Asus Strix 980 not fully stable? Far Cry 4 - Display driver stopped responding

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

litwicki22

Senior member
Sep 13, 2012
340
0
0
You will change card to another and problem can be the same. It can be driver issues. Many peoples have TDR in FC4 on random drivers. It can be problem game\drivers or something else. Even update issue. Like i say when i revert from update 1.7 to 1.6 i dont see single crash.
 

xorbe

Senior member
Sep 7, 2011
368
0
76
FC4/BL + OG Titan + any driver = no crash, that's all I care about. They can keep their Titan X even if the driver is to blame. It's a package deal -- both need to work. If playing with game patches to tip-toe around broken card/driver issues works for you, that's great I guess.
 

Magic Carpet

Diamond Member
Oct 2, 2011
3,477
233
106
FC4/BL + OG Titan + any driver = no crash, that's all I care about. They can keep their Titan X even if the driver is to blame. It's a package deal -- both need to work. If playing with game patches to tip-toe around broken card/driver issues works for you, that's great I guess.
I feel your pain, especially if FC4 happens to be one of your favorites. Personally, I still think it's a software issue of some sort. Are you going to ask for a replacement?
 

litwicki22

Senior member
Sep 13, 2012
340
0
0
FC4/BL + OG Titan + any driver = no crash, that's all I care about. They can keep their Titan X even if the driver is to blame. It's a package deal -- both need to work. If playing with game patches to tip-toe around broken card/driver issues works for you, that's great I guess.
If i was had a broken card, then on any game card will crash or benchmark and on all updates on FC4. Like i say i have crashing display driver on FC4 with 350.12 WHQL but on 1.7 update. On 1.6 is fine. Even that i can have broken card? What you think about it? Of course i can RMA card always.
 

xorbe

Senior member
Sep 7, 2011
368
0
76
I feel your pain, especially if FC4 happens to be one of your favorites. Personally, I still think it's a software issue of some sort. Are you going to ask for a replacement?

The official nVidia store only does refunds. So then I'll need to buy another card later. You may be right about the "software issue" aspect -- it's not a classic TDR / momentary black screen / gpu fan spin down and back up / reset deal. It just drops to desktop and won't let go of the screen. If I then log out and back in, sometimes the OS spontaneously reboots after that. It's not in a happy state after that.

Litwicki, I discovered that the nv control panel said "350.12" after I had installed 350.05 HotFix -- so I uninstalled the driver and re-installed 350.05 for sure this time. FC4 is still going at 3 hours and counting, eh -- nope, finally crashed.

I have one last thing to try, and that's to swap in the ram from my previous build. (Which I did, but broke a tab off of my CM Hyper Evo 212+ ...) No dice, still crashes.
 
Last edited:

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
I have reason to believe that LOTS of those cards, whether from ASUS, EVGA or whomever, especially those who are "factory overclocked" are very much operating at hardware tolerance/border-line.

Short: A dollar saved here and there from using cheap components.

And this is when you get black screens, spontanous reboots etc.

A typical TDR is more like the driver crashing and then recovering (which usually happens from too high overclocks)...but those black screens and cards disappearing from device manager is like that power is insufficient or not stable, so basically the card(s) really crash/shut off and you need a hard-reset to get it to live again. IMHO likely due to problems with voltage regulators and general poor components.

It would be interesting to see whether cards with 8pin+6pin connectors experience this too or whether this issue is mainly with cards with two 6pin PCIE connectors. (Or one 8pin like the Strix has), so basically only with cards that can not draw more than 150W from the connectors.
 
Last edited:

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
>>
I tested Unigine Heaven for 10 hours
>>

If you can really run Heaven on Ultra/Extreme for 10 hours...I'd say it is related to the game. I haven't found a more demanding test than Heaven 4.0. So it might really not be your card after all.
 

litwicki22

Senior member
Sep 13, 2012
340
0
0
Updated not helped. Crash today morning.

Overall I am getting very random driver crash only in Far Cry 4. Black screen and Application stopped . Then i saw on events: Nvlddmkm Display driver stopped responding and has recovered. Its very random , sometimes i can play even 12 hours without driver crash and sometimes like today its crashed driver after 5 minutes of started game.
Its problem with driver or my its just my card? ( i am using stock Asus Strix 980 ). I am using the newest drivers. All other games runs without even single crash. I tested Unigine Heaven for 10 hours, without crash. I tested Metro Last Light , Crysis 3, WatchDogs, for few hours without crash too.

