Message boards :
Number crunching :
Erroneous validations or incorrect CPU time? Part 2
Message board moderation
Author | Message |
---|---|
Send message Joined: 23 Feb 15 Posts: 2 Credit: 53,124,056 RAC: 0 |
I have completed 44 tasks so far. 41 have validated. Computer I'm using hasn't processed any Universe@Home tasks before so I checked the wingmen on all those validated tasks to see if it was performing OK. I discovered something about those wingmen computers that seems strange to me: All are Anonymous Computers. All have highly variable Runtime, sometimes very low. All have no CPU time listed. All are running the same version of Linux: Linux 3.2.0-56-virtual. All use BOINC version 7.5.0. All have the following in Stderr output "Can't set up shared mem: -1. Will run in standalone mode.", sometimes repeated multiple times. Something doesn't seem right here. Similar thing was mentioned in a forum post last October but there was no answer given then. If anyone could explain this I would appreciate it. Seems suspicious to me and I prefer using computer resources and electricity on projects I have confidence in. Thank you. |
Send message Joined: 4 Feb 15 Posts: 847 Credit: 144,180,465 RAC: 0 |
Thats logs shows, that BOINC client is runned on virtual machine. Some of installations of Linux have to low shmem parameter set in system and this is why BOINC starts tasks in standalone mode, but all that results are valid. Also, the very long time on is probably because that machines are started somewhere on server and this server is sometimes under high pressure done by other tasks (one of my computers is on similar environment). Krzysztof 'krzyszp' Piszczek Member of Radioactive@Home team My Patreon profile Universe@Home on YT |
Send message Joined: 23 Feb 15 Posts: 2 Credit: 53,124,056 RAC: 0 |
Thank you for your prompt reply. I was more concerned with very short reported runtimes rather than long runtimes. Having thought about it a bit more the tasks showing very low runtimes all have multiple instances of "Can't set up shared mem: -1. Will run in standalone mode." in Stderr output. Which must mean they have been stopped and restarted a number of times. Thanks again for reassuring me that all results are valid. |
Send message Joined: 11 Mar 15 Posts: 37 Credit: 271,242,973 RAC: 0 |
I've been wondering the same same thing. If I go to any of my PC's & click on a random BHspin v2 valid work unit and look at who the wingman is I have a 98% chance that the wingman will be a anonymous host running Linux 3.2.0-56-virtual with either a i3, i5 or i7 and always using just one core in a VM and usually has an extremely fast run time.(try it yourself) Is this possible? well anythings possible, but what would be the chances? |
Send message Joined: 4 Feb 15 Posts: 847 Credit: 144,180,465 RAC: 0 |
It's not really short execution time, it's just BOINC didn't record it properly In other thread some time ago one of those anonymous users asks me himself to check his results. Krzysztof 'krzyszp' Piszczek Member of Radioactive@Home team My Patreon profile Universe@Home on YT |