Message boards : Number crunching : My rant: Why you should care about the credit-system
Author | Message |
---|---|
![]() ![]() Send message Joined: 31 May 06 Posts: 33 Credit: 97,311 RAC: 0 |
Hi all, I thought you might be interested in my rant about the BOINC credit-systems in general and why scientists should care about it. http://itnomad.wordpress.com/2006/09/06/boinc-why-you-should-care-about-the-credit-system/ Feedback is welcome :) Cheers, Alex. "I am tired of all this sort of thing called science here... We have spent millions in that sort of thing for the last few years, and it is time it should be stopped." -- Simon Cameron, U.S. Senator, on the Smithsonian Institute, 1901. |
STE\/E Send message Joined: 17 Sep 05 Posts: 125 Credit: 4,064,265 RAC: 151 |
Good Article, but there is really nothing anybody can do about the shambles the Credit System is in. As long as new projects are allowed to come on-line allowing almost unlimited Credit to the participants then the Credit System is nothing but a farce...IMO You can go to almost any of the newer alpha projects and a lot of people are getting 100-200 or even more Credits per hour by using the Optimized Clients which do nothing but increase their Credits & inflate their ego's ... |
![]() ![]() Send message Joined: 20 Aug 06 Posts: 9 Credit: 1,037,281 RAC: 0 |
Good Article, but there is really nothing anybody can do about the shambles the Credit System is in. As long as new projects are allowed to come on-line allowing almost unlimited Credit to the participants then the Credit System is nothing but a farce...IMO Actually I think the credit system has improved quite a bit lately, at least in the projects I particiapate in. (SETI, Einstein, Rosetta, CPDN). the only caveat is that I use an optimised client on SETI, which merely seems to bring Seti into line with the others, but credit/hr is now pretty flat (within 20-25% anyway) across all 4 projects. Only 6 weeks ago I was getting at least double, maybe triple the credits per hour of all the others on Einstein. So things aren't perfect, but they seem to be improving..... I also think the article pretty-much gets it. If you want people to crunch your project, you can expect them to give up money, time, sleep, food, sex and grades for you, but don't take away their RAC. --miw --miw |
STE\/E Send message Joined: 17 Sep 05 Posts: 125 Credit: 4,064,265 RAC: 151 |
I agree that some of the Projects are trying to get in line with some of the other Projects to get the Credits in line with them. But the point I was making was that almost all the new Projects allow almost as much Credit as you can stand thus forever Skewing the over all BOINC Credit Ranks. So if you process for the Projects (without using an Optimized Client that does nothing for the Project) that have their Credit System normalized < (for lack of a better word) & in line with other normalized Projects while Joe Blow Jumps from Project to Project that allows unlimited Credit by using an Optimized Client or by whatever other means it takes to get the Credit you are simply getting killed in the Credit Race if thats what you are into. |
![]() ![]() Send message Joined: 18 Sep 05 Posts: 653 Credit: 11,786,769 RAC: 386 ![]() |
Surely, it is simply BOINC combined stats that are meaningless. If project a is giving 10 credits per hour and project b 100 credits per hour, then the BOINC combined credit whores will crunch b, but they will never get a high place at project a. A cruncher with a wide portfolio of projects will compete like with like, that is what I seem to do anyway. I have no idea which of the 12 projects I crunch gives the most credit per hour. Wave upon wave of demented avengers march cheerfully out of obscurity into the dream. |
Dave Wilson Send message Joined: 8 Jan 06 Posts: 35 Credit: 379,049 RAC: 0 |
It sounds kind of like, I will pay you more credits per hour to crunch for my project. Now that's competition. Projects competing for users by offering up more credits per hour. What were those projects you say? There are projects that want users and there are projects that think they can get along without them. |
![]() ![]() Send message Joined: 31 May 06 Posts: 33 Credit: 97,311 RAC: 0 |
PoorBoy wrote: Good Article, but there is really nothing anybody can do about the shambles the Credit System is in. As long as new projects are allowed to come on-line allowing almost unlimited Credit to the participants then the Credit System is nothing but a farce...IMO Yeah, that might be true; although one must admit that this argument is more about combined-stats, not to project-stats as adrianxw already pointed out. You can go to almost any of the newer alpha projects and a lot of people are getting 100-200 or even more Credits per hour by using the Optimized Clients which do nothing but increase their Credits & inflate their ego's ... Oh Christ, yes, you can. Well i just hope that the whole credit-issue will be settled to everyone's satisfaction really really soon as i think that all those flamewars start to tear appart the communities. Thanks for your feedback, Alex. P.S., OT though: I really could use a "preview"-button so that i don't have to edit my posting over and over again when i screwed up the burning-code stuff... :-) "I am tired of all this sort of thing called science here... We have spent millions in that sort of thing for the last few years, and it is time it should be stopped." -- Simon Cameron, U.S. Senator, on the Smithsonian Institute, 1901. |
![]() ![]() Send message Joined: 31 May 06 Posts: 33 Credit: 97,311 RAC: 0 |
Actually I think the credit system has improved quite a bit lately, at least in the projects I particiapate in. I totally agree. One can hope that the whole credit-system will normalize over all (important) projects. If you want people to crunch your project, you can expect them to give up money, time, sleep, food, sex and grades for you, but don't take away their RAC. *lol* Hillarious! I hope i'm allowed to quote you in my .signature... :-) Thanks for reading, Alex. "I am tired of all this sort of thing called science here... We have spent millions in that sort of thing for the last few years, and it is time it should be stopped." -- Simon Cameron, U.S. Senator, on the Smithsonian Institute, 1901. |
tralala Send message Joined: 8 Apr 06 Posts: 376 Credit: 581,806 RAC: 0 |
