Message boards :
Number crunching :
Response to concerns regarding the new credit system.
Message board moderation
Previous · 1 . . . 11 · 12 · 13 · 14 · 15 · 16 · 17 . . . 18 · Next
Author | Message |
---|---|
kevint Send message Joined: 17 May 99 Posts: 414 Credit: 11,680,240 RAC: 0 |
Come join me at Rosetta or SIMAP We will take them, it will just take longer than expected. I expect that sooner or later Crunch3r or someone else will be able to develop a optimized app that will be comparible to the old one. Seti.Usa/Rosetta is in top 5 for daily credits, RAC will catch up soon, and moving up nicely with SIMAP. However I will check you guys out over there - I love Hawaii, lived there for several years back in 78-84 while I was in school. Thanks for the invite. |
Administrator Send message Joined: 26 Jan 06 Posts: 43 Credit: 13,801 RAC: 0 |
Problem is you did not explain it. Go re-read what I wrote. My analogy makes sense if you read it properly. If you want to redefine it to something else that doesn't make sense except to pretend that you're getting ripped off, of course you'll get that result. Spilled water, the way I defined it, is any kind of optimization that hasn't happened. The nuts and bolts of it, are that before, you recieved extra credit for results above the norm. Now that everyone has a much more optimized app, you are not getting results above the norm, so there's no reason to award you any extra credit. Then when someone comes along and decides to make us a society where all should get paid equal not matter what they have invested, the ones like you think that is cool(always the majority) while the ones that have invested in the project cry foul. Let me get this straight... Your contention, is that you think you deserve to have more credits per work unit than the norm, because you have invested more? And that the larger of an investment you make in SETI, the more you should be awarded proportionally, per work unit? Ex) 10 individuals investing 1 identicle computer each, may get, 1,000 credits total, per day. .. but you, as 1 individual investing 10 (of the same as above) computers, should get more than 1,000 credits per day, because you've invested more? So, not only should you recieve more linearly, for your added capacity, but your "pay rate" should be much higher, because you've invested more? Interesting. I disagree with that being a good solution to implement though. I think linearly more is fair, and nothing else. Your solution seems very self-centered to me. I think we should recieve credits in proportion to the amount of results you get. If you can get results faster, you deserve more. If you can't, regardless of "how much you've invested", you haven't earned it. And from what I can see - you joined the project a few months ago, some of us has been with this since the start, I just passed 7 years of crunch time for this project. You've got me there. I can't compare to your attachment or involvement in this project. I don't think it would change my opinion, but, I wouldn't know. And the big problem we have been seeing is WU's looping, and getting stuck Well, that right there is a completely valid concern that should be addressed. Computers hanging or WUs looping is a bug that needs to be fixed. That sucks in the short term, but is acknoledged and being worked on, no? And, doesn't have anything to do with the way credit is being assigned. Damn straight, if my computers were hanging, or, spending all week on the same work unit without getting any credit, that I'd jump ship and go elsewhere. That makes perfect sense. 30 hours of crunch time for 30 credits when it used to take 20 minutes for 30 credits is not what I call better buckets. It is called buckets with holes in them. Agreed. Big gaping holes. But, take that out of the equation. If the bugs get worked out, and WUs stop looping and such, are you fine? ... I'm not some SETI leghumping cheerleader. I've actually mostly abandoned SETI too since Enhanced came out. One of the reasons, is simple economics (for credits), I used make a whole lot of credits on SETI.. now I make less. I'm not bitter. I'm not demanding anyone make changes. I don't think I'm being treated unfairly. I'm happy for SETI being much more productive. Here's my choice: I have in one hand, a fair and well-optimized default SETI. In the other hand, I have a sloppy and poorly-optimized default Einstien (which I've optimized). Einstien feeds me double the credits. Simple credit economics says I should jump ship. I also run quite old machines, which took 24 hours to crunch 1 work unit, and which I don't trust being able to meet deadlines on Enhanced. I also have more of a science interest in Einstien right now. That's my choice. Everyone makes their own choice. I just don't see the rationelle for getting mad at the SETI people for you (the general 'you', not you specifically) not earning credit as fast now that SETI can't be optimized as heavily from the default anymore. |
