| Posts by Ananas |
|
1)
Message boards :
Number crunching :
Not enough disk space allocated????
(Message 5790)
Posted 1101 days ago by Ananas The server side log (if it is activated) does tell which of the 3 HD space values caused the limitation, too bad that the server does not send this information to the BOINC client. Example (randomly choosen from one Einstein log) : 2007-09-05 01:35:03.1064 [PID=4639 ] [debug ] [HOST#496460] MSG(high) (there was work but you don\'t have enough disk space allocated) I\'m using quite oldish BOINC clients though, maybe later clients do tell. |
|
2)
Message boards :
Number crunching :
Why No Credit??
(Message 5427)
Posted 1142 days ago by Ananas Pending credits are those that your PC already did claim, so you don\'t have to do anything once they sit there in pending state. Usually, they get granted when \"minimum quorum\" results (that is 3 in this project) of the same workunit have been returned and those 3 are identical or at least close enough, so the validator (a server side program) is happy with them. As long as uFluids is still Beta, the project admin goes through the invalid results now and then and grants the credits anyway, even for invalid results, as the result difference might be not the users fault. You cannot buy anything for the granted credits (except for ice cream under certain circumstances *), they are just a measure for the work you did for science and can be used for team competitions and stuff like that. * http://hfs.washington.edu/boinc/ |
|
3)
Message boards :
Number crunching :
bridge4 WUs
(Message 5013)
Posted 1205 days ago by Ananas Just finished one, 42566 sec (nearly 12 hours) on a 2xMP2600+ It worked on the 60% part for quite a few hours, the other steps went faster. That Atlon MP has only a small 2nd level cache (256k), which might slow things down. |
|
4)
Message boards :
Number crunching :
Disk space useage of uFlinds....
(Message 5010)
Posted 1205 days ago by Ananas Maybe an old crashed WU? Some WUs increased the result file quite much. If such a result got lost in client_state.xml, the leftover files might still sit there. The project executable isn't that big, the one I have is still just 2MB. If your cache for this project is empty anyway, you can get rid of that stuff with the "reset project" button, it will redownload the 2MB file when you request the next workunits. |
|
5)
Message boards :
Number crunching :
Runtime stats
(Message 5005)
Posted 1206 days ago by Ananas Maybe it would be possible to have a stats web page showing count(*), avg(), min() and max() runtime for all valid(!) results, group by the significant part of the WU name (name part prefix up to the 1st underscore, i.e. SUBSTRING_INDEX(workunit.name, 1, '_') ) It would be helpful for a decision wether the WU got stuck or might still be doing something, when fraction_done hasn't changed for hours. |
|
6)
Message boards :
Number crunching :
Maximum CPU time exceeded
(Message 4893)
Posted 1218 days ago by Ananas Not done yet I guess? At least not for the WUs that already have been generated. Do you have a naming pattern that allows us to selectively abort those drop5 WUs that are still with the old max. CPU time? |
|
7)
Message boards :
Number crunching :
-161 file_xfer_error
(Message 4839)
Posted 1222 days ago by Ananas Connect and -161 are not related on 5.2.13, the WU mentioned above survived. |
|
8)
Message boards :
uFluids@home Science :
Goal of this project = Owner's Master's Degree?
(Message 4838)
Posted 1222 days ago by Ananas Maybe it could be used to design optimized heatpipes for CPU coolers when the current work is done :-) |
|
9)
Message boards :
Number crunching :
-161 file_xfer_error
(Message 4837)
Posted 1222 days ago by Ananas I'm using CC 5.2.13, but still I get the impression, that there is a connection between -161 errors and update connects to the uFluids server. I reduced my cache, which leads to uploads without reporting, as it does not need fresh work so soon. I did one manual update on a different project to report a result, the next uFluids WU showed -161 Now I left the update button alone for several hours, no crashes since and it finished 5 results during this time. I will report the finished uFluids results now, watch the next result and report about its outcome later. If this next result will be a -161, it might be an API problem causing those errors. p.s.: I see a reference to http://www.ufluids.net/files/ufluids_icon.gif in client_state now, that it didn't have before. This file is nowhere in the BOINC directory. |
|
10)
Message boards :
Number crunching :
-161 file_xfer_error
(Message 4831)
Posted 1223 days ago by Ananas I have some -161 errors too This one is "neat" : http://www.ufluids.net/workunit.php?wuid=1632338 3 errors, 1 OK and one still in progress. This shows, that the "max # of error/total/success results" setting triggers a BOINC bug. After the second error results already had been returned, the WU has been delivered one more time, even though 2 error results should have been the end of that WU. |