Actually I think the credit system has improved quite a bit lately, at least in the projects I particiapate in. (SETI, Einstein, Rosetta, CPDN). the only caveat is that I use an optimised client on SETI, which merely seems to bring Seti into line with the others, but credit/hr is now pretty flat (within 20-25% anyway) across all 4 projects. Only 6 weeks ago I was getting at least double, maybe triple the credits per hour of all the others on Einstein. Yeah I too think it has improved recently. The major 4 projects you mentioned have now switched to credit/WU which improves teh situation quite a bit. With the recent reduction on Einstein the credit/hr between the different projects has also approximated. However after a brief look on the results at Einstein it seems to me they still grant about 20% more credit/hour than Rosetta for example. |
![]() ![]() Send message Joined: 20 Aug 06 Posts: 9 Credit: 1,037,281 RAC: 0 |
Yup. Bang on. My main cruncher's recent credit/hr (Athlon 64 X2 4800+) Rosetta: 32.7 Seti: 34.4 (Chicken's optimised app) Climate Prediction: 33.9 Einstein: 47.3 But: Climateprediction just upped credit/trickle by 14.3% today. (226.8/trickle to 259.2/trickle) Einstein flys very fast on CPUs with SSE2 and SSE3 but not so much on slower CPUs, so it may not be so much out of whack on other platforms. --miw |
tralala Send message Joined: 8 Apr 06 Posts: 376 Credit: 581,806 RAC: 0 |
So the race has already started... :-( Well I won't speculate but wait until mmciastro makes his final spreadsheet, that will give factual data. |
Jose Send message Joined: 28 Mar 06 Posts: 820 Credit: 48,297 RAC: 0 |
I respectfully ask: Why should developers subjugate their projects needs as they percvieve them to the whims/dictates/biases or whatver of the compilers of statistics? Why this cult to statistics? When BOINC was developed the credit granting area was left as part of the open source. Can I interpret that as to mean that developers of specific projects were given the freedom to implement a credit system as they see fit? Are people arguing that the use of BOINC should be limited to only those projects that satisfy their views on how that project's credits are granted? Am I reading here that new projects under BOINC should be limited/restricted or even not allowed on the issue of credits and credits alone? People that is what it seems it is being argued? But I can be wrong in reading all the posts . So can anyone explain to me what is so vital about all projects having similar credits granting structure? I hope there is another reasons other than the artificial construct of statistics. What is gained in continuing the credit issue fight in every project? |
![]() ![]() Send message Joined: 31 May 06 Posts: 33 Credit: 97,311 RAC: 0 |
Jose wrote: Are people arguing that the use of BOINC should be limited to only those projects that satisfy their views on how that project's credits are granted? Am I reading here that new projects under BOINC should be limited/restricted or even not allowed on the issue of credits and credits alone? I'm only speaking for myself; i guess you're refering to PoorBoy's statement. I don't think that he meant it in a literal way and i don't think that this was an appeal to restrict it. But I can be wrong in reading all the posts. So can anyone explain to me what is so vital about all projects having similar credits granting structure? I hope there is another reasons other than the artificial construct of statistics. It's the psychological effect; people want to know that their contribution is any worth. Handing out credits is an incentive to keep the people working on the project. If your project looks bad compared to others (less credit) they might wander off for various reason. It's about being part of a team, realizing that you're part of something special and if you get "disappointed" you might turn around and never look back. If I collect all my understanding, I must disappoint you: Yes, it's all about the artificial construction of cross-project-statistics. Not so much the internal statistics in a project (that's something you can solve with more power, to quote Tim Allen). Cheers, Alex. Edit: DOH! Added the missing character "r" to PoorBoy's nickname... sorrysorrysorry......... "I am tired of all this sort of thing called science here... We have spent millions in that sort of thing for the last few years, and it is time it should be stopped." -- Simon Cameron, U.S. Senator, on the Smithsonian Institute, 1901. |
Ananas Send message Joined: 1 Jan 06 Posts: 232 Credit: 752,471 RAC: 0 |
I do not really see a problem with higher credits for CPDN, successfully running a CPDN model is much harder than any other project out there. A credit system that ramps up with the trickes would have been smarter though, so only people with a lot of endurance would get the full bonus. |
tralala Send message Joined: 8 Apr 06 Posts: 376 Credit: 581,806 RAC: 0 |
@ Jose BOINC-Cross-Project-Parity is not a must-have but a nice-to-have. First of all it makes it easier to compare different projects, teams, users and hosts regarding computational power. Second, I find the idea of one big homogenuous framework for volunteering computing with many different projects, but which all share the basic mechanism fascinating. Third, if credit is granted equally they competition between different projects is assured to be friendly and no project can be accused of using dirty tricks or cheats to get more users. Fourth, and that is probably the most important reason, within BOINC it is easier to attract new users and gain in size. If you want that advantage to be sustained each project should not put at risk the cohesion of the BOINC system. |
![]() ![]() Send message Joined: 18 Sep 05 Posts: 653 Credit: 11,786,769 RAC: 386 ![]() |
But: Climateprediction just upped credit/trickle by 14.3% today. The change in credit at CPDN is a result of the current "coupled model", since it's inception, being underated relative to the other slab models. People using the same machine, going from Sulphur cycle to coupled model were seeing their RAC drop, so the factors were adjusted to remain internally consistent within the CPDN project group. I already made the point earlier in the thread, comparing a CPDN credit with a Rosetta credit or a uFluids credit is like trying to equate an apple with a haddock. Wave upon wave of demented avengers march cheerfully out of obscurity into the dream. |
Message boards :
Number crunching :
My rant: Why you should care about the credit-system
©2024 University of Washington
https://www.bakerlab.org