Message boards :
Number crunching :
Completed, marked as invalid
Message board moderation
Author | Message |
---|---|
Send message Joined: 16 Aug 05 Posts: 5 Credit: 2,795,425 RAC: 0 |
From the last batch of WU's, I have about 90 that were 'awarded' with : Completed, marked as invalid. If you look at for example task 47575574, the reports are: Name w2_jobhllhc10_inj_400_w2__11__s__62.28_60.31__8_10__5__82.5_1_sixvf_boinc8503_1 Workunit 22448500 Created 24 Oct 2014, 16:42:28 UTC Sent 24 Oct 2014, 22:55:22 UTC Received 25 Oct 2014, 7:46:47 UTC Server state Over Outcome Success Client state Done Exit status 0 (0x0) Computer ID 10283316 Report deadline 1 Nov 2014, 14:27:36 UTC Run time 4,695.46 CPU time 4,454.09 Validate state Invalid Credit 0.00 Application version SixTrack v451.07 (pni) and Stderr output looks like this: <core_client_version>7.2.42</core_client_version> <![CDATA[ <stderr_txt> 09:26:33 (5548): called boinc_finish </stderr_txt> ]]> My wingmen show the same output, but their results are valid (thus the Validate state). Maybe somebody can shine a light on this. Thanks, Ruud |
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Sorry Ruud; just spotted your message. I shall try and figure out what is going on. Eric. |
Send message Joined: 12 Jul 11 Posts: 857 Credit: 1,619,050 RAC: 0 |
Right, I checked logs and indeed the results is invalid. Cryptic message is 2014-10-25 11:27:31.9695 [debug] [RESULT#47575573 47575574] WU#0 HOST[10333141 10283316] differ: 11 1.0314171966464152e+000 8.6509219752258399e+000 p=0 V: 40517.000000000000e+000 40517.000000000000e+000 This is a real numeric difference; two other Hosts gave the same hopefully correct result. The 11 refers to word 11 in the fort.10 results summary and is the "maximum slope in phase space" which happens to be the most sensitive to any small numeric difference. If this is happening a lot I shall need to check your Host, as I do NOT expect numeric diffs. A one off can just be a memory error or similar glitch. Eric. |
©2024 CERN