David has put new credit calculation code into the beta. It might not work properly. This is a beta project after all.

---

I was going to question this one. ID 7644981 92.86 claimed 221.77 Granted.

David working overtime??

Thanks for the update.

Michael

Possibly as a side-effect, I'm getting failures on all my uploads this morning.

How about WU 2709319?

Genuine -9 overflow, 0.80 claimed, 186.40 granted.

These were also interesting several 500 +

WU 2575577 193.78 claimed 1046.81 granted
WU 2421826 127.61 claimed 1111.97 granted
WU 2559975 193.78 claimed 1119.33 granted


Michael

Yeah, there have to be quite a few bugs in there. One of my results was 116 claimed, 42 granted. The next one was 116 claimed 690 granted.

---

It's getting better. WU 2711994: 0.11 claimed, 365.05 granted in 27.13 seconds.

Rally round, crew. We'll soon get Milkyway licked at this rate.

Edit: Any advance on WU 2450384? 1,465.64 granted.

Edit 2: How far back does this thing go? WU 2542127: 902.65 for a task reported on 17 Jan 2010.

Hi Richard,

The wing person on this WU reported 11 December 2009.

WU 2421826 127.61 claimed 1111.97 granted.

Edit Are all old WU files now being purged?

Michael

Yes, but I'm planning to turn of deletion in the next few minutes and leave it off until next week.

---

Edit 2: How far back does this thing go? WU 2542127: 902.65 for a task reported on 17 Jan 2010.

The important time is when the latest of the valid results was returned.

Eric

---

Edit 2: How far back does this thing go? WU 2542127: 902.65 for a task reported on 17 Jan 2010.

The important time is when the latest of the valid results was returned.

Eric

That's one reason why I posted - the final (of four) task reported 24 Mar 2010 20:18:41 UTC - a week before the new code was loaded.

I think the validators were off for that week.

---

Not all week - I got some credit dumped on the 30th.

There was a script running every few hours to grant credit for work regardless of whether it had been validated. Unfortunately we had forgotten to turn that one off.

Eric

---

I opened up a couple of extra machines as a test on the 26th. March - we were checking out the Fermi-class GPU needed error message over in Astropulse. I thought I saw some genuine validations from that batch, but it may just have been the generosity of your script :-)

What is the estimated time frame for the new validators..

I have A LOT of WU's waiting to report - hundreds of em.. filling up my cache, with still a cache of uncrunched work.

---

Obviously something didn't go as Planned ... http://stats.free-dc.org/stats.php?page=proj&proj=sahb

Yes, I was just looking at [trac]changeset:21064[/trac]:

davea
- backend: fix scaling problem that was producing xe15 size credits.
This had messed up the beta DB, which I had to clean up.
Added a cap (1e5) to prevent this in the future.

Obviously something didn't go as Planned ... http://stats.free-dc.org/stats.php?page=proj&proj=sahb

It only affected the export to DC-Stats, the project top lists looks ok

---

Obviously something didn't go as Planned ... http://stats.free-dc.org/stats.php?page=proj&proj=sahb

It only affected the export to DC-Stats, the project top lists looks ok

There's no such thing as an export to anywhere. This project (like every other project) makes the public version of its stats available freely for everybody - dumps them in a heap at the side of the road, for anyone to collect. The stats sites grab a copy, of course, but you can download exactly the same figures yourself.

I suspect that the project's own top-lists reflect the clean-up after David realised the problem. But unless he also did a special export of the cleaned-up stats, I suspect Beta'a embarassment will be on display for all the world to see, at every stats site: for 24 hours, starting whenever each particular site chooses to do their dumpster-dive and report the results.

PS FREE-DC is showing an update at Fri Apr 02, 23:12:43 2010 BST - later than BOINCstats 2010-04-02 21:00:51 GMT. Maybe David covered his embarassment after all.