GPU Folding Stall?

[Ensign.Fodder]

Guys and Gals,

After getting the SLI 460's folding together and finding the upper limit of the OC's I thought all was good, but I just found that one had "stalled", meaning it was still showing as working but the GPU usage was 0% and the log showed no progress for five hours. A simple shut down and restart for that GPU fixed the problem but I am concerned that my days of babying the two is not over with.

Any idea's on cause and ways to re-mediate?

Thanks!

 

[Filter]

sounds like the classic 'driver crash"

up the voltage.

 

[Ensign.Fodder]

sounds like the classic 'driver crash"

Would that kill one but not both?

up the voltage.

Running at 987mV currently

 

[jebo_4jc]

Would that kill one but not both?





Running at 987mV currently

I agree that giving it some more voltage would probably be my next course of action. I think I have my 460s at approx 1050mv

 

[Filter]

Would that kill one but not both?


Running at 987mV currently

not necessarily on the driver crash. run 1025mv i run that at 851mhz on my 4 460's in my machine

r-type reccomened that voltage and fixed all my instability issues.

 

[sirmonkey1985]

sounds more like the issue others were seeing with the SMP client where it would quit writing to the log file even though the WU was still running fine. which was caused by memory issues.

 

[BassTek]

I have the same problem with my 460s. SLI enabled, after a reboot both will run fine for at least one WU then one will eventually stop at 0%. Not really worth my time to fix it as googling the error didn't turn up much, so I only fold on one now.

 

[Ensign.Fodder]

sounds more like the issue others were seeing with the SMP client where it would quit writing to the log file even though the WU was still running fine. which was caused by memory issues.

Using MSI Afterburner that GPU is @ 0% so I do not think the WU was still being processed.

Thanks!

 

[Filter]

Even in SLI i still need to give them more voltage then just on one card to achieve the same o/c lvl.


im 99% sure more voltage will fix your problem.

 

[Ensign.Fodder]

Trying 1000mV now.

 

[Filter]

Trying 1000mV now.

any results?

 

[Ensign.Fodder]

any results?

Sure UPS started squealing!

Backed down to 987mV.

Stable so far

 

[Ensign.Fodder]

Dropped voltage back down one notch and saw the driver crash.GPU 1 recover but GPU 0 never did. Back to 987mV but again the UPS squeals when external HDD spins up so sounds like time to either get the screwdriver/pliers out or another UPS.

 

[BassTek]

Hmmm, I'll try upping mine as well. They are factory oc'd so maybe they are on the edge of stability and SLI folding puts them over the edge.

Edit: My problem still remains. I can start either/or but if I start both the second one fails with:

[13:59:10] Working on PEPTIDE (1-42)
[13:59:10] Client config found, loading data.
[13:59:14] CoreStatus = 63 (99)
[13:59:14] + Error starting Folding@home core.

And then loops as it tries to get new WUs.

 

[Filter]

Hmmm, I'll try upping mine as well. They are factory oc'd so maybe they are on the edge of stability and SLI folding puts them over the edge.

Edit: My problem still remains. I can start either/or but if I start both the second one fails with:

[13:59:10] Working on PEPTIDE (1-42)
[13:59:10] Client config found, loading data.
[13:59:14] CoreStatus = 63 (99)
[13:59:14] + Error starting Folding@home core.

And then loops as it tries to get new WUs.

when runinng sli you wont get the same o/c as you would on a single card with the same voltage.

i run 1025mv to my 4 460's in hte same box and has fixed all my stability porblems. also the stock voltage for each 460 is different. 2 our from pny and 2 are from galaxy and all have different stock voltages.

i think having all 4 the same also helped fixed stability issues also.

all 4 were at 851mhz.