Message boards :
Number crunching :
Observation of CreditNew Impact
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 . . . 15 · Next
Author | Message |
---|---|
KW2E Send message Joined: 18 May 99 Posts: 346 Credit: 104,396,190 RAC: 34 |
Keith, you're exactly right. Everyone will normalize to the same rankings they had prior to the migration. Not sure how many times I have seen people complain about credits over the years. And Mark, $620? Holy crap dood! Mine never went over $300 before I had to cut back for a while. It's amazing how much getting a divorce eats into your budget. :) |
Sleepy Send message Joined: 21 May 99 Posts: 219 Credit: 98,947,784 RAC: 28,360 |
Keith, you're exactly right. Everyone will normalize to the same rankings they had prior to the migration. Yes, RAC-wise. But absolute distances in credits will effectively double. We have had years were fast crunchers were going twice as fast as anybody can do now and have "stockpiled" credits. Now some distances in credits, effectively doubling, can be so big that some gaps cannot be conceivably closed, even if you make huge upgrades to your rig. There definitely IS a difference, if things do not get back towards as they were. You can even not give a damn about credits and ranks, but if you do, it is not nice to discover that during the night the distance between you and your upper neighbor has (in effect) doubled (considering how much time (the double, actually) it will now take you to overtake him/her). And I am talking about inside SETI. The discussion can get more complicated and probably less interesting for SETI only crunchers like I am myself if you consider what happens in relation to other projects. But I am here pointing out that there are consequences also inside SETI itself. Sleepy |
Keith White Send message Joined: 29 May 99 Posts: 392 Credit: 13,035,233 RAC: 22 |
Yes, RAC-wise. True, hitting 2 million credits is going to take longer. Catching up to some target rank who is X credits ahead of me is going to take longer. So hitting 15,000 in overall rank is going to take me a bit longer now. Nothing I can do about it at my end other than buying more equipment and that's not going to happen anytime soon. "Life is just nature's way of keeping meat fresh." - The Doctor |
JohnDK Send message Joined: 28 May 00 Posts: 1222 Credit: 451,243,443 RAC: 1,127 |
I will be ending all crunching when I hit the 20 mil mark. Electric bill is over 300 a month now with rigs blowing in the wind..... What's wrong with both? Don't see the problem with people wanting "normal" credits. I had a RAC of 35-37k before, I sort of like that better than under 30k, wonder how low I will get... ;) btw I'll stay on SETI no matter the credits :) |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
These are some factors which I think may be influencing the credits. 1. Prior to SETI@home v7 rollout, the x41zc CUDA applications were only being used by those willing and able to do the manual install. Now all CUDA hosts are using it. 2. The adjustment to the estimates for the added Autocorr processing is about a 48% increase at VHAR, 14% at midrange or VLAR. The actual increase in runtime is less than that for CPU processing (based on doing v6 and v7 tasks with the stock 7.00 build), but is more than that for GPU processing with x41zc. The differential may mean that RAC for CPU processing will end up somewhat higher than before but for GPU processing somewhat less. Joe |
tbret Send message Joined: 28 May 99 Posts: 3380 Credit: 296,162,071 RAC: 40 |
When you said that, you really named the poison. There truly is little to keep people interested in the SETI@Home project. In fact, for the past two or three years there's effectively nothing at all. We could all go "compete" with each other on other projects so please, a request of the self-deluded amongst us, don't hit me with the self-righteous and sanctimonious bs about how credits and competition don't matter and only the purity of the science does. There's no reason for any of us to be here except to promote the goals of this research. The fact that we're here means we really don't need that lecture. We're here doing the preliminary sifting and "heavy lifting" but if we, as volunteers, actually found an alien signal, or a pulsar, or something entirely novel, we wouldn't know it and can't declare it. The project scientists do the science. We're the manual labor. We accept that. The competition and camaraderie is what keeps us active and makes