| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: Cooking in Mobile credit lessons... |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 9
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
(my favorite Robert Cray Band live album), but you wonder what brings up this 'coincidence'... the number looked familiar so filtered in WCGDAWS and sure enough, 4 popped up, all granted 168.8, all processed on same device:
E217287_ 090_ J.43.C36H20N2O2S2Se.00068297.4.set1d06_ 0-- Valid 2-12-2013 12:50:00 2-12-2013 11:28:02 12,00 218,9 168,8 27,15 14,07 CEP2 1-12-2013 8:19:06 12,37 E217226_ 337_ I.50.C39H19N7O4.00312608.4.set1d06_ 0-- Valid 28-11-2013 22:56:00 28-11-2013 20:54:13 12,00 239,2 168,8 25,13 14,07 CEP2 27-11-2013 19:46:22 12,07 E217204_ 160_ I.50.C37F3H18N5O5.00361144.3.set1d06_ 0-- Valid 27-11-2013 18:52:00 27-11-2013 12:18:23 12,00 245,0 168,8 24,39 14,07 CEP2 26-11-2013 11:55:03 12,06 E217135_ 693_ I.48.C39H20N4O5.00006557.3.set1d06_ 0-- Valid 24-11-2013 10:54:00 22-11-2013 21:34:20 12,00 244,6 168,8 12,28 14,07 CEP2 22-11-2013 9:17:50 12,08 You pens over what secret recipé s being used [and the one for Coca Cola no longer is] |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Thank you for reporting this. Your workunits that you list here hit a poorly implemented cap on credit awarded for a task. What happens is that there is a limit of 900 BOINC credits for a 100% completed job for CEP2. Your tasks hit the 12 hour timeout and exited earlier so they did not complete the entire task. As an example of what happened, one of your tasks finished 3 out of the 16 jobs int he workunit. After computing the credit for the workunit it estimated that the credit to be awarded if all tasks had completed would be 1167.2 BOINC credits. However, the cap then limited this amount to 900 BOINC credits. The credit awarded for your task was then scaled to (3/16)*900 = 168.75.
Since this limit was put in place, we have other methods to in place to locate excessive credit claims so we are removing this limit. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thank you for that. Since a very long time we actually only do 12 jobs, and the first 3 usually take half or more of the computing time at that if managing to get all 12 done. Now please get rid of "credit new"... it's a dog, one of those naked ones. The picture is not improving... it's name is yo yo and well out of fashion, but who knows some semblance of consistency returns [hope=low]:
 |
||
|
|
Werinbert
Advanced Cruncher United States Joined: Jun 13, 2013 Post Count: 56 Status: Offline Project Badges:
|
Thank you for reporting this. Your workunits that you list here hit a poorly implemented cap on credit awarded for a task. What happens is that there is a limit of 900 BOINC credits for a 100% completed job for CEP2. Your tasks hit the 12 hour timeout and exited earlier so they did not complete the entire task. As an example of what happened, one of your tasks finished 3 out of the 16 jobs int he workunit. After computing the credit for the workunit it estimated that the credit to be awarded if all tasks had completed would be 1167.2 BOINC credits. However, the cap then limited this amount to 900 BOINC credits. The credit awarded for your task was then scaled to (3/16)*900 = 168.75. Since this limit was put in place, we have other methods to in place to locate excessive credit claims so we are removing this limit. I this why my much faster computer has only a marginal increase in points for the same amount of time? For most other projects my fast computer gets a 70% increase in credit per time (comparing app to app). However, for CEP2 work it is only a 25% increase.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Speculation: What OS? CEP2 is 32 bit! Comparing with VINA in 64 bit, on Linux gives a 65-70% speed improvement over Windows, which is the main reason you see that big difference for FAHV result credit.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Contribution to CEP2 dropped by 20% overnight, just 10 years at noon [though overall we did a near noon record of 260 years], and the dog, well, I've just helped it out of it's misery, for I'm uncoupling from CEP2. Your credits now are telling me the device absolutely sucks at computing these [though they run at 99% efficiency]... not worth my micromanagement time. Tired of the credit new random number generator... one doing 15/hr, the next doing 19/hr, the follow 10/hrs.
E217339_ 975_ J.46.C38H22N2O4S2.00057069.1.set1d06_ 1-- Valid 12-05-13 15:49 05-12-13 11:50:33 12,00 183,1 118,3 23,33 9,86 Before claiming around 240 for 12 hours, now 183 and granted 118. Oh wait, the 'claim' is dropping more... things must be really bad with that comp: E217341_ 253_ J.45.C36H19N3S4SeSi.00060131.1.set1d06_ 0-- Valid 12-05-13 16:29 05-12-13 15:24:18 12,00 169,4 169,4 26,18 14,12 Going to do some conservation of energy:  |
||
|
|
Werinbert
Advanced Cruncher United States Joined: Jun 13, 2013 Post Count: 56 Status: Offline Project Badges:
|
Speculation: What OS? CEP2 is 32 bit! Comparing with VINA in 64 bit, on Linux gives a 65-70% speed improvement over Windows, which is the main reason you see that big difference for FAHV result credit. Sorry, I am used to having a link to view the computers of others as in the generic Boinc projects interface. (I wish they would implement it at WCG.) My computers are i7-3770k (Windows 7), i5-2410M (Windows 7), and Core 2 Quad Q6600 (Vista). I have an even older T-2050 (WinXP), but haven't used it for WCG in months. I haven't verified the 32 v 64-bit issue, but I believe on some other 32-bit projects I am still getting the the higher credit ratio. I am worried with CEP2 that the fast computer is crunching for 6 or so hours and then spinning its gears on the failed task for the remaining 6. Only getting credit for the first 6 of work, i.e. before the failed task. Whereas my slower computers are crunching for 9 hours before reaching the failed task, and thus less gear spinning. I am unsure how the credit cap fits in, so I asked. In this case if my computer is loosing 25% (or whatever) of its credits due to the cap then that too can cause a reduction of the cr/effort ratio. * * * * * In the mean time I am crunching CEP2 on my slow computers and MCM on the fast computer. |
||
|
|
Werinbert
Advanced Cruncher United States Joined: Jun 13, 2013 Post Count: 56 Status: Offline Project Badges:
|
Another funky set of results
----------------------------------------E218757_ 658_ I.51.C43H20N4O3S.00239187.2.set1d06_ 1-- XPS Valid 1/10/14 21:15:18 1/15/14 18:43:06 18.00 / 18.82 65.6 / 65.6 E218757_ 716_ I.51.C39H20N6O5S.00195719.3.set1d06_ 1-- XPS Pending Validation 1/10/14 21:15:18 1/13/14 14:46:38 18.00 / 18.85 65.6 / 0.0 E218757_ 700_ I.51.C43H20N6OS.00282724.4.set1d06_ 1-- XPS Pending Validation 1/10/14 21:15:18 1/13/14 14:46:38 18.00 / 18.84 65.6 / 0.0 E218757_ 782_ I.51.C43H20N6OS.00071892.3.set1d06_ 1-- XPS Pending Validation 1/10/14 21:15:18 1/13/14 14:46:38 18.00 / 18.67 65.6 / 0.0 All took the max time allowed, but all asked for the exact same credit amount: 65.6. Some errored on Job #8 some on Job #12, so they should have some variation in requested credits. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Already covered here
|
||
|
|
|