Is there any reason why my model is still running but no credits have accrued for over a week? It seems that I got a new model about 8 days ago but the CPU time didn\'t reset itself to zero. Could this have something to do with it?

Are any tasks running in BOINC Manager? It may be that your slab model (here) has got into a loop or is running slowly. It hasn\'t trickled since 10th September.

If you display the temperature graphics, is the globe blue? And what is the date shown?

Are any tasks running in BOINC Manager? It may be that your slab model (here) has got into a loop or is running slowly. It hasn\'t trickled since 10th September.

If you display the temperature graphics, is the globe blue? And what is the date shown?

Yes I\'m running Rosetta, Predictor and Einstein in BOINC Manager. All have sent in results after 10th September. It does seem odd that my slab model is showing 9.863% completion and yet \"CPU time\" of 271:55:40 and \"To completion\" of 542:12:38. I would have expected the completion % to have been much higher with those figures. The temperature graphics seem fine (most of my globe outside the polar regions is brown or red)and the date shown is 10/05/1815.

Does it look like a \"reset\" is the answer here? And thank you for replying so quickly to my message.

Don\'t do a reset just yet. IIRC the % complete refers to the % complete in the particular phase its crunching. There are 3 phases in the slab models and you\'ve completed phase 1, so it could be 9.863% of phase2 which would put it over a 1/3rd of the way through the entire model.

---

_________________________________

Don\'t do a reset just yet. IIRC the % complete refers to the % complete in the particular phase its crunching. There are 3 phases in the slab models and you\'ve completed phase 1, so it could be 9.863% of phase2 which would put it over a 1/3rd of the way through the entire model.

Yes, in that case it would be silly to reset. If the work is continuing then I don\'t mind not seeing the credits go on to my account. It would be nice to actually finish a model. It\'s certainly worth waiting to see if it corrects itself. Thank you again.

You get credits for each trickle returned.
You haven\'t returned any trickles since September 10, which is why you haven\'t gotten any credits since then.

If the trickles are being created but not being uploaded, then they will accumulate in the cpdn folder, with names starting with: trickle_up_

If there aren\'t any, then your model is probably \'looping\'. See the post in this README on Looping models for more info.

---

Backups: Here

This does look rather strange:

- the date of May 1815 suggests phase 1

- that date is consistent with 9.9% overall completion

- trickles have been received for all of phase 1

- it took 759,901 seconds for phase 1 (~211 hours)

- the 271 hours reported is roughly 211 + (9.9% x 3 x 211).

So, the numbers are consistent with the following odd interpretation: the model has restarted phase 1 having completed phase 1, with the CPU time accumulating. If this is right, the trickles are being repeated and therefore not credited, and none will be until phase 2 starts - if it does!

Are trickles being recorded in the messages tab? On this theory, they should be going out as normal.

I\'ve heard of this behaviour several times with the Slab models, perhaps they\'re more sensitive during post-processing (particularly given that they do it more often than the other models!).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

I\'ve heard of this behaviour several times with the Slab models, perhaps they\'re more sensitive during post-processing (particularly given that they do it more often than the other models!).

Trickles are (now it\'s were)being recorded on the message tab, which did confuse me a little as from my point of view all was proceeding nicely but nothing was happening at the other end. I decided to start another model which is disappointing but I felt it had to be done.

I\'d like to thank you all for your help and God willing we won\'t have to talk to often!

Kind regards.

Duvet

I\'ve heard of this behaviour several times with the Slab models, perhaps they\'re more sensitive during post-processing (particularly given that they do it more often than the other models!).

This occurred occasionally in the olden days of slab. I had one of the new slabs do this as well, but it wasn\'t at phase change. I had a power outage that shutdown the 4 computers at home that were crunching. Only one model on one core of one PC had a problem, and that was about halfway through phase 1 where the timesteps reset to 0 of phase 1. The trickle listing for that model. I\'m not sure if this occurs because the shutdown happens exactly during a disk write to the client_state.xml file but it was, and is frustrating when it happens.