Administrator Send message Joined: 26 Jan 06 Posts: 43 Credit: 13,801 RAC: 0 |
Seems we all turned back in the same claimed credit of 29.73, even though the actaul time crunched on the machines varied from 2.8 hours to 19.1 hours. Hrm.. bear with me, I might not understand how this works.. But, doesn't that make sense? I.E. If I had a P3-500 and it took 19.1 hours to crunch the unit, and you had a P4-2.0 and did it in 2.8 hours.. that regardless of how long it took each computer, both knew how much computing power it took to crunch, and, it's bang-on that they all think it was worth 29.73? Isn't it a good thing that credit claims are so close to eachother? Now everybody can try to explain it away as being fairer to everybody. But I don't see it. If we was still on the older 4.18, that machine would have turned in almost 4 units, that means 4 times the science, and 4 times the credits. Well it depends if this is one of those wonky WUs that screwed up. Unit sizes have changed. If you were running an optimized science app.. yes. You *will* be lower now. Everyone who wasn't running an optimized science app was really getting paid for more than they were worth. That was the only justification for (until they fixed that), paying you 4 times as much. Might as well, because you *were* getting results 4 times as fast. But not anymore. Sucks, but, that's what it is. In terms of science, you should be equal. I think all that's happened is that the bar has been raised. People who were running optimized before are doing equal science now as they were before. But for the other 99%, they're doing ~3 times the science. I suppose it would be nice if BOINC was just SETI and they didn't have to keep everything constant. For credit purposes, all apps are assumed to be fully optimized. When you let a P4-3.0 run for an hour.. credit is given assuming full productivity. So, that's the standard they have to fall back on. Don't look at it as being unfairly (poorly) treated now, look at it as being treated unfairly (well) before, if you weren't running an optimized client. As a bonus, in the interim time period, if you found a way to produce more, you got paid extra for it until they fixed everyone else. So from my stand point, the economics of it is this. The machines are costing to much to run now. They will be converted to a project that has direct impact on the quality of life here on this planet. Well, they're "costing" you the same they were before, right? In real-world costs? You're just not seeing as much credit in return, correct? But, I'm in the same boat as you. Given two equivalent projects, I'll run the one with the most credit given. For as long as that lasts. |
kevint Send message Joined: 17 May 99 Posts: 414 Credit: 11,680,240 RAC: 0 |
No admin, I don't expect to get paid more per WU than others, but I do expect to get paid the same for the same amount of CPU cycles. I.e. On several of my machines I was seeing 2400 credits a day. If I allow these machines to work the same amount of time each day, the same amount of CPU cycles as before then I should be granted the same amount of credits each day for that work. There is no reason why I should want to do the same amount of work for less. Enhanced or not. To prove my point, I take the same machine and move it to say Rosetta, I still get just over 2000 credits a day, while if I keep it crunching Seti, it is getting 1100. I tested it on SIMAP, SIMAP returns about 1800 day, Einstien- 1900. Seti Enhanced did not level the playing field between projects. |
Administrator Send message Joined: 26 Jan 06 Posts: 43 Credit: 13,801 RAC: 0 |
