Did I just blow up my 980's ?

[spacejunk]

I had these in my system for about 2 and a half years but never overclocked them. Getting quite frustrated at poor performance in some poorly optimized games I got desperate and tried overclocking them a bit.

I was using an older version of Afterburner and went for a 192mhz OC on both core and memory. I set power at 110% and crammed in a few more MV's. Fire up the Witcher 3 and things ran smooth for a while, then it crashed with windows reporting a driver crash and said it had recovered.

Now the scary part. When I looked back at Afterburner both core and memory clock speeds were in excess of 32470mhz ... yes you read that right.

I went back into NVCP and my 2nd card was gone, no SLI options. Rebooted PC and SLI was back. Failing that I uninstalled all the old overclocking tools and started fresh.

This time with the latest Afterburner, GPU-Z, NVI I tried again. Set the core at + 200 and memory at +250 with +27mv (27mv being the max available) and 110% power. Fired up Heaven 4.0, let it loop a few times and benched with no issues. I even beat my old score by 300 points totaling 3444 at 1080p max settings and tessellation.

Still displeased with performance, I disabled SLI to test another game. That game would crash intermittently and I got the driver-crash again. When I looked back at Afterburner which I forgot about, the card that was showing the boost was obviously disabled. So I closed Afterburner. I haven't had the driver crash again, but I have had intermittent crashes and I wonder if I damaged both cards or one of them.

I am now testing another game to see if I can isolate the crashing to only one game. So far it's been playing for about 2 hours without issues, but it's Darksouls 1 which is not terribly demanding. Maybe I need to do a session of the Witcher again.

For some reason I was not able to locate the driver error information, the only events that match up are the D3d11.dll which I believe were the normal crashes and not the driver related ones. Here is crash event data:

Faulting application name: LordsOfTheFallen.exe, version: 0.0.0.0, time stamp: 0x54f449cc
Faulting module name: d3d11.dll, version: 6.2.9200.16570, time stamp: 0x5153b56b
Exception code: 0xc0000005
Fault offset: 0x00000000000d6f09
Faulting process id: 0x15a4
Faulting application start time: 0x01d258547e2d3a5a
Faulting application path: H:\Games\Steam\steamapps\common\Lords Of The Fallen\bin\LordsOfTheFallen.exe
Faulting module path: C:\Windows\system32\d3d11.dll
Report Id: b514762c-c448-11e6-bec4-902b343d2ae3


This game is notorious for stability issues and I had not played it to any great length of time before the overlclocking ordeal. I'm hoping it's just coincidence, but the fact that Aferburner had my cards overclocking in the tens of thousands has me worried.

 

[RCFProd]

They're still working so they certainly didn't blow up.

 

[spacejunk]

Lol thanks very helpful.

 

[vrumor]

It is helpful. If they still work, they arent broken. Why be sarcastic when someone is trying to help you.

 

[Shaun o]

Like said you didn`t blow them up.

But like with most types of overclocking you do on graphics cards or cpu`s as a matter of fact.
What happens is you exert more strain on the integrated data path ways of either the cpu or Gpu you are overclocking.
That leads to the circuit pathways degrading due to extra voltage applied and higher temperatures the chips have to run at.

As they degrade they loose there ability to carry a signal at a nominal level, or a reduction in the amount of voltage the path can carry.

The result is you experience more instability when trying to run the cards at there factory default clock speeds and voltages.

Often to keep them stable enough you have to set the card in relation to the Gpu clock speed to a lower value than what was the factory default in speed and voltage.

It`s called electromagnetic migration of circuit pathways spacejunk
To gain stability of the Gpu again.

http://www.tandfonline.com/doi/abs/10.1080/10408439708241261

 

[tatsu99]

modern graphics cards are very resilient against abusive overclocking, your cards are fine. run furmark/kombuster and do a stability test.

 

[junkeymonkey]

still when you overclock you maybe hitting the power /voltage limit a bad / failed OC will crash and can throw the card in a fail safe mode

all that slider does is bump up the % of the target voltage so say when your card gets to 94% of the target voltage it may downclock but with the slider to 110% it may not until you hit with in 98%

like that gpu-z [see all that orange in what should be all blue field ? ]

http://imgur.com/YDTq3pV

my card running valley and gpu-z render test at the same time 3ed loop

http://imgur.com/bsLt1Ci

now not that I recommend this my card stock voltage was 1.193v I increased that to 1.230v plus the 110% with a oc I now got a cleaner voltage limit

any card can only do so much under any voltage and if you tax that and the program demands more then the card bios allows you crash

all you can hope for is with that game/ program your crashing with is start slowly and slightly bump up your clocks and all to the first crash you get and back off a bit from there in till your running stable


now for this ??

