Message boards :
Number crunching :
pending credit problems...again
Message board moderation
Author | Message |
---|---|
makosky Send message Joined: 7 Jul 00 Posts: 56 Credit: 3,908,782 RAC: 0 |
a couple days ago we had a a great catch uo the couple days that folled was not great , but also good but for the past couple days work units seem to be taking longer to complete and pending credit is climbing quickly when from 1300 to 2100 in two days .. total of w/u completed each day is dropping quickly ... my pc seems to be working fine , i have changed nothing ...i use windows xp...also on some of my wu's i notice i have been getting an unusual ammout of computation errors ...and the the number of work units that are waiting to start has also dropped to just one or two normally 4-8 depending on size ...any one else going through this or is it just me , grrr .. thanks |
Valerie Chilton Send message Joined: 14 Jan 08 Posts: 53 Credit: 20,682,710 RAC: 0 |
I'm not seeing another rise in pending credits at this time. I'm usually sitting on 23k to 26K and I'm within those numbers now. During the server problems I got to around 47K. Perhaps your wingmen are at the tavern or something :) |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
A few days ago, the project staff put an almost consecutive run of tapes from 30 March 2007 to 3 April 2007 onto the Server status page. During that period, according to the Arecibo telescope observing schedule, the main observing project was the Arecibo Legacy Fast ALFA Survey - a project which I once falsly accused of generating a high percentage of 'shorty' tasks, but I was swiftly corrected by Joe Segur. In fact, this is a 'drift' survey, generating workunits in the ~0.39 Angle Range. These tasks take approximately six times as long to process as the 'shorty' tasks you may have become used to from projects like the GALFA Study of the Disk-Halo Interface and Mapping HI in a spectacular shell. That explains why your tasks are taking longer to complete, and you are getting through fewer of them: the increase in pending credit is probably because your wingmates are also taking longer to get through the work. Among all this, there are also faulty workunits in circulation from a tape (13fe08ac) not shown on the server status page. See the discussion in Computation Error - Bad Workunit Header. If the computation errors you are seeing come from this tape, don't worry - it's the project's fault, not your computer's. Edit - yes, your computation errors are coming from the tape which was split wrongly. |
David Send message Joined: 19 May 99 Posts: 411 Credit: 1,426,457 RAC: 0 |
Thanks Richard. It appears as though I only got a few of the "bad WU's" but I have noticed a few longer WU's beginning to creap into the list. I dont mind them on my 24/7 PC's but preferred the shorter WU's on the PC that only runs occasionally (Vixen) as it might only be used for a few hrs here & there (It's located in the shed and is only on when I am out there). Ah well, once the longer WU's are processed, back to a nice mix once again :) Edit: And back on topic, yes my pending has risen from a normal 12,000 to 16,000 last time I looked |
KWSN Ekky Ekky Ekky Send message Joined: 25 May 99 Posts: 944 Credit: 52,956,491 RAC: 67 |
On a slightly different question of pending credit, can anyone tell me what happens in cases like this: created 8 Mar 2008 1:47:44 UTC name 02ap07ac.28710.2935.4.7.2 777977608 3554041 8 Mar 2008 20:35:51 UTC 9 Mar 2008 2:31:25 UTC Over Success Done 18,562.86 66.98 pending 779063532 4069706 10 Mar 2008 0:11:37 UTC 3 Apr 2008 2:01:22 UTC In Progress Unknown New --- --- --- 777977609 3806766 8 Mar 2008 20:36:04 UTC 8 Mar 2008 20:55:28 UTC Over Success Done 132.96 0.05 pending I anticipate that when I have completed this unit my result will be the same as the first result. Is the last one some sort of cheat? Will he get the same credit for a tiny fraction of the time? I know little about such matters but is this person gaining credit from the work of others and being of zero benefit to the science as a result? I note that all his (very few) results are claiming low/much lower credit than his wingmen. What can be done about this sort of apparent fraud? |
W-K 666 Send message Joined: 18 May 99 Posts: 19062 Credit: 40,757,560 RAC: 67 |
On a slightly different question of pending credit, can anyone tell me what happens in cases like this: Think you are safe on that WU 234561722. The computer that only did a few sec for a -9 overflow has a problem, that needs sorting. The third result will probably agree with you and grant credit accordingly. The validator will fail the other result. |
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