Only Far Cry 4 sometimes crash after 5 minutes,but mostly often i can play 12 hours without driver crash.

Using 350.12 WHQL. Also i tested 347.88 WHQL and the same problem.


All my pc is stock, cpu is 4790k stock. GPU is on stock too.



Only Far Cry 4 is crashing my driver very randomly , sometimes after 5 minutes, but sometimes i can play 12 hours without crash.
Its something with my card?

Guys i dont know what to do now. RMA my Strix 980 because sometimes its crashing driver only in that game? Is any chance that on another Strix 980 example i will be have the same problem?
 

Magic Carpet

Diamond Member
Oct 2, 2011
3,477
233
106
@litwicki22

You tried underclocking yet? Try underclock and give it a bit more voltage, see if it helps. If not, RMA the card. Just be happy :)
 

litwicki22

Senior member
Sep 13, 2012
340
0
0
No i dont want mess with voltages. So rma card if it crashing on stock very randomly in Far Cry 4?


Also..yesterday i played 12 hours without driver crash. Today i run game morning and after 5 minutes driver crash. Then again for long time game will be stable without driver crash. And after some some hours it will crash again.
Its something with my card? Oc is not stable on my card? ( oc is by factory GPU CORE BOOST is 1324mhz i dont oc manually )
But like i say other games benches are rock stable.
 
Last edited:

Magic Carpet

Diamond Member
Oct 2, 2011
3,477
233
106
You can reduce the Power Limit tab in Afterburner from 100% to about 80%, it will essentially underclock the card. It will not damage your card in any way. See if it helps.

Just do that.
 

litwicki22

Senior member
Sep 13, 2012
340
0
0
>>
I tested Unigine Heaven for 10 hours
>>

If you can really run Heaven on Ultra/Extreme for 10 hours...I'd say it is related to the game. I haven't found a more demanding test than Heaven 4.0. So it might really not be your card after all.
Yes i tested Heaven on extreme 2 times. One time i left running 10 hours, without crash. Second time i left running for 8 hours , without crash too. Also i played other games, like WatchDogs, Crysis 3, Metro LL and all fine.
Only FAR Cry 4 is crashing my display driver very randomly, i will say very rarely. Screen turns to a black for a second and then app crash. Event viewer says about display crash.
One time i can play FC4 even 12 hours without crash, second time it will crash after 5 minutes. It can be game issue even if it triggers TDR?


Well, I accept that an application crashes, but NOT that it triggers a TDR. If Far Cry 4 crashes in a "normal way" (screen freeze, Windows message telling that the application has crashed), fine. But TDRs are a big no-no.

It can be game issue even if it triggers TDR?
 

litwicki22

Senior member
Sep 13, 2012
340
0
0
But all I want to know is: can a hangup in a game cause a TDR? Game can cause TDR? Or its only hardware fault?
 

xorbe

Senior member
Sep 7, 2011
368
0
76
But all I want to know is: can a hangup in a game cause a TDR? Game can cause TDR? Or its only hardware fault?

TDR can literally be caused by anything: psu, motherboard, cpu, system memory, gpu, vram, any driver (network, storage, graphics, etc), software (application or game), etc.

Same problem will the system ram from my previously stable system (old 2x4GB vs new 4x8GB), crashed in 5 minutes this morning.

I ordered another Titan X, so I will know by the weekend. D: I ordered the EVGA SC card and will back-flash it to stock nVidia. Seems like the best odds for a stable card.
 
Last edited:

litwicki22

Senior member
Sep 13, 2012
340
0
0
First Far Cry 4 is crashing display driver very randomly. ( sometimes i can play many hours , sometimes crash after 5 minutes ).
Black screen and Application stopped . Then i saw on events: Nvlddmkm Display driver stopped responding and has recovered. I blame Far Cry 4,ok, nevermind. Many users get TDR crashes in FC4 only.
But today TDR crash happened first time in GTA V.
So what now? I have random ( very rarely ) driver crashes in FC4 and today i notice my first TDR in GTA V.
I am using newest 350.12 WHQL. Is card really unstable? Also i can run 12 hours Unigine Heaven Extreme without crash.I played Crysis 3, Watchdogs, Metro Last Light many hours without crash too.
But today second game ( GTA V ) and TDR crash , makes me really worried. Thx
 

xorbe

Senior member
Sep 7, 2011
368
0
76
I don't know why you post in circles. You never post any results of trying down-clocks or voltage bumps or anything. You had the same problem with ASUS Strix 780 on the [H] forum. Now ASUS 980 Strix here.

