| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Cure Muscular Dystrophy - Phase 2 Forum
Thread: Can someone please explain this............. |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 27
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I can half relate to the "frustration" of the points. But I got past that a long time ago.
I have 3 PCs running. P4 1.8g w/1g memory AMD duo 2g /2g memory P3 1g w/256mg memory Wide range of times between them to do HCC wu. 14+ hr 8+ hr 89+ hr I aborted 1st HCC on the P3 because I was almost to 24hr. And to completion was something like 29hrs and adding. That was after restarting the PC because it was adding what I thought was too much time. But I left the second wu run to finish. Anyway, the AMD and P3 can do HCMD wu in 6.00 hrs and the points don't even come close. But the P3 gets the wu's done. To me it's nice to watch points. But they don't really solve anything. That's why I have my PCs running. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
HCMD2's have a soft limit of 6 hours. If 60% of the included positions have not been completed these jobs get cut off and the remaining positions repackaged and send out in a new job... so that you don't get to run 29 hour jobs and longer.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
To Sekerob,
That's the 1st time that has made sense to me on the %. I've read in the past in other forums about "time limits" that for some reason didn't seem to apply to how my AMD was getting WUs done. I guess sometimes you've got to experience something to understand how it applies to something. That's my story. Thanks for the information. |
||
|
|
Recluce
Cruncher Joined: Nov 22, 2009 Post Count: 9 Status: Offline Project Badges: 
|
Many people here seem to forget human psychology:
- The point system is supposed to motivate people, give a metric for relative contribution of individuals and teams. That is its only function. You cannot buy anything with the points or use them for anything - Any perceived or real unfairness of the point distribution will demotivate people. Demotivating people is the opposite of the intent of the system and in does turn people away from an important community effort. Any system that does not achieve what it is designed to do and may even manage to do the exact opposite, has to be considered broken. Thus, the WCG point system should be considered broken. Also consider this: the one guy that posts "Arrividerci, I am out of here" stands for at least a hundred, probably more, that just quietly remove boinc from their systems. BTW: I only crunch on a single machine and divide my time between all the projects, as I do not care much about badges and points - but even I feel anger, when I see another result returned with a point distribution that I feel is unfair. |
||
|
|
Mysteron347
Senior Cruncher Australia Joined: Apr 28, 2007 Post Count: 179 Status: Offline Project Badges: 
|
No doubt what you say is true, Recluce.
What we have here is a demand for the points system to be "made fair." So what is "fair?" Regardless of the system implemented, someone will be able to find some pathological combination of circumstances to expose some variety of "unfairness." This parallels life, where the politicians are frantically engaged in making law upon rule upon regulation to try to appease those who will find some pressure group to belong to so that they can claim "disadvantage" or "discrimination." Rather than simply leaving in a huff, I'd suggest the correct and reasonable course would have been to engage in the debate and attempt to develop a "fair" system. It's easy to criticise the current system - far harder to devise a replacement that would suit everyone. It's well-acknowledged that if WCG was to allot a larger BOINC score then those who chase these valueless points would be attracted to the science here rather than using their resources on Monty Hall BOINC projects of dubious value. The downside would be that the other projects would respond in kind to recover their "market share" resulting in even greater mountains of points being allocated for the same actual work. All over "points" worth precisely NOTHING. |
||
|
|
Recluce
Cruncher Joined: Nov 22, 2009 Post Count: 9 Status: Offline Project Badges:
|
For people to feel that the points are awarded fairly, more information should be available on demand.
For example: some WUs cut off after six hours, no matter how much of the calculations is completed. So people might see - and not understand - why they got, for example, after 6.00 hours 110 points claimed, 65 granted, while another machine on the same WU claimed 6.00 hours, 87 points and got 142 points granted. So adding a link "More info about points granted calculation" to the individual WU result page would not reduce readability and give the people the chance to pursue further - if they wish to. Real problems (64bit penalty, for example) should be eliminated. The list is probably much longer, but I don't pretend to know all the quirks of the system. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Most of the quirks of points/credits are covered in the Start Here forum FAQ's, substantially replicating posts by the technicians, verbatim. The FAQ index is pinned to the top of all forums, the place where questions are raised, very hard to miss, I think.
