For the job camb_legasy given very little credit. Perhaps we should return to the old system of accrual loans when the assignment was given them 420

For the job camb_legasy given very little credit. Perhaps we should return to the old system of accrual loans when the assignment was given them 420

There may be two issues here. The first is that for some reason, after the upgrade there was a number of camb_legacy jobs which were awarded 111.11 credits. I plan to go back and up these to 420. The second is that we've now switched to the new credit system, which is awarding varying credits per job, but which are all the same as before within say +/- 50% . I'm apt to keep the new system, as it was put in place by BOINC people who have thought about the fairest ways to award credits much more than I have.

"As the boss wants it, so you must do it or act"

:)

I'm apt to keep the new system, as it was put in place by BOINC people who have thought about the fairest ways to award credits much more than I have.

Well, CreditNew is buggy and messy, it's nowhere near to the fairest credit system as you believe. On almost every project it is set up it is causing trouble after time, giving unreasonable credits. That's why most projects change to fixed credits.
In addition, you use quorum 1 here, so a result is never compared to another. That's the easiest way to cheat the system by manipulating the benchmarks of your own computers. It would be harder if you would have either a higher quorum or change back to fixed credits.

And one question:
Is camb_legacy the same app as the old CAMB before? I'm asking because WUprop is counting the hours for this app.

---

Life is Science, and Science rules. To the universe and beyond
Member of BOINC@Heidelberg
My BOINC-Stats

Well, CreditNew is buggy and messy, it's nowhere near to the fairest credit system as you believe. On almost every project it is set up it is causing trouble after time, giving unreasonable credits. That's why most projects change to fixed credits.

Fair, I've gotten a lot of negative feedback about the new credit system. I'm looking into it.

In addition, you use quorum 1 here, so a result is never compared to another. That's the easiest way to cheat the system by manipulating the benchmarks of your own computers. It would be harder if you would have either a higher quorum or change back to fixed credits.

We've got a method of validation that doesn't require inter comparison.

And one question:
Is camb_legacy the same app as the old CAMB before? I'm asking because WUprop is counting the hours for this app.

Yea, camb_legacy is the identical app. Can you point me to one of these results?

Yea, camb_legacy is the identical app.

Okay, then I don't need to "complain" over at WUProp. ;-)

Can you point me to one of these results?

You mean the legacy ones?
Here are two for example: 34245254 and 34245258
No particular "complaint" about them, although for having it crunched with a brand new AMD FX 8320 these are a bit low credits - well, that's CreditNew, I can tell stories about it...

Oh, it's probably a bit off topic, but I had a camb_boinc2docker-task running on my other host which crunched endlessly and never finished. I aborted it after a while just to find out that the next task also never really finished. It took me a while before I checked cpu-usage and memory, then I noticed that one of the vbox-apps (from the first one probably) never really quit, but none of my processors were working on it. Seems it had put the entire process to a halt and new tasks couldn't really crunch.
Here's the link/report file to the "culprit", which has over 15 hours before I could cancel it: 34251176
But it doesn't look like it says anything about the cause of the problem...

---

Life is Science, and Science rules. To the universe and beyond
Member of BOINC@Heidelberg
My BOINC-Stats

Here's the link/report file to the "culprit", which has over 15 hours before I could cancel it: 34251176
But it doesn't look like it says anything about the cause of the problem...

Thanks. A few people have reported similar problems, I think. Can you try upgrading to the latest Virtualbox 5.X and see if that solves it?

I mean, just as the credits awarded are now in march, it's okay. So that they are given depending on the computing time and computing power.
In particular, because the run times vary greatly.

In the same way it does with boinc 2docker.

But recently the run times of planck_param_sims vary extremely by factor 20. Here now you would have to change on this variable Credit System as done for CAMB legacy an oinc2dockers.