On a slightly different question of pending credit, can anyone tell me what happens in cases like this: wuid=234561722 It's looking like that it's the 'bug' of the actually app.. WinVista and stock app BUT, little funny because of the results.. <core_client_version>5.8.15</core_client_version> <![CDATA[ <stderr_txt> setiathome_enhanced 5.27 libboinc: setiathome_enhanced 5.27 Work Unit Info: ............... WU true angle range is : 0.395844 Optimal function choices: ----------------------------------------------------- name ----------------------------------------------------- v_BaseLineSmooth (no other) v_vGetPowerSpectrumUnrolled 0.00053 0.00000 sse2_ChirpData_ak 0.02319 0.00000 v_vTranspose4x16ntw 0.01386 0.00000 BH SSE folding 0.00174 0.00000 SETI@Home Informational message -9 result_overflow NOTE: The number of results detected exceeds the storage space allocated. Flopcounter: 20300727448137.894531 [b]Spike count: 18 Pulse count: 6 Triplet count: 1 Gaussian count: 6[/b] </stderr_txt> <core_client_version>5.10.20</core_client_version> <![CDATA[ <stderr_txt> setiathome_enhanced 5.27 DevC++/MinGW Work Unit Info: ............... WU true angle range is : 0.395844 Optimal function choices: ----------------------------------------------------- name ----------------------------------------------------- v_BaseLineSmooth (no other) v_vGetPowerSpectrumUnrolled 0.00067 0.00000 sse1_ChirpData_ak 0.06037 0.00000 v_vpfTranspose8x4ntw 0.01728 0.00000 AK SSE folding 0.00324 0.00000 SETI@Home Informational message -9 result_overflow NOTE: The number of results detected exceeds the storage space allocated. Flopcounter: 10534206839.728594 [b]Spike count: 0 Pulse count: 31 Triplet count: 0 Gaussian count: 0[/b] </stderr_txt> |
David Send message Joined: 19 May 99 Posts: 411 Credit: 1,426,457 RAC: 0 |
My pending is up to just over 18,000 again (12K normally) so it looks like I'm eithor waiting for the wingmen to complete longer tasks, or I'm just unlucky... Eithor way, Grrrr, as that 18K ALMOST gets me to 300,000 credits ;) OK another 3 to 4 days should do it anyway... |
Mumps [MM] Send message Joined: 11 Feb 08 Posts: 4454 Credit: 100,893,853 RAC: 30 |
Actually, the bastard WU's from 13Feb08 may be partially to blame. I just had two of them on one of my baby crunchers (P4 1.8 Ghz, running Crunchers SSE2 2.4V version) that didn't just gracefully abort. The system started the WU and then hung there. Accumulating no CPU and getting nothing accomplished. Even though Task Manager indicated the task was consuming all avialable CPU. I had to manually abort both of those tasks. Hopefully we don't have many unattended crunchers stuck in that same rut... |
yank Send message Joined: 15 Aug 99 Posts: 522 Credit: 22,545,639 RAC: 0 |
Just really don't want to bring up this 'dead horse' again...but I was just checking out some pending credits and came across this wingman on one of the units. Subject was give 950 work units...640 work units on 7 Feb o8 and another group of 310 work units on 8 Feb 08. The computer is a Quad but seemingly...to me...very low activity. Perhaps someone can explain this to me. workunit http://boinc.mundayweb.com/teamStats.php?userID=14824 |
Mumps [MM] Send message Joined: 11 Feb 08 Posts: 4454 Credit: 100,893,853 RAC: 30 |
Just really don't want to bring up this 'dead horse' again...but I was just checking out some pending credits and came across this wingman on one of the units. Subject was give 950 work units...640 work units on 7 Feb o8 and another group of 310 work units on 8 Feb 08. The computer is a Quad but seemingly...to me...very low activity. Perhaps someone can explain this to me. Well, I probably can't explain it to you, but I'll share how I interpret what I see. I pulled up this host using boincstats and see it was freshly "Comissioned" on Feb 8th. Even though it has credits from January. And there's been just a trickle of WU's being credited. If it's a new host, with an as-of-yet undetermined DCF, it may have grabbed a large queue, but then is only getting a very small percentage of time to crunch. But then the system is discarding (maybe reseting?) and losing most of the WU's. I see a 13 day run with no granted credit. Followed by another of 3 and the current workless run of 9 more days. And those 2 bumps could almost be attributed to work turned in back on Feb 8th that is only now having it's wingman reporting in. Last bad bit of news is if you go look at http://setiathome.berkeley.edu/hosts_user.php?userid=8354686 you'll see that host is no longer even showing as associated with the owning user. So it must be detached again and those work units most likely will never be coming back... |
David Send message Joined: 19 May 99 Posts: 411 Credit: 1,426,457 RAC: 0 |
