|
41)
Message boards :
News :
Badges description change
(Message 538)
Posted 14 Sep 2015 by  Ananas
Post: Badges on the team pages show the same 'kk' error Currently badges in the team member lists are damaged (duplicate "img src") : <img src="<img src="img/badges/mars.jpg" title="Mars (500k)"></img>"> |
|
42)
Message boards :
Number crunching :
Not getting much work
(Message 537)
Posted 13 Sep 2015 by Ananas
Post: When I do rarely get A download it shows up as download failed! Is there something wrong? I think I have 12 wu's and all show Error while downloading. It says "wrong size" for universe-xray_sources_v3_2_windows_intelx86.exe in your workunits, so something must have gone wrong when your BOINC client downloaded the file You could try to download it from here through your web browser : http://universeathome.pl/universe/download/ and then copy the downloaded file into the Universe project folder(s). Work supply is currently very low so you might get a "no tasks available" message at the moment. |
|
43)
Message boards :
Number crunching :
Project Temporarily Down
(Message 415)
Posted 16 Jul 2015 by Ananas
Post: Something is still wrong. Tasks ready to send = 10000 but the sched_reply is <message priority="low">Project has no tasks available</message> Otoh. the server status page says under "Tasks by application" : unsent = 0 This somehow doesn't go together. (Wintel x64 box / 16 threads, core client 7.4.36) |
|
44)
Message boards :
News :
Badges description change
(Message 314)
Posted 14 May 2015 by Ananas
Post: Jupiter ALT text has one "k" too much : title="Jupiter (250kk)" |
|
45)
Message boards :
News :
Work Units cancelled
(Message 280)
Posted 22 Apr 2015 by Ananas
Post: It always creates _0, _1 and _4 files but never _2. _3 and _5 (same existing / missing numbers for all 16 "xray_13" results I had today). Files in the slot directory of a result at 97.5% : boinc_task_state.xml checkpoint.dat2 checkpoint.dat3 <= same content as checkpoint.dat2 data.dat error.dat <= empty error.dat2 <= empty error.dat3 <= empty gwMT.dat gwMT.dat2 gwMT.dat3 <= same content as gwMT.dat2 gwWIND.dat gwWIND.dat2 gwWIND.dat3 <= same content as gwWIND.dat2 init_data.xml job.xml param.in stderr.txt tmp5467.dat <= empty I saved those so if you need any of them for the analysis, I can put it somewhere for download. I wonder what purpose all the duplicate files might serve :-/ The boinc_lockfile was locked (couldn't copy it) but that never has any contents so it plays no role. (btw.: if I would test my software like this, it would probably not be easy for me to get a job) |
|
46)
Message boards :
News :
Badges description change
(Message 255)
Posted 2 Apr 2015 by Ananas
Post: Still not back to normal? April 1st is over. Conan's still says "Uranus". Should be fixed. sometimes |
|
47)
Message boards :
News :
Badges description change
(Message 252)
Posted 2 Apr 2015 by Ananas
Post: Still not back to normal? April 1st is over. According to Dr. Hubert J. Farnsworth Uranus has been renamed at some point. |
|
48)
Message boards :
News :
Work Units cancelled
(Message 244)
Posted 1 Apr 2015 by Ananas
Post: What is going here? ... Explanation in posting 226 (earlier in this thread). |
|
49)
Message boards :
Number crunching :
WU's not stopping
(Message 237)
Posted 31 Mar 2015 by Ananas
Post: Check the thread "Work Units cancelled" in the "News" section |
|
50)
Message boards :
News :
Work Units cancelled
(Message 234)
Posted 31 Mar 2015 by Ananas
Post: I have no idea.... BOINC's "suspend" doesn't work on your application (p.s.: on windows, not sure about *ix), suspended tasks keep running. So if you have 7 tasks running and the host goes into panic mode and starts 7 different tasks with an earlier deadline, you will end up with 14 active tasks (BOINC doesn't notice that and marks the paused ones as "suspended" in the GUI). The same problem exists at FiND@Home, still unsolved, so if you find a solution, maybe you could share it with them. Unfortunately they have not created their application properly so the executable doesn't contain the BOINC API version string (I supposed that it might be a version issue but could not verify it). p.s.: This behaviour might have a bad side effect, when a workunit ends while it is supposed to sleep, it reports an error "result file present too long" (the wording is from my memory, might be slightly different) as BOINC doesn't pick up the result immediately (it doesn't watch suspended applications) |
|
51)
Message boards :
News :
Work Units cancelled
(Message 229)
Posted 31 Mar 2015 by Ananas
Post: I'm not sure if this information helps : The _9_ series had checkpoint files that consisted of a single line containing only the character "0" (0x30, not 0x00) followed by CRLF after 15+ hours |
|
52)
Message boards :
Number crunching :
WU deadlines
(Message 216)
Posted 31 Mar 2015 by Ananas
Post: Be careful with your cache sizes, v2 v0.02 runs much longer than v2 v0.01. Currently I have some at 13.5 hours, not finished yet, the last %%% take forever. It used to be ~4.5 hours. I hope that RSC_FPOPS_BOUND is set high enough p.s.: actually even the last fractions of % take forever, those things are stuck at 99.91% to 99.93% for quite some time now. Still using the full CPU time for each thread, so they are at least not sleeping. And while I was typing, one jumped to 99.94%, so I hope they still do something ... something useful that is ;-) |
(English)
(Polski)