Message boards : Number crunching : Stats a bit haywire after model upload...
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Send message Joined: 17 Sep 04 Posts: 10 Credit: 314,724 RAC: 0 |
> yeah... something still isn't right IMHO. Look at <a> href="http://macg.no-ip.info:5520/boinc/team_hist_graphs.php?proj=cpdn&teamid=45&pmode=r">this > graph</a> on my stats site. You can clearly see several big jumps when people > return work units. If the credit that is granted along the way doesn't match > up to the final claimed credit, I could see some adjustment going on when you > finish but this seems a bit extreme. Mine shows a very good example for the bad <a href="http://macg.no-ip.info:5520/boinc/team_hist_graphs.php?proj=cpdn&teamid=1122&pmode=r"> RAC</a>, as does this <a href="http://macg.no-ip.info:5520/boinc/team_hist_graphs.php?proj=cpdn&teamid=1122&pmode=dp"> graph</a> for Daily Prod. I normally average about 600 credits a day and it jumped to almost 4800. This was a day after I had 2 boxes puke and lost the WUs and each started over with new WUs. |
Send message Joined: 10 Sep 04 Posts: 3 Credit: 653,178 RAC: 0 |
I noticed some of the huge jumps on the leader board as well (happening on an almost daily basis). Couldn't work it out. But after discovering this thread did a bit of looking around. Could it be caused when a w/u is completed with errors? I completed a w/u the other day *my first he says with a big grin*. <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=267689"> Result 267689 </a> It got an outcome of "Success" and there does not appear to be any incorrect credit given. I found this other completed w/u from last night <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=33983"> Result 33983 </a> It got an outcome of "Client error" and a huge jump in credit. Prior to the 72nd (last) trickle being posted the total credit was 5369 (exactly correct for being one timeunit short of the standard 5444 credits. After posting the last trickle, a total of 10748 extra credits were granted (even though the record only shows 5444 credits for the w/u). This is very close to double the 5369 they were on previously. The only other difference I can see between the two work units is the creation dates. 25th Aug vs 6th Sep. Maybe the old w/u caused the error. A good mystery is alway fun. :) Another result that has this week finished in error and generated a 8000+ credit bonus' <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=31559"> Result 31559 </a> |
Send message Joined: 10 Sep 04 Posts: 3 Credit: 653,178 RAC: 0 |
LOL. So much for those theories. Here is another w/u that got completed on the 27/10. <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=263441"> Result 263441 </a> It was created on the same day as mine (only 14 minutes earlier) and completed on the same day as mine (just 12 hours later). It achieved an outcome of "Success". And the total credit at the end of the 27/10 was 10888 (2 x 5444). All looks good. Then the next day (on the 28/10) they get an extra 6404 credit. And the really funny thing about that is that they didn't post any results on the 28/10 at all. All very strange :) |
Send message Joined: 5 Aug 04 Posts: 1496 Credit: 95,522,203 RAC: 0 |
Hi, everyone, On a whim, I decided to follow the lead of others and put the numbers from my 'Results' page into a calculator. 129,980.49 Granted Credit 139,033.62 Top Participants page 9,053.13 Excess over Granted Credit Given that I received ~2,500 extra on 26 September, that means the two runs on 24 October didn't each receive a full 5444.21 extra credit -- as I thought they did. (Or, I may have fat-fingered the numbers for the calculator.) It seems the extra credit may sometimes be 5444.21 but be a random number at other times. Strange. (Random gremlins are the worst kind, eh?) We have met the enemy and he is us -- Pogo |
Send message Joined: 5 Aug 04 Posts: 907 Credit: 299,864 RAC: 0 |
so I think people have reported these credit anomalies from other BOINC projects? I still think they are "leftovers" from when I botched credits up in September (when I thought I would "correct" things). I guess that wouldn't be true if an entirely new run has started and finished since when I tried to correct the old credits. I may update the server software (waiting to hear from David A. of BOINC/SETI) soon so maybe that will fix things. I'm gone in 4 weeks so I would rather do it now and iron things out than wait too much longer. |
Send message Joined: 17 Aug 04 Posts: 56 Credit: 63,814 RAC: 0 |
Hmm. I didn't think about your tinkering possibly causing this... I'm certain that the model that got uploaded by my team mate was downloaded before your fix so that could still explain it. Don't know about others. The ones linked to in this thread are a little of both. Guess you have more data than us so we leave it in your capable hands :) <br> ---------------------------- A member of <a href="team_display.php?teamid=45">The Knights Who Say Ni!</a> Yet another stats page: <a href="http://boinc-kwsn.no-ip.info">http://boinc-kwsn.no-ip.info</a> |
Send message Joined: 6 Sep 04 Posts: 4 Credit: 9,798,557 RAC: 0 |
Oh K. This then explains why I am seeing people with a similar level of machine as mine getting credits of 6000 per day and yet I only achieve an average of 300 a day on the exactely same machine and O.S. So how does this magically get fixed or just the luck of the draw ? |
Send message Joined: 10 Sep 04 Posts: 3 Credit: 653,178 RAC: 0 |
Yes, the 6000 credits per day was what grabbed my attention. Hehe. Hard not to notice when someone leapfrogs you in the list coming out of no where. :) If the root of the problem has been fixed and we are just seeing some trailing odd results work their way out of the system, then it should be possible to recalculate all our totals based on our actual credits from the individual workunits. Not that I know anything. That would be up to system admin. Though, looking at some of the funny results, I'm not sure we can assume that the problem has been resolved. I have found 3 w/u's that were all created on 6th Sep (within a 40 minute period) and returned 27th Oct. All 3 returned a "successful" result. The 1st and 3rd ones got heaps of extra credit but the middle one did not. Does that indicate the Boinc version being used could make a difference? My w/u started off at version 4.05, got upgraded to 4.09 in the middle and again to 4.13 before it completed. Maybe others could post here what versions they were running and whether their final result was granted "extra" credits... |
Send message Joined: 6 Sep 04 Posts: 4 Credit: 9,798,557 RAC: 0 |
So what is going on ? Are the accounts that got the extra credits going to have them removed or what ? I know of several who jumped up major amounts without actual working being done just getting extra credits. |
Send message Joined: 17 Sep 04 Posts: 10 Credit: 314,724 RAC: 0 |
A few days ago another WU (257473) went belly up just after the third trickle of phase 3. Tonight I noticed a hugh jump again in <a href="http://macg.no-ip.info:5520/boinc/tgraphs.php?proj=cpdn&teamid=1208">RAC and Total Credit</a>. I've noticed similar jumps with a few other crunchers also. It makes the stats look good but offers no incentive to complete and turn in a valid WU.I figure about 4 or 5 more of these bad WUs and I'll be in the top 5 or 10. Is this extra credit normal even though the WU did not finish? sholzy <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=358"> |
©2024 cpdn.org