the long wait bearable. We have to wait and wait and wait for someone to look at all the data we've returned. Our rivalries and credit-goals are, for us, like singing-in-the-fields while we harvest the crops. But... and this is something we would be unwise to forget: This project really does not rely on people who feel that way. There are tens of thousands of people who will "set and forget" the SETI@Home screensaver who never visit a forum, never compare their work with anyone else, who don't notice their video cards are throwing trash, who really don't care if the science is ever done on the huge collection of data, and really don't think about this project one way or the other. If those of us who are active here leave, nobody will even notice. That's the way this thing was set-up to run in the first place. |
Ron Send message Joined: 24 Aug 99 Posts: 42 Credit: 34,544,679 RAC: 0 |
BOINC is software that can use the unused CPU and GPU cycles on a computer. It seems that credits given for said cycles should have the same value whether used for V6 or V7. |
bill Send message Joined: 16 Jun 99 Posts: 861 Credit: 29,352,955 RAC: 0 |
BOINC is software that can use the unused CPU and GPU cycles on a computer. It seems that credits given for said cycles should have the same value whether used for V6 or V7. Be careful Dr. A. doesn't read that and decide to lower the water instead of raise the bridge in the name of fairness, equality, and uniformity. IOW, reduce V6 credits to match V7. |
Ron Send message Joined: 24 Aug 99 Posts: 42 Credit: 34,544,679 RAC: 0 |
BOINC is software that can use the unused CPU and GPU cycles on a computer. It seems that credits given for said cycles should have the same value whether used for V6 or V7. I sense the staff isn't all that interested in our opinions and are not likely to read my ventings! Here's an example: "Tomorrow I'll post explaining why we wanted version 7 and what's different about the analysis it does. 30 May 2013, 0:39:58 UTC" |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13855 Credit: 208,696,464 RAC: 304 |
I sense the staff isn't all that interested in our opinions and are not likely to read my ventings! Here's an example: "Tomorrow I'll post explaining why we wanted version 7 and what's different about the analysis it does. 30 May 2013, 0:39:58 UTC" You're confusing BOINC with SETI. And if you'd noticed, there have been some issues relating to the new version. Personally i'd prefer the staff work on those issues than spending time writing up something which could be posted at any time. Grant Darwin NT |
kittyman Send message Joined: 9 Jul 00 Posts: 51478 Credit: 1,018,363,574 RAC: 1,004 |
Please do not insult Eric. I would like to see his dissertation on the new v7 work as much as anybody else. But I also know how busy the man is at times. He does have a life outside of Seti too, and more than that, has other scientific endeavors he is involved with. As well as doing other chores such as writing numerous grant proposals to try to keep the doors open at Seti, given the miniscule percentage of participants that donate funds to the project. He has when his time has allowed both listened and responded to my personal inquiries over the years. He has also posted in the threads here in NC from time to time regarding user's posts and questions. "Time is simply the mechanism that keeps everything from happening all at once." |
Donald L. Johnson Send message Joined: 5 Aug 02 Posts: 8240 Credit: 14,654,533 RAC: 20 |
And here is Dr. K's response to the issue of reduced credit and falling RACs... Donald Infernal Optimist / Submariner, retired |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Why are the Astropulse Credits so low again? Nothing was changed with Astropulse, it's still at Version 6. Usually a completed AP task was good for around 700+. Now you're lucky if you see 700. It seems to be coming from being paired with a Host running the Stock GPU App, you're usually awarded higher Credits when paired with a CPU Host. Look at my most recent WU; Workunit 1259671742 3029730600 6891736 6 Jun 2013, 5:26:53 UTC 6 Jun 2013, 9:40:09 UTC Completed and validated 2,484.57 622.52 491.24 AstroPulse v6 v6.04 (opencl_ati_100) 3029730601 6797524 6 Jun 2013, 5:26:38 UTC 8 Jun 2013, 20:22:24 UTC Completed and validated 2,206.98 169.73 491.24 AstroPulse v6 Anonymous platform (ATI GPU) Now look at the Wingperson's 'awards'; 3031207918 1260338286 7 Jun 2013, 9:22:18 UTC 7 Jun 2013, 12:51:56 UTC Completed