Message boards :
Number crunching :
Can anyone stop this machine ?
Message board moderation
| Author | Message |
|---|---|
 Yeti Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,569,815 RAC: 13,099    |
I just checked my work and found some results, that where not yet validated. When I followed the Tasks, I found a machine that is trashing round about 5.000 Tasks. http://lhcathomeclassic.cern.ch/sixtrack/results.php?hostid=10298557 Please, can anyone stop this ?  Supporting BOINC, a great concept ! |
|
Send message Joined: 27 Sep 08 Posts: 807 Credit: 651,900,632 RAC: 291,407  |
You could try to PM the owner? |
|
Send message Joined: 9 Oct 10 Posts: 77 Credit: 3,671,357 RAC: 0 |
It's not the only one ... I've seen a few AMD CPUs as my wingmen that are trashing WUs too ... :( |
|
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Sorry about this; my fault as usual. We are trying to run boinc tests for the new sixtrack. I think they are all failing. We have to validate "manually". I have the fix and we shall try again. They will be very short so as not to waste CPU tiime. More real work coming soon. Eric. |
|
Send message Joined: 9 Oct 10 Posts: 77 Credit: 3,671,357 RAC: 0 |
Hi Eric, In this case we're not talking about SixTrackTest but Sixtrack production WUs ... And I double checked, it's also Skyhawk's machine linked by Yeti which is trashing work :( |
|
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Right; I checked again and indeed it looks like production.........the plot thickens. I will investigate more closely tomorrow when Danilo (the user) is back from vacation. Very strange. (In fact the "test" short jobs are returning results.) I need to have a look at "computation error" I think srderr just says too many exits. (Maybe you could see something in fort.6 which I don't get back.) Anyway thanks for your help and messages. Eric. |
|
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
I already had a look at a failure; I see: Stderr output The segment is already discarded and cannot be locked. (0x9d) - exit code 157 (0x9d)ues have been in the results directory...... forrtl: severe (157): Program Exception - access violation Image PC Routine Line Source sixtrack_windows_ 005A58E9 Unknown Unknown Unknown sixtrack_windows_ 005A3F83 Unknown Unknown Unknown sixtrack_windows_ 0059BFDB Unknown Unknown Unknown sixtrack_windows_ 0057BE85 Unknown Unknown Unknown sixtrack_windows_ 0057BA1B Unknown Unknown Unknown sixtrack_windows_ 00599D50 Unknown Unknown Unknown ]]> Never seen this before! Perhaps a bug in SixTrack??? I have an open mind. When I get to work I'll look at his case with Danilo. However I also see my colleagues have been active in the results directory. There are 3344 results (look OK) to be downloaded . Have you updated client recently? More news soonest. Thanks. Eric. |
 Send message Joined: 12 Mar 12 Posts: 128 Credit: 20,013,377 RAC: 0 |
Eric Most of my processing computers have no active tasks or just finalising 1-2 WU received 31 Aug or 1 Sep. What do we expect now? |
|
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Just finishing the installation of a new SixTrack. More work coming. Just want to try and understand these failures as well. Eric. |
|
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
I tested a "failing" WU at CERN Linux and Windows. Running OK. We shall see. Eric. |
©2024 CERN