|
1)
Message boards :
Number crunching :
Can't get work
(Message 21284)
Posted 3 Mar 2009 by Mike.Gibson Post: There has been a trickle of work but I can\'t get any of it. Sometimes it is because it arrives between updates and disappears before the next scheduled update. Other times when I issue an update and there is a little work then it rejects because it is less than 15 mins since the last update. Can\'t the 15 min. restriction be based on when we last get work rather than the last update which failed to get work, please? |
|
2)
Message boards :
Number crunching :
WUs returning zero credit despite successful completion
(Message 21283)
Posted 3 Mar 2009 by Mike.Gibson Post: Because 10 people have had errors the WU has been aborted, so you will not get any credit for it. Credits are only granted when a quorum of correct results have been received. |
|
3)
Message boards :
Number crunching :
Long Term Debt
(Message 21183)
Posted 16 Feb 2009 by Mike.Gibson Post: Depends which version of BOINC you have. There is a bug in some of the 6.6.x alpha releases. I also believe they changed the LTD calc to take into account when a project has no work for extended periods of time (but I could be making that up). Thanks Gas Giant, version is 6.4.5 I have had the reverse problem in the past when I had a much higher LTD and it was preventing me from getting anything else although nothing was available from LHC. The current situation is much better than that, but it doesn\'t allow a big backlog to build up so that I can catch up later. However, we don\'t get enough WUs to fulfil the resource share that I have set, if backlogs are to count. |
|
4)
Message boards :
Number crunching :
Long Term Debt
(Message 21181)
Posted 15 Feb 2009 by Mike.Gibson Post: I have my resource share set for LHC to crunch for an average of 6 hours per day whenever there is work. This equates to 21,600 seconds. My LTD was rising by about that sort of number until recently when it suddenly started to rise by only about 200 per day, although I haven\'t had any work to crunch for about 2 weeks. My LTD amount is only just over +80.000 so is not excessive. Can anyone explain what is happening, please? |
|
5)
Message boards :
Number crunching :
Problem with WU
(Message 21152)
Posted 31 Jan 2009 by Mike.Gibson Post: If it was aborted by the project, it hadn\\\'t started running yet. So, there is no credit to lose. There are some which slip through the net. If a WU has started but there hasn\'t been an update since it started, then it can still abort because the project doesn\'t know it that it has started. Or it may be the BOINC Platform rather than the project. Either way, hit the update and it doesn\'t abort. |
|
6)
Message boards :
Number crunching :
Problem with WU
(Message 21147)
Posted 31 Jan 2009 by Mike.Gibson Post: I had 2 WU\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\'s finish so far and they both were \\\\\\\\\\\\\\\"Aborted by project\\\\\\\\\\\\\\\"; why is this? Will I still get whatever credit(s)? It is not quite as simple as rembertw said. Each WU is sent out to 5 people and once 3 identical replies have been received, an abort order is sent out to anyone who they don\\\'t know as having started the WU. The idea is so that you can move onto a more productive WU. Unfortunately they don\\\'t know that a WU has started unless you have had a project update since it was started. You will not get any credits for those WUs. If they know that someone has started the WU, they are allowed to continue and get the credit. Many people have their computers on permanently and have a high resource share for LHC@home, so they pick up the WUs as soon as they become available and start crunching immediately. Many WUs also finish very quickly so it can be that the Abort message is issued only half an hour after the WU has been issued. |
|
7)
Message boards :
Number crunching :
Is it possible to inform us..
(Message 20809)
Posted 23 Nov 2008 by Mike.Gibson Post: We have just had 100,000 WUs dished out on a first 3 from 5 basis to get results back quicker and then the website was taken off-line. What is the sense in that? If there is a 25% error rate then that means you need an IR of 4. Or it did when I was at school many moons ago. You can always issue an extra if one of the first 3 returns is an error in case number 4 is slow. It might increase the return time slightly for individual WUs but it would mean more WUs could be started sooner so should result in a better overall rate of completion. |
|
8)
Message boards :
Number crunching :
Please note: this project rarely has work
(Message 20703)
Posted 28 Oct 2008 by Mike.Gibson Post: Jim This was my first time. With 600,000 seconds owed to LHC, all my projects had hit their limit of time owed. All I had left was CPDN coupled modules but they take over 100 days of continuous running to finish. I\\\\\\\'ve had my debts all one sided for LHC at one point, but never had the problem that you described. All of the other projects requested work when they needed it. I never ran out of work. |
|
9)
Message boards :
Number crunching :
Please note: this project rarely has work
(Message 20696)