''Now the scary part. When I looked back at Afterburner both core and memory clock speeds were in excess of 32470mhz ... yes you read that right. ''

may of just been a glitch cause I never experienced that

thing is a cards overclock is a lot of luck of the draw some card do it well some may not at all and a oc that's stable in a lot of thing can still be bad under that one all you can do is hope for the best result or you may find you cant got anything much at all

 

[spacejunk]

It is helpful. If they still work, they arent broken. Why be sarcastic when someone is trying to help you.

Actually, the title thread was supposed to be "creative" or "entertaining" He was being facetious, his answer was blatantly anal and of no help whatsoever and he obliviously knew that. So I should be thankful ?

In fact don't answer that. I'm done talking about the correctness of a thread title.

Like said you didn`t blow them up.

But like with most types of overclocking you do on graphics cards or cpu`s as a matter of fact.
What happens is you exert more strain on the integrated data path ways of either the cpu or Gpu you are overclocking.
That leads to the circuit pathways degrading due to extra voltage applied and higher temperatures the chips have to run at.

As they degrade they loose there ability to carry a signal at a nominal level, or a reduction in the amount of voltage the path can carry.

The result is you experience more instability when trying to run the cards at there factory default clock speeds and voltages.

Often to keep them stable enough you have to set the card in relation to the Gpu clock speed to a lower value.
To gain stability of the Gpu again.

Thanks for the insight. I do know a little about silicone degradation caused by heat stress, but I am more worried about the immediate ramifications of Afterburner's little trick on me.

I just wanted to know if what happened there could be considered potentially lethal to a video card. Perhaps I should be less adventurous with my thread titles as some people don't appreciate it all lol.

 

[spacejunk]

still when you overclock you maybe hitting the power /voltage limit a bad / failed OC will crash and can throw the card in a fail safe mode

all that slider does is bump up the % of the target voltage so say when your card gets to 94% of the target voltage it may downclock but with the slider to 110% it may not until you hit with in 98%

like that gpu-z [see all that orange in what should be all blue field ? ]

http://imgur.com/YDTq3pV

my card running valley and gpu-z render test at the same time 3ed loop

http://imgur.com/bsLt1Ci

now not that I recommend this my card stock voltage was 1.193v I increased that to 1.230v plus the 110% with a oc I now got a cleaner voltage limit

any card can only do so much under any voltage and if you tax that and the program demands more then the card bios allows you crash

all you can hope for is with that game/ program your crashing with is start slowly and slightly bump up your clocks and all to the first crash you get and back off a bit from there in till your running stable


now for this ??

''Now the scary part. When I looked back at Afterburner both core and memory clock speeds were in excess of 32470mhz ... yes you read that right. ''

may of just been a glitch cause I never experienced that

thing is a cards overclock is a lot of luck of the draw some card do it well some may not at all and a oc that's stable in a lot of thing can still be bad under that one all you can do is hope for the best result or you may find you cant got anything much at all

Great info man thanks!

I noticed you too were limited to how much you can boost the voltage. I see some people having much larger boosts, but in NVI and AB it only lets you overvolt by +27mv or was that +37mv, either way it's much less than what I've seen other people achieve.

Is that limitation directly related to which model card you have ? Can it be increased further ?

 

[junkeymonkey]

a lot of folks like afterburner over evga precision x , but I got better results from evga's tool ? not to say its better but seemed so for my card ? then that furmark test had to be a joke seems if your looking to damage a card use it [opinion] lol.....


I assume you got enough PSU to hold up to that sli running at full load ?


