| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Outsmart Ebola Together
Thread: OET WU taking Longer than usual progress. Stuck? (31May 2019) |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges: 
|
Greetings,
----------------------------------------I have an 8-core CPU and and running 7 of them doing OET WU. Lately , i do noticed that they take longer to run. But today... They start off with an estimated 5 1/2 hours to run. Three of them show 4 Hours:24 minutes remaining (Not changing) with:
4:07 elapsed 1:37 elapsed Three WU show declining remaining time - but less than 1 hour elapsed. And one wu showing 5:33 elapsed and 22:15:38 remaining - but the remaining time is INCREASING. Anyone else seeing this? How long should I wait before aborting? Or wait until they self-abort? Thanks!! Jay  |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7581 Status: Recently Active Project Badges: 
|
I have seen this goofy behavior a few times, but I think it is some kind of internal problem with the progress indicator. I have left them go and they seem to finish in about the right time. I had one which indicated 20% and 24+ hours to run after about 4 hours, but it seemed to finish in the usual range of 6 to 8 hours. I usually check my machines about once a day, so if some unit started racking up over 24 hours of run time with no progress, I would eventually notice it, but so far that has not happened. If one did run for more than 24 hours of run time, I would abort it.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges:
|
THANKS!!!
----------------------------------------As you say - part of progress indicator. I will be more patient. 6 of the WU took 6 or 7 hours. That's OK - just was anxious.. :-) |
||
|
|
AlexGV
Cruncher Russia Joined: Apr 6, 2018 Post Count: 16 Status: Offline Project Badges: 
|
I just faced the same issure. I run two notebooks with 4c/8t i7s and one of them shows significant slowdown all of a sudden. The estimated time just going up prolonging overall working time for each task. I still see the overall progress rising but it is way too slow compared to how it was before. The speed reduced almost twice. Tasks taking more than one day instead of 14 hours.
----------------------------------------The CPU not being working at full power as before while prcessing tasks, it barely goes higher than 40-50%. I can see that with monitoring software, fan speed and generated heat alltogether, also in the screensaver animation (molecule rotates slower indicating overall processing speed has been reduced). I tried to restart machine, pause work and resume it but it seems like it helps only for several minutes, after that the processing speed just goes down again. Both notebooks are on the wall plugs (not on the batteries, though they are plugged in) and no energy saving modes - just maxed performance settings 24/7. Is it only me or tasks are more heavy now? Which logs I possibly could show to find an issure? This is really frustrating.
i7-3632QM 4C/8T, 2.2GHz (2.9GHz Turbo), 35 W;
----------------------------------------i7-3630QM 4C/8T, 2.4 GHz (3.2GHz Turbo), 45 W. [Edit 4 times, last edit by AlexGV at Oct 25, 2018 8:50:05 AM] |
||
|
|
AlexGV
Cruncher Russia Joined: Apr 6, 2018 Post Count: 16 Status: Offline Project Badges:
|
I've noticed some slow down on the second notebook too but it is not that high compared to the first one, just few hours (19-23 hours against usual ~18 hours).
----------------------------------------
i7-3632QM 4C/8T, 2.2GHz (2.9GHz Turbo), 35 W;
----------------------------------------i7-3630QM 4C/8T, 2.4 GHz (3.2GHz Turbo), 45 W. [Edit 1 times, last edit by AlexGV at Oct 25, 2018 4:56:33 PM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7581 Status: Recently Active Project Badges:
|
You say you are monitoring the temperatures, but your slowdown has all the earmarks of overheating causing the the unit to throttle down due to thermal issues. This is a known problem with notebooks, especially as they get thinner and do not have good heat dissipation. Somehow the manufacturers just do not expect the users to run their laptops at 100% for hours on end.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
AlexGV
Cruncher Russia Joined: Apr 6, 2018 Post Count: 16 Status: Offline Project Badges:
|
You say you are monitoring the temperatures, but your slowdown has all the earmarks of overheating causing the the unit to throttle down due to thermal issues. This is a known problem with notebooks, especially as they get thinner and do not have good heat dissipation. Somehow the manufacturers just do not expect the users to run their laptops at 100% for hours on end. Cheers Hi, the problem is definetely not in overheating. Both notebooks ran on full workload non-stop without any problems for a long time. The processing power went down all of a sudden and it doesn't even close to how it would be like with a throttling (I know how it would be like). At full load the temperature is high but not critical for the CPUs I have so they don't throttle most of the time at all. One of possible reasons I thought about might be that something is wrong with the system power management due to the defective windows update (I run Windows 10 on a machine where the problem exists for 100%) and how BOINC reacts on that (false system trigger about working with the battery instead of the AC). There were numerous reports from notebooks owners about their batteries starting to work inappropriate - notifying they are missing or discharged in a second. I tried many things to locate the issure. Since my power settings are set for maximum performance and machines working with AC, I just pulled out the battery "hot" way to see if system will trigger anything to switch the modes correctly. Once I did that, calculations in BOINC client went for the maximul performance again, and made the fan to spin at the appropriate speed fo that generated heat (good sign of returned maximum workload). But problem then comes back as calculations drop down again and never go higher than around 50% of the CPU usage. Next, I plug the battery back in "hot" way again and voila - calculations are back to normal. But, not forever, sadly. After almost 24 hours it went down back again... I retried the battery trick and again, it helped. And I'm monitoring for how long it will work properly again. Apart of the corrupted system update managing power settings and the batteries in a wrong way, I thought about faulty battery itself or maybe some laptop issure, but this is unlikely since everything worked fine before and, as you can see, even working only on AC doesn't solve the issure completely. I might play with some more settings to see if this will lead me to anything useful. Might as well confirm the OS power management issure and it's impact on how BOINC client treats the calculations. My guess is that there is some software issure. For the second notebook, it seem like the calculations are back to normal and giving me appropriate ~ 18 hours of work per the task, as usual again.
i7-3632QM 4C/8T, 2.2GHz (2.9GHz Turbo), 35 W;
----------------------------------------i7-3630QM 4C/8T, 2.4 GHz (3.2GHz Turbo), 45 W. [Edit 4 times, last edit by AlexGV at Oct 26, 2018 6:38:26 AM] |
||
|
|
|