BTW I finished FC4 south island with my older OG Titan. It just doesn't crash.
 
Last edited:

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,250
136
Have you tried using Display Driver Uninstaller when switching drivers? Clean install option?

Far Cry 4 works fine on my gimped 970 boosting to around 1500 +/- a bin or two here and there.
 

ocre

Golden Member
Dec 26, 2008
1,594
7
81
Has the OP ran memtest overnight? if he doesnt want to do that, How about downloading the free intel burn test, set it to maximum and run 10 loops
 
Last edited:

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
* run your card underclocked about 50 or more mhz and test
* run your card at 80% power limit and test
* test the rest of your system (OCCT, memtest etc.)
* make sure you have pagefile on your system drive activated and it's set to "automatically manage" and also there is plenty of space (8gb+ left)
Does it crash on any of those settings?
 

deanx0r

Senior member
Oct 1, 2002
890
20
76
My GTX970 has been plagued with TDR crashes in StarCraft2. It's stable everywhere else except in that game. The crashes are so random and occurs when the card isnt being pushed hard (50% at most) and the system is stock. I had the card since launch and was hoping this issue would be fixed by now.

I might jump back to AMD much sooner if this issue persists.
 

xorbe

Senior member
Sep 7, 2011
368
0
76
If the worst comes to the worst, I'll flash one of the Titan X cards with the "no boost" rom with constant voltage. The problem is the darn boost and 40 frequency bins and multiple voltage targets. They just don't qual the cards well enough.
 

flexy

Diamond Member
Sep 28, 2001
8,464
155
106
The problem is the darn boost and 40 frequency bins and multiple voltage targets. They just don't qual the cards well enough.

I actually think this is LIKELY. There are reports/rumours that some cards are stable at max. boost but have problems with the lower clocks.

It is POSSIBLE that the voltages in the voltage table (and yes, it's a bunch since it's 40 something clocks) for the lower clocks are too low or that the common voltages used in the BIOSes are just not tested enough.

Making sure 40 different clocks having the right voltage sure is a different story than years ago where all we had was ONE clock and didn't need to bother with that boost nonsense... you set your max overclock, voltage and then you were all set...
 
Last edited:

litwicki22

Senior member
Sep 13, 2012
340
0
0
I have question.
If Far Cry 4 is rock stable on 344.75WHQL and not stable ( TDR issues ) on 350.12 WHQL rma card or not?
 

Deders

Platinum Member
Oct 14, 2012
2,401
1
91
I actually think this is LIKELY. There are reports/rumours that some cards are stable at max. boost but have problems with the lower clocks.

It is POSSIBLE that the voltages in the voltage table (and yes, it's a bunch since it's 40 something clocks) for the lower clocks are too low or that the common voltages used in the BIOSes are just not tested enough.

Making sure 40 different clocks having the right voltage sure is a different story than years ago where all we had was ONE clock and didn't need to bother with that boost nonsense... you set your max overclock, voltage and then you were all set...

I have seen a thread somewhere where a guy flashed his Asus DCUII with a gigabyte Bios to get it to run stable at higher clocks. (this was after unlocking etc) Asus's uses a different voltage controller which has lower voltages than most 780's, so the Gigabyte bios treated the card like it was a normal card voltage wise so yes too much or too little voltage can be a problem.

I have question.
If Far Cry 4 is rock stable on 344.75WHQL and not stable ( TDR issues ) on 350.12 WHQL rma card or not?

I'm going to get on to Nvidia's Driver Feedback page and tell them I've had problems for the last 4 driver sets in various games on my 780. You should do the same for you 970.
 

amenx

Diamond Member
Dec 17, 2004
4,406
2,727
136
I have question.
If Far Cry 4 is rock stable on 344.75WHQL and not stable ( TDR issues ) on 350.12 WHQL rma card or not?
If it works with 344.75, why RMA it??!! I think you've RMA'd several cards before this too, eh? Seeing your history, whatever you get to replace it will probably end up being RMA'd just as well, so whats the point? Just keep it and stick to 344.75 for time being. Drivers and patches come and go, eventually something will work. If you RMA it, you will probably be taken less and less seriously in all future posts.