Posted 26 Oct 2008 by Mike.Gibson Post: Hi, Folks I only have LHC set to a resource share of 12.5% of total capacity but, because there has been so little work available, my LTDs fouled up my Boinc Manager to such an extent that it wouldn\'t ask for any work for anything but LHC and that had nothing to give. This was such a waste of resources that I contacted Boinc and the stuff below is the correspondence which has cured my lockup. Please note that you need to close down Boinc Manager before starting BoincDV. Apologies for the sequence being in reverse order. Message Mike Gibson Send message Joined: Mar 5 08 Posts: 13 ID: 3437 Message 20986 - Posted 26 Oct 2008 10:20:43 UTC - in response to Message 20981. Edit Thanks,Kathryn I had just closed down Boinc Manager when BoincDv told me to. Everything worked perfectly when I closed Boinc down, then made breakfast and then ran BoincDv. I still reckon Boinc Manager should allow 1 WU per project even when the debts are high. Cheers ID: 20986 | Rating: 0 | rate: / ReplyQuote Dagorath Send message Joined: Jun 13 07 Posts: 385 ID: 2516 Message 20982 - Posted 26 Oct 2008 4:56:39 UTC - in response to Message 20980. 1) Did you read the read.me that comes with BoincDV? 2)Maybe you didn\'t point BoincDV to the right directory? Maybe you pointed it to that other BOINC directory you have hanging around? 3) When you run BoincDV, does it give any feedback or any indication that it was able or unable to find client_state.xml and reset the debts? 4) BOINC keeps a backup of client_state.xml in the BOINC data directory. The backup is named client_state_prev.xml. If BOINC thinks something is not quite right it sometimes restores client_state.xml from client_state_prev.xml. Maybe that\'s how it\'s putting the old debts back. 5) Maybe you\'re trying to reset the debts while BOINC client is still running? Quit BOINC client first by clicking Advanced -> Shutdown connected client. Then exit the manager by clicking File -> Exit. I strongly encourage you to make a backup copy of the BOINC data directory so that if you screw something up you won\'t lose your climate prediction model. Shutdown the client as described above before you make the backup. To make the backup just drag a copy of the BOINC data folder to an empty area, Windows will tell you a folder of the same name already exists and ask for a new name, just call the copy BOINC-BKP. Now, with BOINC client shutdown, delete client_state_prev.xml. Then run BoincDV. Watch for any messages BoincDV gives you. If it says anything about error or failed or anything that makes you think it didn\'t work then write it down and report the message here. After you run BoincDV then start BOINC manager which will start the client. ID: 20982 | Rating: 0 | rate: / ReplyQuote KSMarksPsych Forum moderator Project administrator Send message Joined: Oct 30 05 Posts: 961 ID: 229 Message 20981 - Posted 26 Oct 2008 4:55:23 UTC - in response to Message 20980. Jord Thanks. It seems that the current files were hidden . I must have found an early version which was not hidden. However, when I clear all debts on the current files and restart Boinc Manager, I still can\'t get Einstein & Rosetta WUs and then when I check BoincDv I find that all the debts have been restored. What can I be doing wrong, please? Mike Make sure BOINC is stopped, not just the manager is shut down. ____________ Kathryn :o) The BOINC FAQ Service The Unofficial BOINC Wiki The Trac System More BOINC information than you can shake a stick of RAM at. ID: 20981 | Rating: 0 | rate: / ReplyQuote Mike Gibson Send message Joined: Mar 5 08 Posts: 13 ID: 3437 Message 20980 - Posted 26 Oct 2008 2:34:24 UTC - in response to Message 20978. Last modified: 26 Oct 2008 2:35:49 UTC Jord Thanks. It seems that the current files were hidden . I must have found an early version which was not hidden. However, when I clear all debts on the current files and restart Boinc Manager, I still can\'t get Einstein & Rosetta WUs and then when I check BoincDv I find that all the debts have been restored. What can I be doing wrong, please? Mike ID: 20980 | Rating: 0 | rate: / ReplyQuote Ageless Forum moderator Project administrator Send message Joined: Aug 29 05 Posts: 2903 ID: 8 Message 20979 - Posted 26 Oct 2008 1:43:07 UTC - in response to Message 20977. I tried BoincDv but found that the client_state files hadn\'t been updated since last November when I was only crunching 3 projects (CPDN, Seti & WCG). As it did not include LHC, I thought I had better not proceed with that. Make sure you check for the client_state.xml file in your BOINC Data directory. By default this one can be found at: Windows 2000/XP: C:\\Documents and Settings\\All Users\\Application Data\\BOINC\\ Windows Vista: C:\\ProgramData\\BOINC\\ (*) (*) It may be hidden, so make sure you changed Windows Explorer to show hidden files and directories. ____________ Jord BOINC FAQ Service ID: 20979 | Rating: 0 | rate: / ReplyQuote Dagorath Send message Joined: Jun 13 07 Posts: 385 ID: 2516 Message 20978 - Posted 26 Oct 2008 0:40:39 UTC - in response to Message 20977. Dagorath I tried BoincDv but found that the client_state files hadn\'t been updated since last November when I was only crunching 3 projects (CPDN, Seti & WCG). Impossible. I have no idea which files you looked at but client_state.xml