and validated 2,621.43 813.95 565.73 AstroPulse v6 v6.04 (opencl_ati_100) 3031207912 1260181592 7 Jun 2013, 9:22:18 UTC 8 Jun 2013, 4:08:12 UTC Completed and validated 2,468.16 653.93 539.48 AstroPulse v6 v6.04 (opencl_ati_100) 3031207841 1260338199 7 Jun 2013, 9:22:18 UTC 8 Jun 2013, 4:08:12 UTC Completed and validated 2,474.49 639.07 243.60 AstroPulse v6 v6.04 (opencl_ati_100) 3031051672 1260267898 7 Jun 2013, 6:25:01 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 2,473.56 660.04 259.42 AstroPulse v6 v6.04 (opencl_ati_100) 3031051651 1260267892 7 Jun 2013, 6:25:01 UTC 7 Jun 2013, 12:51:56 UTC Completed and validated 2,456.36 662.26 259.61 AstroPulse v6 v6.04 (opencl_ati_100) 3031004942 1260246782 7 Jun 2013, 5:29:06 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 580.35 97.11 189.12 AstroPulse v6 v6.04 (opencl_ati_100) 3030889757 1260037657 7 Jun 2013, 2:49:46 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 2,682.97 885.76 478.26 AstroPulse v6 v6.04 (opencl_ati_100) 3029970773 1259781698 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 17:31:29 UTC Completed and validated 1,580.83 390.78 164.21 AstroPulse v6 v6.04 (opencl_ati_100) 3029970766 1259781680 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 13:53:20 UTC Completed and validated 2,430.17 617.87 543.04 AstroPulse v6 v6.04 (opencl_ati_100) 3029970719 1259781663 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 17:31:29 UTC Completed and validated 1,573.20 412.51 165.78 AstroPulse v6 v6.04 (opencl_ati_100) 3029730704 1259671777 6 Jun 2013, 5:26:53 UTC 6 Jun 2013, 13:53:20 UTC Completed and validated 2,436.37 601.38 288.99 AstroPulse v6 v6.04 (opencl_ati_100) 3029730600 1259671742 6 Jun 2013, 5:26:53 UTC 6 Jun 2013, 9:40:09 UTC Completed and validated 2,484.57 622.52 491.24 AstroPulse v6 v6.04 (opencl_ati_100) 3029464716 1259554039 5 Jun 2013, 22:53:05 UTC 6 Jun 2013, 5:26:53 UTC Completed and validated 2,497.21 610.88 360.18 AstroPulse v6 v6.04 (opencl_ati_100) 3029464707 1259554037 5 Jun 2013, 22:53:05 UTC 6 Jun 2013, 5:26:53 UTC Completed and validated 2,422.53 594.74 407.55 AstroPulse v6 v6.04 (opencl_ati_100) 3026226363 1257999586 3 Jun 2013, 1:32:08 UTC 3 Jun 2013, 4:22:57 UTC Completed and validated 5,568.82 4,002.19 655.86 AstroPulse v6 v6.04 (opencl_ati_100) 3026044998 1258044551 2 Jun 2013, 23:13:22 UTC 3 Jun 2013, 0:42:20 UTC Completed and validated 2,353.37 621.77 499.96 AstroPulse v6 v6.04 (opencl_ati_100) There's not a single 700 in there. If we have to wade through all these 'Errors', at least award the Normal credit. Again, NOTHING has changed with AstroPulse...except the credits, and the credits are all over the place. |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
Why are the Astropulse Credits so low again?. Because Eric reset the Stock AP stats, now they need to be repopulated?, and every host needs to repopulate their stats too? I've only managed 6 out of 10 v6 AP CPU validations so far, hosts with single or dual cores are going to take weeks or months to get their validations done, Claggy |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
Why are the Astropulse Credits so low again?. I've got a plan that will work much faster. Since I receive the 'normal' credit when my GPU is matched with a CPU; 3029788914 6810201 6 Jun 2013, 6:09:22 UTC 7 Jun 2013, 15:54:43 UTC Completed and validated 98,747.98 93,402.47 723.68 AstroPulse v6 v6.03 3029788915 6797524 6 Jun 2013, 6:08:57 UTC 8 Jun 2013, 21:52:42 UTC Completed and validated 2,590.73 517.73 723.68 AstroPulse v6 Anonymous platform (ATI GPU) 3028974419 6797524 5 Jun 2013, 13:56:17 UTC 8 Jun 2013, 0:54:16 UTC Completed and validated 2,481.36 327.61 751.57 AstroPulse v6 Anonymous platform (ATI GPU) 3028974420 6871372 5 Jun 2013, 13:56:01 UTC 7 Jun 2013, 8:14:43 UTC Completed and validated 104,714.37 104,385.20 751.57 AstroPulse v6 v6.01 I'll just Abort all the Tasks that match my GPU against a low scoring GPU. That should clear that up. Or...someone could tweak the Scheduler to do the same, match a GPU with a CPU until this mess is cleared up. Unless you can explain why that wouldn't work, it sounds like a plan to me. Good Luck explaining why it wouldn't work ;-) |
Lionel Send message Joined: 25 Mar 00 Posts: 680 Credit: 563,640,304 RAC: 597 |