I.e. On several of my machines I was seeing 2400 credits a day. Unless you're talking about differences from when you were optimized, yes, I would agree. That's the whole entire point of using a common credit system. To prove my point, I take the same machine and move it to say Rosetta, I still get just over 2000 credits a day, while if I keep it crunching Seti, it is getting 1100. I tested it on SIMAP, SIMAP returns about 1800 day, Einstien- 1900. Seti Enhanced did not level the playing field between projects. Interesting. My mistake, that does sound quite valid. It seemed to me all the complaints here were of the "Look at how much I was making before on S@H Optimized, and now I'm not! I made assumptions based on those, and it's up to you make sure my assumptions are still valid!" nature. Ruling out the bugs with looping WUs and stalls, default applications on everything, and making sure it's not due to some kind of validating lag that'll just take a week or two to straighten itself out... No explanation has been given for this yet? |
W-K 666 Send message Joined: 18 May 99 Posts: 19078 Credit: 40,757,560 RAC: 67 |
No admin, I don't expect to get paid more per WU than others, but I do expect to get paid the same for the same amount of CPU cycles. I agree at the moment that large amounts of credits can be obtained on Einstein at the moment and for about the next three or four weeks. But we have one computer doing work for Rosetta, using BOINC 5.4.9 and that is getting approximately the same credits/time as Seti and CPND. Might I think you are using a Trux or Crunch3r client with the wrong settings. Andy |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
May I ask everyone to watch their language? You can have a heated discussion all you want, as long as you keep the expletives and offensive language out of it. |
Douglas Send message Joined: 18 May 06 Posts: 1 Credit: 3,328 RAC: 0 |
May I ask everyone to watch their language? You can have a heated discussion all you want, as long as you keep the expletives and offensive language out of it. Don't forget the "colorfull metaphors" :D What would life be like without them? |
kevint Send message Joined: 17 May 99 Posts: 414 Credit: 11,680,240 RAC: 0 |
I agree at the moment that large amounts of credits can be obtained on Einstein at the moment and for about the next three or four weeks. But we have one computer doing work for Rosetta, using BOINC 5.4.9 and that is getting approximately the same credits/time as Seti and CPND. Might I think you are using a Trux or Crunch3r client with the wrong settings. No, I do use trux on most but since ehnanced came out, I have it set to calibrate none. And on the machines I don't use trux I still get higher credit ratings by over 2-1 than with the current enhanced seti. And I don't consider what I am getting on Einstien a large amount - 1700-1800 day barley equates to large amounts, I should optimize this einstien machine and see what it brings. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
What a load of bull take the donkey and go home. Berkely is not paying on a one for one basis. The analogy is about how credits are granted. If you are having trouble with errors, that is an entirely different issue. You can help get errors fixed by reporting as much as possible about them. Or, you can whine about lost work due to errors. Whining about errors will get the attention of the developers, but it probably won't lead to fixes unless you actually describe them. ... and, this thread is about the credit system, not bugs in the SETI app. |
Xaak Send message Joined: 22 May 99 Posts: 32 Credit: 22,636,357 RAC: 0 |
I see about the same with respect to Rosetta. I have both, split 50/50 on a Pentium 630 and am getting roughly the same credits per hour crunched on both projects. Seti varies more by AR, some wus get more per hr than ohers, but on average it's within 2 credits/hr (factoring in 2 instances). I don't run CPDN so I can't comment on that project. XaaK |
ML1 Send message Joined: 25 Nov 01 Posts: 20331 Credit: 7,508,002 RAC: 20 |
I agree at the moment that large amounts of credits can be obtained on Einstein at the moment and for about the next three or four weeks... And this aspect of which project gains the best optimised clients before those optimisations get released to everyone adds a whole new dimension to the competition game. Brilliant! Rationalising the credits for everyone for each new release of the standard client (for whichever project) is a fantastic way to spur new developments to gain whatever new optimised advantage. Rather than buckets and water, a better analogy is that of F1 motorcar racing where optimisations are taken to the extreme and the game is won by whichever car has the best new optimisations for that season. Happy crunchin', Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65764 Credit: 55,293,173 RAC: 49 |
My last response to this thread as I see It's not stickyed anymore. Even with some of the Uncle B's crunchin It's still goin down, I wonder when It'll change? The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
