Overclock.net › Forums › Overclockers Care › Overclock.net Folding@Home Team › GTX 460 has failed 5 WU's in a row
New Posts  All Forums:Forum Nav:

GTX 460 has failed 5 WU's in a row

post #1 of 19
Thread Starter 
This has been happening recently, and now I can't get any PPD for the work my GTX 460 is doing. redface.gif

I am using -advmethods.

Edit: The card is stable, and even if I up the voltage it continues to happen.
Edited by 1keith1 - 2/6/12 at 8:35pm
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
post #2 of 19
Where is it failing the work unit at? Right when it loads, or after it has been on after awhile? Also is it throwing any error codes?
My System
(15 items)
 
The Expermint
(7 items)
 
 
CPUMotherboardGraphicsRAM
2 X 4386 Asus KCMA-D8  560 SE Wintec 
CoolingCoolingOSPower
stock stock Win 7 PC Power and Cooling 750W 
CPUMotherboardGraphicsRAM
intel 5820K X99-UD4 ati 5750 corcair  
Hard DriveOSPower
ScanDisk SSD win 8.1 CORSAIR Builder Series CX600  
  hide details  
Reply
My System
(15 items)
 
The Expermint
(7 items)
 
 
CPUMotherboardGraphicsRAM
2 X 4386 Asus KCMA-D8  560 SE Wintec 
CoolingCoolingOSPower
stock stock Win 7 PC Power and Cooling 750W 
CPUMotherboardGraphicsRAM
intel 5820K X99-UD4 ati 5750 corcair  
Hard DriveOSPower
ScanDisk SSD win 8.1 CORSAIR Builder Series CX600  
  hide details  
Reply
post #3 of 19
Thread Starter 
No errors, sometimes it gets just about to finish before failing, other times it happens very quick into the WU. I have disable -advmethods because I think that is causing it, but I would like to able to use it to up my PPD.
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
post #4 of 19
There are two things I can think of,

1, heat, how long has it been since you have cleaned your 460 out of dust?

2, power, how old is the power supply? Or has the tower taken a jolt (by cat, dog, kids, or vaccume cleaner) in the past couple days?


Lets try the first one, take out the card, and clean the card and look for anything bad, this also solves the reseating of all the electrical connectors.
My System
(15 items)
 
The Expermint
(7 items)
 
 
CPUMotherboardGraphicsRAM
2 X 4386 Asus KCMA-D8  560 SE Wintec 
CoolingCoolingOSPower
stock stock Win 7 PC Power and Cooling 750W 
CPUMotherboardGraphicsRAM
intel 5820K X99-UD4 ati 5750 corcair  
Hard DriveOSPower
ScanDisk SSD win 8.1 CORSAIR Builder Series CX600  
  hide details  
Reply
My System
(15 items)
 
The Expermint
(7 items)
 
 
CPUMotherboardGraphicsRAM
2 X 4386 Asus KCMA-D8  560 SE Wintec 
CoolingCoolingOSPower
stock stock Win 7 PC Power and Cooling 750W 
CPUMotherboardGraphicsRAM
intel 5820K X99-UD4 ati 5750 corcair  
Hard DriveOSPower
ScanDisk SSD win 8.1 CORSAIR Builder Series CX600  
  hide details  
Reply
post #5 of 19
Thread Starter 
But its completely stable, it doesn't go above 80C either, its a fermi it can take on 95C easy.
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
post #6 of 19
I would say its the overclock, turn it back 10mhz see if it happens again
Main rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
Icore7 2600k P8P67LE ASUS EVGA 660ti 8 gb Corsair Veng 
Hard DriveOSMonitorKeyboard
Silicon Power ssd 120 gb Windows 10 Acer P237HL microshaft ECR 4000 
PowerCaseMouse
ULTRA LSP 650 thermailtake element t Razor Naga  
  hide details  
Reply
Main rig
(13 items)
 
  
CPUMotherboardGraphicsRAM
Icore7 2600k P8P67LE ASUS EVGA 660ti 8 gb Corsair Veng 
Hard DriveOSMonitorKeyboard
Silicon Power ssd 120 gb Windows 10 Acer P237HL microshaft ECR 4000 
PowerCaseMouse
ULTRA LSP 650 thermailtake element t Razor Naga  
  hide details  
Reply
post #7 of 19
Quote:
Originally Posted by 1keith1 View Post

This has been happening recently, and now I can't get any PPD for the work my GTX 460 is doing. redface.gif

I am using -advmethods.

Edit: The card is stable, and even if I up the voltage it continues to happen.

The 7620, 7621, 7622, and soon to arrive 7623 WUs are by far the most stressful thing that you can throw at a Fermi card. They will find instability in your OC that you will not find with a normal WU, with any game, or with any benchmark.