----------------------------------------Yes we can propose for a text bit to be added to all the Result distribution headers, similar to what is already shown on the HPF2 & RICE headers: The large number of copies sent out for this workunit is due to the unique nature of this project. And yes, 6 hours, 110 points claimed, 65 granted could happen, but if at all, for 64 bit clients quad owners you may have a look at my last 15 HCMD2 results and conclude that there may be the case, on the whole though, for a 32 bit compiled application, a client with a pumped up 64 bit integer benchmark (dhrystone), it not being as bad, the below totaling to 1272 claimed / 1295 granted (to a 64 bit client.We encourage you to read the FAQs about this project for more information. CMD2_ 0344-HXK2A.clustersOccur-1YNS_ A.clustersOccur_ 14_ 6178_ 6218_ 0-- 1112084 Valid 17-2-10 11:57:25 23-2-10 14:29:28 5.99 106.4 / 124.9 Yes at the top the job finished slightly before 6 hours and got more than full marks, the wingmen also not getting a shabby amount, probably 64 bit too, claiming 143... an obscure 6.10.16 alpha release.CMD2_ 0333-ATPB.clustersOccur-1RXS_ C.clustersOccur_ 11_ 32384_ 33142_ 1-- 1112084 Valid 16-2-10 22:34:42 23-2-10 09:33:09 1.86 33.1 / 27.1 CMD2_ 0348-MYH3.clustersOccur-1IKN_ A.clustersOccur_ 418_ 0-- 1112084 Valid 16-2-10 05:05:46 23-2-10 01:40:22 5.60 99.1 / 96.6 CMD2_ 0348-MYH3.clustersOccur-1IKN_ A.clustersOccur_ 279_ 0-- 1112084 Valid 16-2-10 04:40:52 22-2-10 22:58:40 5.01 88.7 / 81.9 CMD2_ 0348-MYH3.clustersOccur-1IKN_ A.clustersOccur_ 381_ 0-- 1112084 Valid 16-2-10 04:40:28 22-2-10 19:13:22 5.22 92.0 / 84.3 CMD2_ 0343-HXK2A.clustersOccur-2RGN_ B.clustersOccur_ 4_ 9350_ 9994_ 1-- 1112084 Valid 14-2-10 20:44:28 22-2-10 13:42:46 4.15 73.8 / 71.4 CMD2_ 0329-1433GA.clustersOccur-2YU1_ A.clustersOccur_ 123_ 110560_ 110831_ 1-- 1112084 Valid 14-2-10 20:02:31 22-2-10 07:45:05 4.74 84.2 / 96.4 CMD2_ 0344-HXK2A.clustersOccur-3D0G_ A.clustersOccur_ 188_ 14892_ 14911_ 1-- 1112084 Valid 14-2-10 19:03:20 22-2-10 07:36:10 5.68 100.9 / 93.3 CMD2_ 0337-CUL3A.clustersOccur-2HD5_ A.clustersOccur_ 11_ 35655_ 36571_ 1-- 1112084 Valid 14-2-10 19:01:59 22-2-10 02:24:18 4.76 84.5 / 74.9 CMD2_ 0342-HXK2A.clustersOccur-1EZX_ A.clustersOccur_ 0_ 330_ 541_ 1-- 1112084 Valid 14-2-10 18:28:48 22-2-10 01:51:21 6.17 109.5 / 103.5 CMD2_ 0343-HXK2A.clustersOccur-2YU3_ A.clustersOccur_ 3_ 24440_ 27047_ 0-- 1112084 Valid 14-2-10 18:26:53 21-2-10 21:04:52 5.54 98.4 / 89.5 CMD2_ 0337-CUL3A.clustersOccur-1GUW_ A.clustersOccur_ 15_ 1-- 1112084 Valid 12-2-10 00:45:51 16-2-10 04:52:17 4.56 85.5 / 77.4
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|