W-K 666 Send message Joined: 18 May 99 Posts: 19078 Credit: 40,757,560 RAC: 67 |
I agree at the moment that large amounts of credits can be obtained on Einstein at the moment and for about the next three or four weeks... It is actually a cunning plot to get everyone to crunch for all the projects, and so that the whingers can meet some of the moderators on the other sites. Andy |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
This graph is for one team, not S@H as a whole. My last response to this thread as I see It's not stickyed anymore. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
This graph is for one team, not S@H as a whole. Yep. So let's have a look at mine, as i've only ever crunched Seti & haven't chopped & changed & started & stopped crunching depending on what sort of mood i'm in at the time. And i have only ever used the official software. Appart from a few peaks & troughs due to project outages & recoveries & one of my computers having a hiccup or 2 for a few hours, the number of credits per day this month doesn't appear to be that different from last month. EDIT- fixed up bbcode, then spelling. Grant Darwin NT |
rainmaker Send message Joined: 20 Jan 04 Posts: 22 Credit: 10,898,070 RAC: 0 |
Reading a lot of administrators excuses a lot of colorful compoarison to bathhouses and so on - but it keeps to be very "buggy" credit-system. Dear Administrator - explain this (results of the same machine): 330214935 79215371 22 May 2006 8:18:42 UTC 29 May 2006 4:58:17 UTC Over Success Done 111,355.00 64.89 64.53 330214906 79215362 22 May 2006 8:18:42 UTC 30 May 2006 21:10:28 UTC Over Success Done 41,082.00 64.89 64.53 nearly 3time longer WUs do the same credit - the exactly same ?!?!) another machine: 333726069 80039254 29 May 2006 8:46:39 UTC 31 May 2006 4:36:03 UTC Over Success Done 8,273.55 12.52 12.52 333726064 80039253 29 May 2006 8:46:39 UTC 31 May 2006 4:36:03 UTC Over Success Done 12,238.97 12.52 12.52 333726061 80039257 29 May 2006 8:46:39 UTC 31 May 2006 4:36:03 UTC Over Success Done 11,385.02 12.52 12.52 333726046 80039255 29 May 2006 8:46:39 UTC 31 May 2006 4:36:03 UTC Over Success Done 7,931.09 12.52 12.52 semms, it doesn matter how big and long the WU needs - the credit is easy to calculate (its not CALCULATEd - its fixed!!!) sometimes you really "calculate" (my 3rd machine): 333420911 79966108 28 May 2006 18:57:37 UTC 29 May 2006 8:49:00 UTC Over Success Done 19,827.11 49.30 49.30 330942640 79386188 23 May 2006 23:04:25 UTC 27 May 2006 9:21:26 UTC Over Success Done 37,394.81 45.75 45.77 double the work - reduce credit ??? Sorry Folks - You haven't done a good job, it seems you are using us all as beta-tester or we are member of a psychologic study "how much boring of the guys who du the crunching work ist needed, before the turn of there machine ..:" looking forward to your "excuses" and "explanations" or better: your BUG_FIXES ! best regards Rainy |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13746 Credit: 208,696,464 RAC: 304 |
Dear Administrator - explain this (results of the same machine): With the computers hidden it makes it almost impossbile to check what's going on with your processing. looking forward to your "excuses" and "explanations" or better: your BUG_FIXES ! Don't know why you're posting all that here- from what i was able to see you're using an optimised client for which the BOINC & Seti developers aren't responsible. I suggest you email the developer of the optimised client. Grant Darwin NT |
Pooh Bear 27 Send message Joined: 14 Jul 03 Posts: 3224 Credit: 4,603,826 RAC: 0 |
looking forward to your "excuses" and "explanations" or better: your BUG_FIXES ! The answer has been given. It's no longer about time. It's about FLOPS. Some FLOPS take longer than others. It's a simple concept, that people do not seem to get. My movie https://vimeo.com/manage/videos/502242 |
Clyde C. Phillips, III Send message Joined: 2 Aug 00 Posts: 1851 Credit: 5,955,047 RAC: 0 |
It appears that recent average credit has been dropping due to the advent of Enhanced. I believe it's because of the different credit counting system. Also there may be more errors, perhaps ten percent, that pull down RAC. There is plenty of disparity among the Enhanced units - short ones yield more cobblestones per hour. Yes, it does look like, a lot of the time, each machine claims nearly the same amount, for that same unit. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.