time over - cputime - claimed - granted

23 Aug 2006 16:27:12 UTC - 2,732.33 - 9.64 - 9.93
23 Aug 2006 14:13:55 UTC - 5,004.38 - 17.65 - 9.58
23 Aug 2006 19:45:31 UTC - 4,316.08 - 15.22 - 8.82
23 Aug 2006 19:15:59 UTC - 2,642.31 - 9.32 - 9.84

All results from the same system.
So I become more points from a job that has used less cputime ?

Stranger, It doesn't matter how much cputime I spend on a job, the outcome is almost the same granted credits.

I'm not amused and if this is the outcome of the new creditsystem, I don't believe that I will crunch much longer for rosetta.

---

Claimed credit was changed to Granted Credit, Granted credit was changed to "granted work credit" the new system. You're still getting credit via the old system, or what you show as "claimed credit".
Eek, I see they haven't change the column title here yet. They did on ralph. Perhaps they should change the titles to avoid confusion???

very confusing !

---

I'm confused too. did they just adopt the new system? Perhaps I better go read the science board or something.

Anyone one know for sure???

It now appears the same at Ralph as it does here, and all the extra columns which showed "Granted Work credit" and "average granted work credit" are GONE

The new system isn't live yet. . . it's still using the old credits.

Edit: I was incorrect in this post, I hadn't checked my email. My apologies.

---

Some models will give you above-average credit while others give below-average on the same system (not every model takes the exact same amount of time)... it's supposed to even out over time.

You can probably decrease those fluctuations on a per-WU basis by increasing the runtime, right now most of your WUs only get to run one model. However it's going to even out over time one way or the other.

boincstats and other sites use the 'old' credit, so the new credit system doesn't affect your ranking there.

---

i think it's just the headings that need changing...

---

i think it's just the headings that need changing...

that's what I thought, look at ralph, All the new columns are GONE (granted work credit, and RAC granted work credit)

The new system isn't live yet. . . it's still using the old credits.

I challenge that. What's being exported to the stat sites, appears to be the new, smaller numbers.

I think it is more than just "wrong column headers".

---

Reno, NV
Team: SETI.USA

From David Baker's thread yesterday:

David Kim will be posting the definitive description of the new system tomorrow on the boards.

So - tomorrow is here - where's the "definitive description"?

---

Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)

Some models will give you above-average credit while others give below-average on the same system (not every model takes the exact same amount of time)... it's supposed to even out over time.

Sorry, but I don't accept less credits for almost double the CPU-time on the same system which is doing nothing else then crunching.
I don't call that a fluctuation but a misbehaviour of the creditsystem.

---

The new system MUST be in place.



I just recieved an increase of 27.77 credits after sending in ONE wu. from 40726.44 to 40754.21. IF it were still Benchmark based, I would have gotten the claimed credit value of 25.35.

tony

I'm pretty sure that the thread titles aren't what's in question here. It's the fact that he is getting/would be getting 9.58 points for a WU that ran 5,004.38 seconds and then 9.84 points for a WU that ran 2,642.31 seconds under the new credit system. Getting slightly higher points for running a WU a bit longer than half as long?

No Thanks

---

On this WU you get almost twice the normal credit: https://boinc.bakerlab.org/rosetta/result.php?resultid=33986546

The only reason those anomalies "pop" out is because of the short WU run time - in one case you probably got a particulary complex model, and the other one aborted early because it didn't lead anywhere.

---

The new crediting system went live since yesterday. I am working on a brief description to post on the technical news section. I will also post on the boards.

23 Aug 2006 22:57:36 UTC

The new system isn't live yet. . . it's still using the old credits.

...and then less than 2 hours later...

24 Aug 2006 0:21:56 UTC

The new crediting system went live since yesterday. I am working on a brief description to post on the technical news section. I will also post on the boards.

*Boggle*

How do you expect to be able to communicate with us effectively, if you cannot with each other?

---

Reno, NV
Team: SETI.USA

On this WU you get almost twice the normal credit: https://boinc.bakerlab.org/rosetta/result.php?resultid=33986546

The only reason those anomalies "pop" out is because of the short WU run time - in one case you probably got a particulary complex model, and the other one aborted early because it didn't lead anywhere.

Another problem, I can't edit my prefs anymore, as some usersadvised I tried to set it from 1 to 4 hours, but after I pressed update nothing else happens then that my explorer shows a empty page.

---

On this WU you get almost twice the normal credit: https://boinc.bakerlab.org/rosetta/result.php?resultid=33986546

The only reason those anomalies "pop" out is because of the short WU run time - in one case you probably got a particulary complex model, and the other one aborted early because it didn't lead anywhere.

Another problem, I can't edit my prefs anymore, as some usersadvised I tried to set it from 1 to 4 hours, but after I pressed update nothing else happens then that my explorer shows a empty page.

Same here, I even tried with Firefox and get an empty page as well.

I can't edit my prefs anymore, as some usersadvised I tried to set it from 1 to 4 hours, but after I pressed update nothing else happens then that my explorer shows a empty page.

This is likely a bug. David Anderson made a global announcement of a security hole and told projects to update some files ASAP. I updated the files on both ralph and R@h but have not had a chance to really test the changes.