SMP clients hanging?

K

KMac

[H]ard|DCer of the Month - June 2012
Joined
Dec 30, 2002
Messages
549
In the last week I have had several SMP clients hang, still at 100% CPU usage. The hang has occurred on four separate systems, all Win7 64 on UPS, all dust free, both regular SMP and bigadv. Only two of these systems are overclocked, and only one (stock clock) has a GPU client as well. The GPU client continued when the SMP was hung.
Because I thought it was a fluke rather than a trend I did not save the logs, but at least two of them mentioned something to the effect of 'shutdown failed, please use task manager to kill process'.
Is this becoming more common or does it indicate hardware or other issues on my end?
 
KMac said:
In the last week I have had several SMP clients hang, still at 100% CPU usage. The hang has occurred on four separate systems, all Win7 64 on UPS, all dust free, both regular SMP and bigadv. Only two of these systems are overclocked, and only one (stock clock) has a GPU client as well. The GPU client continued when the SMP was hung.
Because I thought it was a fluke rather than a trend I did not save the logs, but at least two of them mentioned something to the effect of 'shutdown failed, please use task manager to kill process'.
Is this becoming more common or does it indicate hardware or other issues on my end?
Click to expand...

I've got 2 smp clients and have no issues.

I would suggest getting some more complete logs - especially the core status error code if it generates one. If you just get a windows dialog box you should still be able to save the WU, using switch to the SMP command box and ctrl-C it in the normal way.
Also may be worth running AV/Anti spam just in case you collected something nasty recently from the hacking attacks
 
KMac said:
In the last week I have had several SMP clients hang, still at 100% CPU usage. The hang has occurred on four separate systems, all Win7 64 on UPS, all dust free, both regular SMP and bigadv. Only two of these systems are overclocked, and only one (stock clock) has a GPU client as well. The GPU client continued when the SMP was hung.
Because I thought it was a fluke rather than a trend I did not save the logs, but at least two of them mentioned something to the effect of 'shutdown failed, please use task manager to kill process'.
Is this becoming more common or does it indicate hardware or other issues on my end?
Click to expand...

I had two of these happen to me last week on my rig. It stayed at 36% for over 12 hrs. I ended up cancelling the project and attempted to grab a new project and it hung again. I ended up not starting another SMP client until 24 hrs have passes and now the current one is working with no issues.

I've just started this SMP client on my I7 950 setup, so I was thinking it was a problem with my rig. Unfortunately I didn't take note to which project was causing me an issue.
 
Update... The most recent hang was unrelated. Another process was using 100% so FAH had no resources.
A/V scans are clear and there were no Windows messages.
Here is one of the problem logs (too long to embed).
 
Last edited:
id delete the client and set it up again, see if that helps. also make sure it has admin access.
 
I had this hapen last night on my win 7 x64 955be system it sat at 63% for 7 hours still uesing 100 cpu
 
sirmonkey1985 said:
id delete the client and set it up again, see if that helps. also make sure it has admin access.
Click to expand...

It is running fine now, and if it was just a bad unit so be it. I just got concerned after something happened on a few different systems.
 
KMac said:
Update... The most recent hang was unrelated. Another process was using 100% so FAH had no resources.
A/V scans are clear and there were no Windows messages.
Here is one of the problem logs (too long to embed).
Click to expand...

Thats one thing i like about GPU tracker, you can stop the client easilly, and see which program is hogging one of the cores because an instance half crashed or something (im looking at you winamp and firefox) then kill whatever is using the CPU core and restart it.

If only one core is being used 100% by something else then all the other ones will stall while they wait for that one to finish the fragment it is working on.
 
argg just got on my pc and its been stuck at 69% all dam day the unit is 6702 (Run 6, Clone 48, Gen 91) send time on this unit it just got stuck i never had this problem before
 
windows update maybe people? sometimes a question box can hold everything up.
 
There is something going on with the log files in SMP. Same thing happened to me today. The log stopped at 56% but when I checked the unitinfo.txt file it reported 60%. The work was being done but the log file isn't being updated. I stopped unit and then started it again and the log finally updated. Check the unitinfo.txt file before you delete anything, it may be running fine just not being updated.
 
ok my dam unit is stuck at 69% been like that for hours even restated pc how can i fix it?
 
thax said:
ok my dam unit is stuck at 69% been like that for hours even restated pc how can i fix it?
Click to expand...

Rebooting does not help. Neither does 'Stop CPU' in GPU Tracker v2.
You need to kill the FAH.exe process using the Windows task manager (I rebooted too) and then restart the client.
You may need to do this repeatedly until you finish the WU if the client hangs again. After my WU finally completed there were no further problems on that machine.
 
Last edited:
If it happened on different computers, i guess it must have been a bad batch of work.
 
You must log in or register to reply here.
Back
Top