The Einstein BOINC plugin fails to honor the CPU usage contraints placed upon it, and periodically causes EVERYTHING, even mouse and keyboard events, to grind to a complete halt for up to ten seconds at a time. I know Einstein is at least complicit in creating this problem because: (a) BOINC service and Manager alone don't cause it, and (b) BOINC and SETI@Home plugin together don't cause it. Every time the problem has been occurring I've checked the BOINC Manager, and EVERY single time it's been the Einstein project that was active at that instant. Whether this event occurs at the point of transition to Einstein or in response to some other state of the computations, I cannot say; I am unable to monitor it that precisely. Further, the symptom only began when I started running the Einstein plugin during SETI@Home's recent outage.
I don't deny the possibility that some other aspect of my environment might also be contributory, but if so it is indeed merely contributory. I am reporting this issue, but I will not be providing further details as I'm removing the Einstein plugin shortly. This is a production machine and I cannot afford to use it to help beta test this plugin. I doubt that I'm the only participant who has experienced this behavior, so presumably you can find someone with a less critical system willing to let you use it as a test bed to troubleshoot the problem. I can no longer afford the disturbance.
Mark Craig
Copyright © 2024 Einstein@Home. All rights reserved.
Einstein plugin fails to honor CPU constraints
)
FYI, these other threads are likely describing the exact same symptom:
http://einsteinathome.org/node/190412
http://einsteinathome.org/node/190375
http://einsteinathome.org/node/190061
http://einsteinathome.org/node/189921
http://einsteinathome.org/node/188603
And who knows how many of the "crashes" and "lockups" other less patient people have encountered and reported here might actually be the same symptom, and that might have been reported differently had they waited a bit longer before hitting the Reset button?
Instead of dismissing people that report such things as crackpots or deprecating their observations, perhaps they should be taken at face value so that the problem gets proper attention?
Mark