 

Your overclock is NOT STABLE if you are getting NANs on your Fermi card.

Quote:
Originally Posted by 1keith1 View Post

No errors, sometimes it gets just about to finish before failing, other times it happens very quick into the WU. I have disable -advmethods because I think that is causing it, but I would like to able to use it to up my PPD.

Your overclock is causing it. The -advmethods flag is simply enabling an advanced WU that is more stressful on your hardware than a standard WU and uncovering an instability that is not present when the card is not being run at maximum potential.

Quote:
Originally Posted by 1keith1 View Post

But its completely stable, it doesn't go above 80C either, its a fermi it can take on 95C easy.

It's not completely stable, even if you're at 75C. I fold 24/7 on my GTX 580 and my "Gaming stable" OC of 950 and 962 MHz will fail within 2 minutes of starting an -advmethods WU. I failed a WU after 9 successful WUs at 925 MHz and then I failed a WU after 30 successful WUs at 918 Mhz.

 

I'm now down at 904.5 MHz and not failing WUs and my voltage is quite high, almost as high as it needed to be for "gaming stable" at 962 MHz...

 

My temperatures max out at 76C in 24/7 folding situations.

 

EDIT: Just because it's a Fermi and can handle 95C doesn't mean it's going to fold at 95C nicely. I am only stable where I'm at because of my high fan speed profile. If I allow the card to heat up to 90C, the fans kick in at 100%, cool it down to 85C, where they throttle down, and then it heats back up to 90-95C where the fans kick in again. It will cycle like this constantly and it will have NANs.

 

Yet, by ramping up my fan profile and keeping it at 76C, it's stable at the same voltages and frequencies that I had without the custom fan profile.

Quote:
Originally Posted by behappy View Post

I would say its the overclock, turn it back 10mhz see if it happens again


^^^^

 

This.

 

It's the overclock. Turn it back 25 MHz without reducing your voltages and watch for another NAN. You might even be able to run 30 WUs in a row and have the 31st one fail. That's still unstable, it's just "more stable" than it currently is. Ideally you want to be able to fold 24/7 for a few months without getting NANs.

 

If you're serious about folding 24/7 on the big point -advmethods WUs with your Fermi card, then you're going to have to make very large reductions in core clocks and leave your voltages high, then make very small reductions in voltage until you get a NAN, then increase the voltage up a notch to where you get no NANs.

 

It might take you three months to find a truly stable spot and hopefully you can complete at least 10 to 20 WUs for every one that you fail.


Edited by shad0wfax - 2/6/12 at 9:49pm
B2
(8 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-5930K Haswell-E 6-Core 3.5GHz LGA... EVGA X99 Classified (151-HE-E999-KR) LGA 2011-v... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... 
RAMHard DriveCoolingCase
G.SKILL Ripjaws 4 series 16GB (4 x 4GB) 288-Pin... SAMSUNG 850 EVO MZ-75E1T0B/AM 2.5" 1TB SATA III... CORSAIR Hydro Series H100i Extreme Performance ... Phanteks Enthoo Luxe Series PH-ES614L_BK Black ... 
  hide details  
Reply
B2
(8 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-5930K Haswell-E 6-Core 3.5GHz LGA... EVGA X99 Classified (151-HE-E999-KR) LGA 2011-v... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... 
RAMHard DriveCoolingCase
G.SKILL Ripjaws 4 series 16GB (4 x 4GB) 288-Pin... SAMSUNG 850 EVO MZ-75E1T0B/AM 2.5" 1TB SATA III... CORSAIR Hydro Series H100i Extreme Performance ... Phanteks Enthoo Luxe Series PH-ES614L_BK Black ... 
  hide details  
Reply
post #8 of 19
Thread Starter 
Okay, I've turned my OC down to 875Mhz while keeping it at 1.137V. I ruled out stability because 890Mhz @ 1.137V has appeared to be stable from all the test I've thrown at it so far. I even upped it to 1.15V and it still failed, hopefully it will be stable now.

I hate my GTX 460, its not stable at 890Mhz with even 1.15V, these cards by MSI have the worst binning. mad.gif

I might run the full 1.21V through it before the warranty is up just to kill it and receieve one of the newer ones that OC's better, then I can get decent PPD at a decent voltage.
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
post #9 of 19
Quote:
Originally Posted by 1keith1 View Post

Okay, I've turned my OC down to 875Mhz while keeping it at 1.137V. I ruled out stability because 890Mhz @ 1.137V has appeared to be stable from all the test I've thrown at it so far. I even upped it to 1.15V and it still failed, hopefully it will be stable now.

I hate my GTX 460, its not stable at 890Mhz with even 1.15V, these cards by MSI have the worst binning. mad.gif

I might run the full 1.21V through it before the warranty is up just to kill it and receieve one of the newer ones that OC's better, then I can get decent PPD at a decent voltage.


