Is it just me, or are the credits a lot lower now than before the glitch? Before the glitch a successful HadAM3P PNW WU would get me around 3000 credits, now that is only 1000-1250...

I also had a HadAM3P with MOSES II WU that recently crashed after 1.3 million sec which only got me less than 1100 credits. A similar unit that crashed earlier after a similar amount of CPU time got me 8900 credits...

Not entirely sure the credit script has run properly since the glitch. Still, at 0900Hrs today there should be some activity in Oxford after the long weekend so once anything vital/more important has been dealt with we might get some update.

I suspect that the answer is:

Credit here is based on trickle_up files received.
Those of your models with lower credit, were given that from when the trickles-to-credits script last ran, and some trickle_ups from after the script last ran have yet to get any.
So those models, over all, only have part of their full credit.
The rest will come in the fullness of time.

The phantom credits have trashed the in Boinc cumulative stats by ruining the scale. I guess this will also resolve in a few weeks time when the glitch moves off the screen.

I suppose no credits are better than the whole system being down again...

Chris -

1. You can stop the BOINC Manager and all of the applications.

2. Edit the file statistics_climateprediction.net.xml in the BOINC folder.

3. Delete the group of lines with the obviously wrong info.

4. Save.

5. Re-start BOINC.

---

Hello,

Have been credited for 478795 points without crunching ...

I think you have the same problem as 2 weeks ago.

Kind Regards,

Sorry for "double post"

Now it get's strange:

As I said in the "Another Upload Problem"-Thread the zips of my latest eu-Task don't load up, only the small trickle files have uploaded.

However, I am getting points today. How that?

16726183 8954673 1329544 1 Aug 2014 7:43:32 UTC 14 Jul 2015 13:03:32 UTC Im Gange --- --- 2,386.39 2,386.39 UK Met Office HadAM3P-HadRM3P Europe v6.09

Folks, this entire CPDN-Project needs a serious rework: From completely disappearing web access to weekend service absenty to delayed credit updates to errant credit updates to various inaccessible upload-servers to errant jobs and on and on every day some problem seem to be solved and two new arise. It's is getting a monstrous task for us users to generally handle all these pitfalls.

Sadly it seems this project has gotten to a point where its structure has simply gotten too difficult for the available resourced folks to manage.

The issues are less a case of solving a problem and two new ones showing up than a case of solving a problem and then one or two OLD ones resurface.

Most of the issues reported over the past months (or in some cases years), are long standing repetitively. RAID failures, storage space issues, updating issues, scripts and the like. They appear to get patched and then the tape used to patch things up frays and breaks.

The only relatively new problem that has surfaced (and now resurfaced over the weekend) is the 'mega credit' reporting. Even it is now in its second appearance.

My sense here again is simply that this is a project which has grown far larger than the original design envisioned and that available resources are inadequate to the growth in size *and* complexity required to maintain it effectively.

I've supported this project with CPU's for over nine years.

---

As suggested by the two previous posts, maybe it's time to get a combined undergrad/grad project for the Oxford Computer Science department to recode a brand new infrastructure from scratch. It would give hands-on to the students and solve the real-world problems we're seeing with the project. As an added bonus, the code would probably be more efficient and easy to troubleshoot. Win-Win.

---

Michel
-----------------------------------------------
Always keep an open mind, even Einstein could not
explain how the universe worked,...

I think that the number of lines, currently in Fortran which would need recoding would keep a fair number of post doc coders busy for a few years. Plus the Met office who own the code would have to agree to release it.

Increased funding, enough to provide a half dozen extra full time staff would probably sort out most of the problems though, especially if combined with some infrastructure money as well.

Let me keep these credits! :0

hmmm
Increased funding?
more staff?

Flying pigs as well mayhap?

Bob (in cynical mode)

---

Hello,

Have been credited for 478795 points without crunching ...

I think you have the same problem as 2 weeks ago.

Kind Regards,

Sorry for "double post"

My Mistake, sorry !

Now the credit for the day = 6703133

If you don't amend this TODAY, we will AGAIN have a negative RAC for a while :/

:(

A half-dozen more techs is wishful thinking but two are possible. The first should be in-place in the foreseeable future but that person faces a steep learning-curve when he/she is aboard.

As to recoding, the basic model is about a million lines of Fortran developed by UK Met. Office scientists over approximately twenty years. Daunting to re-do? Then comes massive testing...

Techs are currently wrestling with current-model integration with a boinc server update. A further complication is coordination with remote project managers -- this is no longer a purely in-house project. And need I mention fun and games with hardware now located where CPDN techs have access only during normal work hours?

It's enough to keep two guys in near-constant fire-fighting mode, eh? It would be nice if they could devote full time to "the small stuff."

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

The money is here :

http://www.ox.ac.uk/about/organisation/finance-and-funding

� 50m budget surplus for 2012/2013

The money doesn't come from that budget, it comes from the researchers who give the work to Oxford to farm out to us.

I don't really mind the credit blips and the server issues. But I do hope that the actual code that we run is solid.

I volunteer because I would like to aid the scientific progress. And I would be very disappointed if it was in vain.

I don't think you need to worry about the code being solid. It has undergone a lot of testing and statistical analysis to check that out over the years.

My understanding is that we use the same model as used by the UK Met. Office for routine forecasts. (Different resolution and timelines, of course.) It is a highly-rated model -- I've been running it since July 2003, a year before the switch was made to boinc; not something I would have done if it seemed a dodgy undertaking.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

The only use for credits is for bragging.
So when there's a big jump like this, quickly grab a copy of your certificate.
Then you can tell everyone that you've got LOTS of credits, and you have an official document to prove it.

I've got 5,531,729,197,210 Cobblestones of computation on cpdn. :)