gets updated every few minutes if BOINC is crunching. As it did not include LHC, I thought I had better not proceed with that. I then went onto the Projects page of Boinc Manager and Reset Project on LHC as I had no WUs on it. This resulted in several WUs downloading from WCG only. Still nothing on Einstein & Rosetta or Seti (but I wasn\'t worried about Seti as I still had an Astropulse WU). Doing Reset Project on Einstein & Rosetta still didn\'t result in any downloads. Help! Lots of people think that if they don\'t have a task from each and every one of their projects in their list of tasks then something is wrong. Abandon that thought completely. I can think of 2 explanations for your current state. Neither of them is wrong, bad, abnormal or indicative of a bug. 1) the LTDs are still such that BOINC thinks it should not crunch Einstein and Rosetta for a while in order to respect your resource shares 2) BOINC thinks that if it downloads more work then 1 or more of your tasks will miss deadline ID: 20978 | Rating: 0 | rate: / ReplyQuote Mike Gibson Send message Joined: Mar 5 08 Posts: 13 ID: 3437 Message 20977 - Posted 26 Oct 2008 0:01:04 UTC - in response to Message 20974. Dagorath I tried BoincDv but found that the client_state files hadn\'t been updated since last November when I was only crunching 3 projects (CPDN, Seti & WCG). As it did not include LHC, I thought I had better not proceed with that. I then went onto the Projects page of Boinc Manager and Reset Project on LHC as I had no WUs on it. This resulted in several WUs downloading from WCG only. Still nothing on Einstein & Rosetta or Seti (but I wasn\'t worried about Seti as I still had an Astropulse WU). Doing Reset Project on Einstein & Rosetta still didn\'t result in any downloads. Help! ID: 20977 | Rating: 0 | rate: / ReplyQuote Mike Gibson Send message Joined: Mar 5 08 Posts: 13 ID: 3437 Message 20975 - Posted 25 Oct 2008 21:30:23 UTC - in response to Message 20974. Thanks, Dagorath I would have thought that BOINC Manager should be able to request a single WU per project where LTD is excessive and that project has no work. That way we would be able to continue crunching something until the owed project produced work. at which point BOINC Manager would then be able to restrict the others until either the LTD was cleared or the project ran out of work again. It should be a simple matter to enable just 1 WU to be downloaded when a project has run out of work. An automatic request for, say, 5 seconds of work should do the trick. The other amendment that is needed is for the Project page to show LTD data so that we can see if there is a problem building up. Cheers ID: 20975 | Rating: 0 | rate: / ReplyQuote Dagorath Send message Joined: Jun 13 07 Posts: 385 ID: 2516 Message 20974 - Posted 25 Oct 2008 20:47:49 UTC - in response to Message 20973. The scheduler doesn\'t work well when you give too high of a resource share to projects that don\'t have work all the time, for example, LHC and SIMAP. The scheduler will let projects borrow CPU time from other projects and pay the debt back later. What has happened is that all your projects now owe LHC so much CPU time, the scheduler won\'t get work for them until LHC collects its debt by running some LHC tasks. Since you can\'t get any LHC tasks, LHC can\'t collect its debt. What you need to do is reset the LTDs (long term debts) to 0 then reduce LHC\'s share down to a number that LHC can meet. I can\'t tell you exactly what that number should be but I would say it should be less than the smallest share for projects that have steady work. In other words, LHC\'s share should be less than Einstein\'s share which is 40. One easy way to reset the LTDs is to use BoincDV. ID: 20974 | Rating: 0 | rate: / ReplyQuote Mike Gibson Send message Joined: Mar 5 08 Posts: 13 ID: 3437 Message 20973 - Posted 25 Oct 2008 19:29:47 UTC Last modified: 25 Oct 2008 19:45:14 UTC Hi I am running CPDN/Einstein/LHC/Rosetta/Seti/WCG using 6.2.19 with resource shares of 280/40/80/60/80/100 repectively on a dual-core machine. I have 2 CPDN WUs left, both with over 2000 hours remaining and one of which I have suspended, requiring only 12% & 13% of the allotted CPDN tme. Also 1 Seti Astropulse showing 53 hours remaining but I expect that to take about 90 instead needing 44% of Seti allotted time and 1 WCG(HPF) with only 13 mins left needing 0.2% of WCG allotted time. My cache is set for 10 days but I can\'t get any more units. LHC is out of work and Boinc is vainly trying to get 99,000 secs of work for that, I have set CPDN not to get any more WUs, but the other 4 projects seem to be a problem. I have tried doing a manual update on all 4 but the message each time says \'Requesting 0 seconds of work\'. Why won\'t BOINC Manager let me ask for work, please? Is it a bug? ID: 20973 | Rating: 0 | rate: / ReplyQuote |
|
10)
Message boards :
Number crunching :
Please note: this project rarely has work
(Message 19414)
Posted 14 Apr 2008 by Mike.Gibson Post: I suppose it is like their WORLD Series which is only competed for by US teams. There are 300M of them and none of them seem to be aware that are 8B others out here. They also don't know how to spell English on their side of the pond, either. |
©2024 CERN