Hello, somewhere long ago I read that missing trickles eventually were regenerated, or re-credited, or something like that. Never happened to me until my CM3 model. When it seemed time for its 2nd trickle, I looked at the graphics to check and they were a bit scrambled; could see the globe but it wasn\'t in the right position and I couldn\'t get it to show the model progress etc. (I have never had graphics issues with CPDN before.) Nor did I see trickle info in the logs. I decided to restart the BOINC client; when it came back up everything looked fine, the model date was still well into 1922, and it sent its hello trickle. Now it\'s sent the 3rd trickle, butl ooking at the graphs they are blank for the period of the missing trickle. And the 2nd trickle wasn\'t credited, though that\'s secondary.

So I\'m just curious how this works; will it \"catch up\" the data and/or credit eventually? Or is part of this model\'s data missing, never to be recovered?

---

Since you are in 1922, you should have one or two credit-worthy trickles. Since the model started Dec 1 1920, you would get credit trickles on Dec 1 in 1921 and 1922 (once each model year).

The webpage listed credits is only updated once a day (something like 1000 GMT) so even after the next trickle, the webpage credits will not update immediately.

Thanks for the quick response and sorry, I meant to post the result to demonstrate the \"hole\" in the graph. That\'s what I\'m wondering about, will it ever get filled in?
5104426

The graphs data is way behind in getting added because of the huge load on the servers.

Also, are you running the new model?
There was a problem found recently, which resulted in ALL models being killed off, and new models sent.
New models start with: hadcm3lm

It looks like the information is currently missing, though you got the credit for it (3 trickles worth of credit at 226.xx each), which is correct.

I can only assume, but I would believe that this data will fill back in after the phase is done, once the larger data is sent back to the servers. There is a lot of data pushed up at the end of each phase, and I believe it should be able to fill it\'s own information back in at that time.

The trickles are just so that the system knows it\'s still alive (and provide some data), but those end of phase ones are the important ones.

---

Thanks for the helpful responses. Since this seems to be my day to display ignorance I\'ll ask one more: how many phases are there in cm models?

I just finished a seasonal model which had only one phase, and the graphics display for this cm3 model looks similar i.e. showing trickle x of y, where x/y = the same % complete shown on boincmgr\'s progress bar; to me that implies one giant phase. Nor did it show a phase designator, as it did in the slab and sulphur models I ran.

At least my current model passed the (small) amount of crunch time I\'d achieved on the bad model, so it feels like I\'m in new territory again. :)

There is only one phase. It\'s being described as a 70 (or 80) year hindcast to 1990 (2000) and the rest is a forecast to 2080. But for how it is listed in the display, it is being treated as one phase.