''I noticed you too were limited to how much you can boost the voltage. I see some people having much larger boosts, but in NVI and AB it only lets you overvolt by +27mv or was that +37mv, either way it's much less than what I've seen other people achieve. ''

luck of the draw ?

also becareful on what you see they maybe using a custom modded bios as well to cheat all that some thing you got to take with a grain of salt


check your cards bios limits

Board power limit
Target: 180.0 W
Limit: 225.0 W
Adj. Range: -33%, +25%
Thermal limits
Rated: 79.5C
Max: 91.0C
https://www.techpowerup.com/vgabios/162173/evga-gtx980-4096-140910

Board power limit
Target: 216.0 W
Limit: 270.0 W
Adj. Range: -44%, +25%
Thermal limits
Rated: 79.5C
Max: 91.0C

https://www.techpowerup.com/vgabios/168541/evga-gtx980-4096-141111

 

[spacejunk]

modern graphics cards are very resilient against abusive overclocking, your cards are fine. run furmark/kombuster and do a stability test.

I really do hope so. Interesting is the fact that it benched quite well but the game made it crash. Maybe heaven 4.0 wasn't pushing it hard enough ?

 

[junkeymonkey]

fail safes are built in to the cards bios by NVidia it would hvve to be bios mod abuse or physical damage like a pcb diode or transistor just burned up

a card just using simple tol like afterburner will / should not allow you to go more then NVidia speced and what you crash when you get with in x-amount of it before any damage occurs

 

[spacejunk]

a lot of folks like afterburner over evga precision x , but I got better results from evga's tool ? not to say its better but seemed so for my card ? then that furmark test had to be a joke seems if your looking to damage a card use it [opinion] lol.....


I assume you got enough PSU to hold up to that sli running at full load ?


''I noticed you too were limited to how much you can boost the voltage. I see some people having much larger boosts, but in NVI and AB it only lets you overvolt by +27mv or was that +37mv, either way it's much less than what I've seen other people achieve. ''

luck of the draw ?

also becareful on what you see they maybe using a custom modded bios as well to cheat all that some thing you got to take with a grain of salt


check your cards bios limits

Board power limit
Target: 180.0 W
Limit: 225.0 W
Adj. Range: -33%, +25%
Thermal limits
Rated: 79.5C
Max: 91.0C
https://www.techpowerup.com/vgabios/162173/evga-gtx980-4096-140910

Hmm, I might try switching to that EVGA tool. It would seem Afterburner doesn't like me very much.

The PSU is a Corsair AX-1200 which I think is enough for my system.

I'll check out the limits for these 980's, never really thought to do that before. As a general rule I never have them exceeding 70c, but I don't know about the power limits.

There's a video of a guy on Youtube overclocking a 980 and his card can boost by 230mv. I want to learn his secret, or maybe it's better that I don't.

Cheers!

 

[spacejunk]

fail safes are built in to the cards bios by NVidia it would hvve to be bios mod abuse or physical damage like a pcb diode or transistor just burned up

a card just using simple tol like afterburner will / should not allow you to go more then NVidia speced and what you crash when you get with in x-amount of it before any damage occurs

The funny thing about it is, the OC I set wasn't all that crazy, what I am concerned about is the possibility that when Afterburner glitched, it wasn't playing by the rules of the cards limitations.

Is it possible for Afterburner to override that safe-guard while in an erroneous state ?

 

[junkeymonkey]

http://forums.evga.com/EVGA-980TI-Burnt-out-m2571295.aspx

well you could keep your eye on your psu when running something

http://imgur.com/N4V6t5Q

http://imgur.com/AvRnRgU

may help to monitor your values max and min

even the best of psu can fail

 

[vrumor]

Sorry to inform you, it was neither. You asked a yes or no question, got an answer and then down voted it when it was correct. And down voted me. LOL. Pretty sure he isnt the one being anal here.

 

[junkeymonkey]

this did not post above [sorry]

http://imgur.com/Ak24Pbi

 

[spacejunk]

http://forums.evga.com/EVGA-980TI-Burnt-out-m2571295.aspx

well you could keep your eye on your psu when running something

http://imgur.com/N4V6t5Q

http://imgur.com/AvRnRgU

may help to monitor your values max and min

even the best of psu can fail

Holy crap! Now that card really did blow up lol.

I need that tool you are running looks very handy indeed. Thanks for posting all that very helpful!

As a side-note, my maximums look very similar to what I see in your screenshot. Perhaps this is normal then.

Back to stress testing hehe.

 

[RCFProd]

Lol thanks very helpful.

I'm saying it because it makes your title look ''click-baity''. You mention both GPU's work, yet your title is if you blew them up. Makes no sense.

 

[Saga Lout]

I think this thread has run its course and I'm closing it before it degenerates into a farce.