That's still a respectable clock speed. Those cards are also not "binned" by hand. They are simply tested for QC compliance to MSI standards. MSI (and evga) both have some of the best QC standards in the industry for their Fermi cards.

 

The only cards that are truly binned, are the MSI Lightning Extremes and the EVGA Ultra Classifieds. Those cards are literally hand-picked cores (the VRAM is not binned) and those boards are over-engineered with superb VRM circuitry. (In the case of evga, this statement has been verified by the evga product manager who posts on these forums.)

 

Also, if you've overclocked the memory on it, that could be the cause of it. The memory on our cards is much more sensitive to OC than the cores are, so you can try turning the memory down some.

 

875MHZ @ 1.137 V sounds very reasonable for a Fermi on -advmethods, even though it's a crazy voltage for "gaming stable" at that frequency.

 

My GTX 580 is at 904.5 MHz @ 1.138V on -advmethods and our temperatures are similar as well.

 

-advmethods is just hellishly tough on our cards.

 

 

EDIT: If you're curious, what's happening in reality, is that you were unstable in your gaming benchmarks and you were getting artifacts. You probably just never noticed them, because they may have been as small as a single pixel and been on the screen for only a single frame. Your naked eye is not going to pick out one pixel out of millions on your monitor, especially if it flashes past in less than half a second during a benchmark. It's only when you get into extreme clocks that the artifacts become visibly obvious to the naked eye and gamers say "unstable here."

 

Now, Furmark and evga OC Scanner and OCCT (which uses Furmark) have "artifact detection" software capabilities that will detect artifacts that you and I can't see easily with the naked eye. The problem is that the Nvidia drivers throttle the power on our cards if we run Furmark, and so we never get to actually thrash on the card.

 

That leaves folding -advmethods as the only reliable way to detect minute instabilities that other benchmarks can't detect (or that our naked eyes can't detect).

 

In reality, you and I were unstable all along and we just didn't know it. Most gamers are "unstable" too, at their high OCs, but they can't see it and so they aren't aware of it. Unless it crashes the drivers or throws a giant sunburst on their screen or covers the textures in pink triangles or something, no one is going to notice it, and so they say "I'm stable at 962 MHz and your card just sucks."

 

Don't worry so much about it. Look at it this way, folding got your card overclocked to a truly stable point and you can have the confidence that your gaming is going to be at 100% quality with not even a single pixel of error. :)


Edited by shad0wfax - 2/6/12 at 10:05pm
B2
(8 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-5930K Haswell-E 6-Core 3.5GHz LGA... EVGA X99 Classified (151-HE-E999-KR) LGA 2011-v... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... 
RAMHard DriveCoolingCase
G.SKILL Ripjaws 4 series 16GB (4 x 4GB) 288-Pin... SAMSUNG 850 EVO MZ-75E1T0B/AM 2.5" 1TB SATA III... CORSAIR Hydro Series H100i Extreme Performance ... Phanteks Enthoo Luxe Series PH-ES614L_BK Black ... 
  hide details  
Reply
B2
(8 items)
 
  
CPUMotherboardGraphicsGraphics
Intel Core i7-5930K Haswell-E 6-Core 3.5GHz LGA... EVGA X99 Classified (151-HE-E999-KR) LGA 2011-v... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... EVGA 04G-P4-3975-KR GeForce GTX 970 SSC ACX 2.0... 
RAMHard DriveCoolingCase
G.SKILL Ripjaws 4 series 16GB (4 x 4GB) 288-Pin... SAMSUNG 850 EVO MZ-75E1T0B/AM 2.5" 1TB SATA III... CORSAIR Hydro Series H100i Extreme Performance ... Phanteks Enthoo Luxe Series PH-ES614L_BK Black ... 
  hide details  
Reply
post #10 of 19
Thread Starter 
Yes, but that kind of voltage was what I needed for it to appear stable in games, I understand that it was probably artifacts I didn't see, I try to test with OCCT for that purpose.

I think the memory is stable, I'll see how it folds now. I just hope I don't get set back as much as I did by all the failures it had today, I'm trying to get in the 500,000 points per month slot for prizes. tongue.gif
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
Only 9" Tall
(8 items)
 
  
CPUMotherboardGraphicsRAM
i5 4570 3.6Ghz Turbo Asrock Z87E-ITX Powercolor 7750 Samsung 
Hard DriveCoolingOSCase
Mushkin Chronos Scythe Big Shuriken v2 Windows 8.1 64bit Lian Li PC-Q02B 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Overclock.net Folding@Home Team
Overclock.net › Forums › Overclockers Care › Overclock.net Folding@Home Team › GTX 460 has failed 5 WU's in a row