The credits currently appearing, are those created by the BOINC software, which isn't correct. For this project, credit is based on returned trickles, but BOINC occasionally, for some computers/models, also gives credit for the completion of a model. Credit for completed work is only supposed to happen with other projects.

Back in 2004, when this project switched from it's pre BOINC beginnings to using the BOINC concept, this extra credit caused lots of screaming from those people not getting it, and it came to be known as "The random credit lottery".

This is why Carl came up with the idea of recalculating credit, first almost constantly, then hourly, and eventually daily, as the amount of stored data grew into gigabytes. Currently, the script is taking well over half a day to run, and was tying up the server in the process.

The current idea is for the script to be run on a backup copy, and for the data to be linked back to the export scripts on the main server.
This was tested a few weeks ago, with the data being manually copied back to the main server.

But LOTS of files are being created, and the new people aren't sure which are still needed. (Apparently, different stats sites use different data structures for their import files.)
And as you may know, the two new project people are up to their rapidly receding hairlines at present, trying to get the next stage of the RAPIT/RAPID models working. Probably with the researchers screaming at them from a few feet away.


So, yes, it really IS difficult/hard/time consuming, to get the credit export files created.

---

Backups: Here

Interesting - Only a 1/10 chance of getting a hadcm3n. I must be lucky as I have bagged 2, the second one when an eu model crashed on me.

Dave

Not sure,
I know I was unable to merge an old CPDN id with my current one. - It may depend on whether you still have access to the password or the account key for your old account. I didn't have either and also no longer had access to the email I used at the time.

I know I have made a contribution to the project that predates my present id and for me that is more important than the stats.

Dave

Lex

There's no way to 'merge' the two.
That would have had to have been done when the BBC CCE project was still active, because after you make the change, you then need to wait until the next export file is created, and for the various stats sites to pick these up and incorporate the data.
The BBC project was officially closed down at the end of 2007, since when it has been a classed as a Retired project by stats sites. This means that they no longer look for data from them.

The BBC CCE server was kept for a few years, but has now been turned off. All model data was moved elsewhere some time before this.

And posting your email addresses on web sites is just what spammers love. So good luck with that too.

---

Backups: Here

On the 1/10 chance of hadcm3n tasks:
I must have been very lucky then; on my last request for work (about 2weeks ago) I only got hadcm3n tasks! Just asked for 5 on one system and all 5 were hadcm3n.
One has completed, one is at 68% (running) and the others are between 20 and 40% - Now just running between 2 and 3 at a time; the returned task lost 14h CPU - runtime delta. It's not a particularly well CP optimized system, yet, but that will hopefully change for many in due course... 'Are we there yet' time is around 313h for each task. i7-2600 8GB, only using 2nd drive for Boinc.

GL

---