Actually, the bastard WU's from 13Feb08 may be partially to blame. I just had two of them on one of my baby crunchers (P4 1.8 Ghz, running Crunchers SSE2 2.4V version) that didn't just gracefully abort. The system started the WU and then hung there. Accumulating no CPU and getting nothing accomplished. Even though Task Manager indicated the task was consuming all avialable CPU. I had to manually abort both of those tasks. Hopefully we don't have many unattended crunchers stuck in that same rut... I wonder if this is affecting more people that they realise. Unless you were actively watching, you may not realise the issue has happened. My pending is over 20K now, but looking at some of the WU's, it appears as though a lot of the wingmen have stopped crunching, or at least seriously slowed down the S@H crunching, so I have to wait until the WU's time out & then get resent to someone else... No big deal, but it appears to becoming a more common problem. Add that to the -9 crashes for noisy WU's on Vista (I seem to have had my share of wingmen that experienced that too lol), and it all adds up. |
AndyW Send message Joined: 23 Oct 02 Posts: 5862 Credit: 10,957,677 RAC: 18 |
|
KWSN Ekky Ekky Ekky Send message Joined: 25 May 99 Posts: 944 Credit: 52,956,491 RAC: 67 |
Interesting to see that my usual paltry credit pending has fallen from an all-time high of 2,6000 to around 1,600. No one moans when pending credit falls! |
AndyW Send message Joined: 23 Oct 02 Posts: 5862 Credit: 10,957,677 RAC: 18 |
|
Dr. C.E.T.I. Send message Joined: 29 Feb 00 Posts: 16019 Credit: 794,685 RAC: 0 |
Pending credit: 1,893.66 . . . dropping ever so slowly - though < it was around 3-4000 last week . . . Thanks Wingmen (women) ;) . . . BOINC Wiki . . . Science Status Page . . . |
Dr. Bob Send message Joined: 1 Apr 03 Posts: 78 Credit: 623,977 RAC: 0 |
After reading the posts here, I am beginning to see why my 9 work units have about 5 that are all pending though completed. The earliest is Feb 8, one of the times mentioned when some "bad" WUs were released in the system, I believe. Shame it is and I hope everything catches up and is corrected. While it is nice to get the credit, it is most important to the project to finish the crunching for good information to be usable for analysis. I wouldn't be opposed to deleting the pending WUs for me and trying to do them over, if the powers that be decide on that course of action. Those who work on SETI should be commended for their commitment to scientific inquiry; best wishes... Dr. Bob Robert L. Hanson, Ed.D. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
After reading the posts here, I am beginning to see why my 9 work units have about 5 that are all pending though completed. The earliest is Feb 8, one of the times mentioned when some "bad" WUs were released in the system, I believe. Shame it is and I hope everything catches up and is corrected. While it is nice to get the credit, it is most important to the project to finish the crunching for good information to be usable for analysis. I wouldn't be opposed to deleting the pending WUs for me and trying to do them over, if the powers that be decide on that course of action. Bob, You're looking at this way too closely. The "powers that be" is a very small group of people. They've set some policies, written code to implement those policies, and from there, it is up to BOINC to "decide." They won't look at those units until they've been sorted out by BOINC, and the applications that search the resulting signals (the near-time persistency checker). "Pending credit" generally means you returned the work faster than someone else. If the result doesn't match, it may be "pending" a bit longer while a third result is crunched as a "tiebreaker." It may be pending a long time if the other machine (your "wingman") goes missing, or if the work is otherwise lost. I looked at the five work units you have that are pending, and in all five cases, you're the first to return the work, and you'll get credit when a second machine returns a matching result. Deleting your pending work would be a complete waste: that would mean a third computer would have to crunch them again to pair-up for a proper result. It would mean that your wingman would have pending work units until that third computer finished it. You can run a bigger cache (so you keep work longer and return it later) or you can not worry about it. I'd recommend not worrying. -- Ned |
Mike O Send message Joined: 1 Sep 07 Posts: 428 Credit: 6,670,998 RAC: 0 |
Are all WUs that have the prefix of 13fe08 bad? I have a few Q'ed to be started... 13fe08ac on two of my machines and 13fe08ab on one. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Are all WUs that have the prefix of 13fe08 bad? I have a few Q'ed to be started... 13fe08ac on two of my machines and 13fe08ab on one. No. There was a bad batch of specifically the 13fe08ac variety, but the problem has been fixed and they are being re-issued. All others (13fe08ab) should be OK. |
©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.