Note: Data for last four days added and revised running average. Thankyou I’ve had a look at the data around v7 and v6 WUs and below is a quick observational analysis of my data. Under v6, I was roughly averaging 100 credits per Work Unit (WU). Under v7, it seems that the average is sitting around 75-80 credits per WU. In looking at run-times and taking the outliers out, cpu run time was around 600-660 seconds (10-11 minutes) per WU for v6, and appears to be around 800-1100 seconds (13+ to 18+ minutes) for a v7 WU. CPU time seems to have gone up by a factor of 2-3 from 50-60 seconds for v6 to 90-180 seconds for v7. So doing a quick Back of the Envelope (10.5/15.83=0.66) shows that from a WU processing/throughput capability, I can expect to do roughly 66% of the volume of WUs that I did before (for example, if I was doing 400 WUs per day under v6, I can now expect to do around 264 WUs per day under v7). Looking at the impact on credit gives 0.66*0.775 = 0.514 or 51.4%. In essence I can expect that daily credit for v7 will drop to circa 51% of what I was getting under v6. I am aware of the comments around “that the system needs time to settle down†and that “it thinks all the WUs coming back at the moment are easy, hence the low credit†however, if the system continues to perform as is, then I can expect to see no change from current trajectory. To test the assumption, I have looked at credit per day for the last 20 days. Below is the data: [list=] 2013.05.16 – 244,130 2013.05.17 – 220,168 2013.05.18 – 231,098 2013.05.19 – 226,353 2013.05.20 – 224,723 2013.05.21 – 210,477 2013.05.22 - 0 2013.05.23 – 431,485 2013.05.24 – 229,312 2013.05.25 – 228,767 2013.05.26 – 239,021 2013.05.27 – 231,271 2013.05.28 – 231,050 2013.05.29 – 0 2013.05.30 – 392,635 2013.05.31 – 209,556 2013.06.01 – 123,072 2013.06.02 – 94,061 2013.06.03 – 102,333 2013.06.04 – 99,896 2013.06.05 - 65,653 2013.06.06 - 112,209 2013.06.07 - 102,538 2013.06.08 - 110,760 2013.06.09 - 89,757 2013.06.10 - 96,018[/list] The average daily credit prior to migration was 221,878. Following migration on 1st June, the average daily credit is showing as 99,629 which is circa 44.9% of the previous daily average under v6. I suspect that many are starting to see their RAC decline, but as RAC is a lagging indicator and is currently composed of v6 and v7 numbers, the end effect is being masked. As this mask evaporates and RAC asymptotes towards the underlying v7 values, I suspect that the grumblings in the forum may get louder. Comment on comments Whilst some are focusing just on credit, the issue is not about credit as such. It’s about recognition. There are many distributed computing projects to which people contribute resources. The manner and means in which those projects recognise individual contribution is through a system that is based on and allocates credits. Some projects choose to recognise a person’s contribution more than other projects, thus they grant a higher credit rate per contribution for that project. In short, credits are effectively an indication of a person’s contribution to a project. In the case of "the New Credit System" implemented by Seti, recognition of personal contribution has been reduced. At present, the indication is that recognition for effort is effectively half that of what it was prior to the new recognition system being employed. |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
... Although a CreditNew design goal was to have credits comparable between projects, the implementation clearly hasn't achieved that. And the change from Sah Enhanced to Sah v7 seems to have demonstrated that even for very similar work the system doesn't stabilize at the same credit rate. Those are BOINC issues, and perhaps the close relationship between Sah and BOINC will allow improvements to flow back into BOINC. What I see as a possible short-term improvement would be to recognize the contribution in a different form, RAC relative to the total project RAC. That way if the overall RAC declines by 50% and an individual's RAC also by 50%, it is obvious that user is still contributing as before. It would need to be scaled up more than just converting to percent, the current top host is doing about 0.1% of the total RAC and most hosts far far less. Users wouldn't be comfortable with scoring having a lot of zeroes after the decimal indicator. Then a name for that score is needed... Joe |
betreger Send message Joined: 29 Jun 99 Posts: 11416 Credit: 29,581,041 RAC: 66 |
I'll just Abort all the Tasks that match my GPU against a low scoring GPU. That should clear that up. Or...someone could tweak the Scheduler to do the same, match a GPU with a CPU until this mess is cleared up. Unless you can explain why that wouldn't work, it sounds like a plan to me. Good Luck explaining why it wouldn't work ;- TBar, not a good idea, the data base needs all results so it has the opportunity to get squared away. |
