Why are World Community Grid points not consistent?

[Deleted member 88227]

Can WCG WU's fail for whatever reason? I ask because my production seems to be all over the place.

For example the past three days:
03/15/2017

0:185:21:12:27

938,632

3,313
03/14/2017

0:209:20:23:22

1,039,674

3,524
03/13/2017

0:148:06:51:43

862,664

4,404

Look how far off the results returned are. Even today my numbers seem really, really low.

03/16/2017

0:176:09:29:05

720,369

2,945

Since I have almost all my boxen set to only use the network between midnight and 5am my next 7PM update today will not be anywhere near enough to make up 100k+ points. I'll probably get around 20k points at most.


03/12/2017

0:083:13:57:33

638,559

2,087
03/11/2017

0:071:13:18:07

334,554

1,408
03/10/2017

0:083:08:28:52

358,077

1,319
03/09/2017

0:071:04:34:19

303,151

1,028
03/08/2017

0:078:21:00:43

314,160

1,074
03/07/2017

0:076:21:37:34

394,675

1,012
03/06/2017

0:075:08:03:49

397,457

1,002

On 03/11 I returned from my vacation and turned all my down'd boxen back on so hence the low numbers up to the 11th. Since it was the middle of the night on the 11th I can understand why the 12th was rather low, however they've been running 100% without any issue since then yet my numbers are all over the place.

Is this normal for WCG?

 

[Gilthanis]

pending validation? pending verification? running different sub project? running more of one sub project than you had been previously? Results returned does not mean verified, so the numbers can easily shift. What you need to look at is number validated. Let us know how those numbers look. Also, points for each sub projects work can shift a bit since they use a modified version of BOINC's credit new system.

 

[Deleted member 88227]

How do I find validated numbers?

 

[Gilthanis]

Not easy... you have to look in your account's results page and filter them.

 

[Gilthanis]

Here is an example

 

[Deleted member 88227]

Well crap. I've got 12 pages of pending verification.

So you're telling me if I were to stop crunching on all my boxen right now that I would still be generating points as those results are validated?

 

[Gilthanis]

That is what I'm saying. Some (well most) sub projects require a wingman to validate the work. However on the server from time to time will require systems to be double checked just to verify they are still a "reliable host" and thus is where the verification work comes in.

 

[Deleted member 88227]

Well that answers my question. Thanks man.

Just wanted to be sure something wasn't wrong on my end. On my quest to world domination. Muahaha

 

[Gilthanis]

Another thing to keep in mind is that most work units have a 10 day deadline. If someone does not finish in that time, another is sent out. These typically have a 4 day deadline. However, that person may not return them in time either. So, some work units could take a while to finish because of this.

 

[Gilthanis]

Something else to watch out for. If you look at the charts that are in the very first post of our WCG thread here: https://hardforum.com/threads/world-community-grid.1833345/ There is one graphic that has a break down of the projects. One of the helpful stats on that chart are what work units (all of them together not individually) from each sub project are averaging. Last category on the chart. Talk to pututu on how to maximize points running FAAH phase 1. there are certain work units that were crediting much better than others.

 

[mmonnin]

PV jail as I've heard it called before on their forums. For some reason their servers can starting sending all your tasks to require validation even when in the past it has been quorum 1. I had one error due to a lockup, reboot or something and my production dropped for a couple days. All my tasks now required validation from someone else. After a couple of days the server decides your PC is ok and goes back to quorum 1.

I know Zika gives better RAC than MCM or SCC. Those apps I've run exclusively for awhile to see the RAC difference.

 

[ChelseaOilman]

Since I have almost all my boxen set to only use the network between midnight and 5am

Satellite internet? That's what I have with a 15 GB data cap. Thankfully between midnight and 5am doesn't count against the cap or I wouldn't be able to run DC projects.

 

[Deleted member 88227]

Nah, 1/1Gbps fiber connection. I just like all my points updating at once in the AM.

 

[Deleted member 88227]

Up to 17 pages now.

Also noticed I've got a partial page of invalid with 10 entries all from the same host. Should I be concerned about this? They're all from the same sub project also.

 

[mmonnin]

Thats probably what is causing the PV jail. Anything happen to that system? Was it all at once or randomly over time.

 

[Deleted member 88227]

Seems that it was all at once. If I am reading them correctly it appears they took too long to process. 20+ hours per result which is really, really high for that system.

Result Name Device Name Status Sent Time Time Due /
Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit
FAH2_ 000947_ zinc01459336_ 1_ 000009_ 0012_ 025_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 20.29 / 20.30 379.9 / 0.0
FAH2_ 000947_ zinc01459336_ 1_ 000008_ 0008_ 022_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 20.17 / 20.19 377.8 / 338.5
FAH2_ 000947_ zinc01459336_ 1_ 000003_ 0002_ 025_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 21.72 / 21.73 406.6 / 0.0
FAH2_ 000947_ zinc01459336_ 1_ 000001_ 0006_ 029_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 20.11 / 20.13 376.7 / 0.0
FAH2_ 000948_ zinc01661685_ 000006_ 0000_ 030_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 22.55 / 22.56 422.2 / 0.0
FAH2_ 000944_ zinc01413906_ 000005_ 0019_ 027_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 20.16 / 20.17 377.5 / 357.7
FAH2_ 000947_ zinc01459336_ 1_ 000008_ 0019_ 023_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 19.91 / 19.92 372.7 / 305.0
FAH2_ 000947_ zinc01459336_ 1_ 000009_ 0005_ 030_ 1-- localhost.local
domain Invalid 3/16/17 04:27:19 3/17/17 04:28:41 21.59 / 21.60 404.2 / 392.3
FAH2_ 000918_ zinc01187822_ 1_ 000006_ 0009_ 027_ 2-- localhost.local
domain Invalid 3/16/17 04:25:09 3/17/17 04:28:41 20.32 / 20.33 380.5 / 360.4
FAH2_ 000914_ zinc01178200_ 000001_ 0005_ 030_ 2-- localhost.local
domain Invalid 3/16/17 04:25:09 3/17/17 04:28:41 20.38 / 20.40 381.7 / 0.0

Guess that's just a garbled mess, but you can see that they all took almost a .01 more time than was allowed to complete the unit. To my knowledge nothing is wrong with that system. It's still crunching away without any issues.

My production is continuing to go down the drain. Now I am up to 25 pages in the PV jail. le sigh