prevent further downloads of data sets from cpdn, go to the Projects tab, click on a project, then click No new work.

After 24 hours, it appears it will take 3000 hours, or 125 days (4 months) to complete.

That not be an accurate guess as to how long it will take. Versions 5.2.x of BOINC can vastly overestimate the time to completion of cpdn models.