TBar Send message Joined: 22 May 99 Posts: 5204 Credit: 840,779,836 RAC: 2,768 |
I'll just Abort all the Tasks that match my GPU against a low scoring GPU. That should clear that up. Or...someone could tweak the Scheduler to do the same, match a GPU with a CPU until this mess is cleared up. Unless you can explain why that wouldn't work, it sounds like a plan to me. Good Luck explaining why it wouldn't work ;- The Database need the correct results so it has the opportunity to get squared away. By assuming Astropulses complete rather quickly on all hardware, ie GPUs, it is gathering Incorrect Data. The Database needs to be informed that Astropulses take much longer on most CPUs than GPU tasks. As long as the Database keeps accumulating the Wrong Assumptions, the problem will continue to worsen. By matching the same tasks on different hardware, the Database May get a clue. Someone around here needs to get a clue. If you look at the large differences in Astropulse Credits, it's clear the Credit Guesser doesn't have a clue. How can you rationalize getting 200 pts on one AP task and 700 on another? Look around, it's not that difficult. The current AP credits are a joke. 3031051569 1260267858 7 Jun 2013, 6:24:49 UTC 10 Jun 2013, 21:58:42 UTC Completed and validated 18,283.23 1,657.85 213.96 AstroPulse v6 v6.04 (opencl_nvidia_100) 3024643812 1257437959 1 Jun 2013, 17:33:21 UTC 10 Jun 2013, 21:58:42 UTC Completed and validated 163,910.69 114,720.00 753.72 AstroPulse v6 v6.01 3031207918 1260338286 7 Jun 2013, 9:22:18 UTC 7 Jun 2013, 12:51:56 UTC Completed and validated 2,621.43 813.95 565.73 AstroPulse v6 v6.04 (opencl_ati_100) 3031207912 1260181592 7 Jun 2013, 9:22:18 UTC 8 Jun 2013, 4:08:12 UTC Completed and validated 2,468.16 653.93 539.48 AstroPulse v6 v6.04 (opencl_ati_100) 3031207841 1260338199 7 Jun 2013, 9:22:18 UTC 8 Jun 2013, 4:08:12 UTC Completed and validated 2,474.49 639.07 243.60 AstroPulse v6 v6.04 (opencl_ati_100) 3031051672 1260267898 7 Jun 2013, 6:25:01 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 2,473.56 660.04 259.42 AstroPulse v6 v6.04 (opencl_ati_100) 3031051651 1260267892 7 Jun 2013, 6:25:01 UTC 7 Jun 2013, 12:51:56 UTC Completed and validated 2,456.36 662.26 259.61 AstroPulse v6 v6.04 (opencl_ati_100) 3031004942 1260246782 7 Jun 2013, 5:29:06 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 580.35 97.11 189.12 AstroPulse v6 v6.04 (opencl_ati_100) 3030889757 1260037657 7 Jun 2013, 2:49:46 UTC 7 Jun 2013, 9:22:18 UTC Completed and validated 2,682.97 885.76 478.26 AstroPulse v6 v6.04 (opencl_ati_100) 3029970773 1259781698 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 17:31:29 UTC Completed and validated 1,580.83 390.78 164.21 AstroPulse v6 v6.04 (opencl_ati_100) 3029970766 1259781680 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 13:53:20 UTC Completed and validated 2,430.17 617.87 543.04 AstroPulse v6 v6.04 (opencl_ati_100) 3029970719 1259781663 6 Jun 2013, 9:40:09 UTC 6 Jun 2013, 17:31:29 UTC Completed and validated 1,573.20 412.51 165.78 AstroPulse v6 v6.04 (opencl_ati_100) 3029730704 1259671777 6 Jun 2013, 5:26:53 UTC 6 Jun 2013, 13:53:20 UTC Completed and validated 2,436.37 601.38 288.99 AstroPulse v6 v6.04 (opencl_ati_100) 3029730600 1259671742 6 Jun 2013, 5:26:53 UTC 6 Jun 2013, 9:40:09 UTC Completed and validated 2,484.57 622.52 491.24 AstroPulse v6 v6.04 (opencl_ati_100) 3029464716 1259554039 5 Jun 2013, 22:53:05 UTC 6 Jun 2013, 5:26:53 UTC Completed and validated 2,497.21 610.88 360.18 AstroPulse v6 v6.04 (opencl_ati_100) 3029464707 1259554037 5 Jun 2013, 22:53:05 UTC 6 Jun 2013, 5:26:53 UTC Completed and validated 2,422.53 594.74 407.55 AstroPulse v6 v6.04 (opencl_ati_100) 3026226363 1257999586 3 Jun 2013, 1:32:08 UTC 3 Jun 2013, 4:22:57 UTC Completed and validated 5,568.82 4,002.19 655.86 AstroPulse v6 v6.04 (opencl_ati_100) 3026044998 1258044551 2 Jun 2013, 23:13:22 UTC 3 Jun 2013, 0:42:20 UTC Completed and validated 2,353.37 621.77 499.96 AstroPulse v6 v6.04 (opencl_ati_100) |
betreger Send message Joined: 29 Jun 99 Posts: 11416 Credit: 29,581,041 RAC: 66 |
TBar, IMO all the data needs to be collected or it is not as valid as it could be. Your arbitrary editing of the incoming data screws it up to some small degree. |
©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.