Crashfest with MSI GXT 980 Ti Gaming 6G

jyi786

Supreme [H]ardness
Joined
Jun 13, 2002
Messages
5,757
Bought an MSI GTX 980 Ti Gaming 6G. So far, here are the games I've played on it:

1. Crysis 3
2. COD Advanced Warfare
3. Borderlands: TPS
4. Far Cry 4

Out of the those 4, Crysis and Borderlands have been completely stable. But COD and now Far Cry 4?

An absolute f****** crashfest mess.

Both of them are completely unstable. I get random crash to desktop (CTDs), nVidia driver TRDs, freezing, you name it, it goes with those two games. I've tried everything with the card: backing down clocks, placing an artificial limit on the temps, going extreme and setting power limit/temp to 109/91, nothing can get these stable. I've also tried multiple beta nVidia drivers, setting power to prefer max performance instead of adaptive. Nothing can get those two games stable.

Now I know for certain that I was able to play COD:AW on my old GTX 570 with minimal to no crashes. I also know that Far Cry 4 is a POS and that Ubisoft still hasn't fixed it and that it crashes so much. What I can't narrow down is if it's the card or just two very bad examples of games that are buggy by default that is causing these crashes.

Thanks, and any help/suggestions that could be provided would be much appreciated.

Edit: system specs.

CPU: 2600k @ 3.4Ghz
RAM: Corsair 32GB
HDDs: Samsung SSD 830, Samsung 1TB spinner, WD 750 spinner and WD 500 spinner
PSU: XFX 1250W Pro
Case: Corsair 600T
Cooling: 200mm intake, 200mm output (at top), 120mm at back
Ambient temp: 70F to 75F

Windows 7 (been on the current install for about 2 years), and it's perfectly stable. I'm sort of an expert/pro in keeping shitty drivers at bay and doing proper installations of drivers/software. All games I've played have been stable...until I got this 980 Ti.
 
Last edited:
Bought an MSI GTX 980 Ti Gaming 6G. So far, here are the games I've played on it:

1. Crysis 3
2. COD Advanced Warfare
3. Borderlands: TPS
4. Far Cry 4

Out of the those 4, Crysis and Borderlands have been completely stable. But COD and now Far Cry 4?

An absolute f****** crashfest mess.

Both of them are completely unstable. I get random crash to desktop (CTDs), nVidia driver TRDs, freezing, you name it, it goes with those two games. I've tried everything with the card: backing down clocks, placing an artificial limit on the temps, going extreme and setting power limit/temp to 109/91, nothing can get these stable. I've also tried multiple beta nVidia drivers, setting power to prefer max performance instead of adaptive. Nothing can get those two games stable.

Now I know for certain that I was able to play COD:AW on my old GTX 570 with minimal to no crashes. I also know that Far Cry 4 is a POS and that Ubisoft still hasn't fixed it and that it crashes so much. What I can't narrow down is if it's the card or just two very bad examples of games that are buggy by default that is causing these crashes.

Thanks, and any help/suggestions that could be provided would be much appreciated.

Are you using 353.30 or below? If so, switch to 353.49. Much more stable. You can find it here.
 
Try downclocking the card - sometimes those factory overclocked variants are clocked too high for chip inside.

edit: if it helps then it's RMA or refund time
 
Last edited:
It would also help if you posted what your setup was like so we can have an idea.

Things like CPU, MOBO, RAM, Hard drive configuration, which version of windows you have installed and for how long have you been running this particular installation (6mo? a year?) since your last install/reinstall/reformat, what 3rd party apps/drivers you have running apart from the video card, what's your ambient room temperature vs what your components are reporting in speccy or speedfan, and what's your cooling setup like?

Your video card might be defective and MSI has a pretty good return policy, not as great as some other vendors, but I'd get it to them ASAP if it turns out that lowering your clockspeeds is what fixes the issue. :cool:
 
Sounds exactly like the default high clocked 980 card I just sent back.
At stock clocks, the only way I could get it 100% stable was to put power % at max, downclock it 100MHz and give it a small voltage boost.
(I'm now on a 980ti and its rock stable)

Try the same things, see what it takes to get it stable.
Its possible your problem is similar.
 
Test the card in another system with the same games. Farcry 4 has never crashed for me, and i've finished it using old 970 + new 980 setup.

I'd do a fresh windows install on another hdd or smth, put only fc4 in there and see if still crashes. If it does, i would return the card asap.
 
with 353.30, I had driver recovery over and over and over again, someone here suggested moving to 353.38, i tried it, all fixed, btw, mine was with a GTX 970.

GL! hope you get it resolved..
 
Bought an MSI GTX 980 Ti Gaming 6G. So far, here are the games I've played on it:

1. Crysis 3
2. COD Advanced Warfare
3. Borderlands: TPS
4. Far Cry 4

Out of the those 4, Crysis and Borderlands have been completely stable. But COD and now Far Cry 4?

An absolute f****** crashfest mess.

Both of them are completely unstable. I get random crash to desktop (CTDs), nVidia driver TRDs, freezing, you name it, it goes with those two games. I've tried everything with the card: backing down clocks, placing an artificial limit on the temps, going extreme and setting power limit/temp to 109/91, nothing can get these stable. I've also tried multiple beta nVidia drivers, setting power to prefer max performance instead of adaptive. Nothing can get those two games stable.

Now I know for certain that I was able to play COD:AW on my old GTX 570 with minimal to no crashes. I also know that Far Cry 4 is a POS and that Ubisoft still hasn't fixed it and that it crashes so much. What I can't narrow down is if it's the card or just two very bad examples of games that are buggy by default that is causing these crashes.

Thanks, and any help/suggestions that could be provided would be much appreciated.

I returned mine for instability as well. Replacing it with a G1. I'm pretty disappointed with MSI. I think it's not so much the chip, but that the power delivery system of the PCB having issues. I replaced my PSU due to random reboots, and now it's no longer rebooting, but it straight up loses signal on the HDMI. It was rock solid stable and all of a sudden it's crashing 30 seconds into Valley benchmark with no change in drivers or anything else.

Run Valley Benchmarks for 2 hours. If it doesn't crash, then it's the game. If it crashes, then it's probably the card.
 
Are you using 353.30 or below? If so, switch to 353.49. Much more stable. You can find it here.

Tried that already. Matter of fact, I'm on 353.49 right now.

Try downclocking the card - sometimes those factory overclocked variants are clocked too high for chip inside.

edit: if it helps then it's RMA or refund time

Definitely will do this.

It would also help if you posted what your setup was like so we can have an idea.

Things like CPU, MOBO, RAM, Hard drive configuration, which version of windows you have installed and for how long have you been running this particular installation (6mo? a year?) since your last install/reinstall/reformat, what 3rd party apps/drivers you have running apart from the video card, what's your ambient room temperature vs what your components are reporting in speccy or speedfan, and what's your cooling setup like?

Your video card might be defective and MSI has a pretty good return policy, not as great as some other vendors, but I'd get it to them ASAP if it turns out that lowering your clockspeeds is what fixes the issue. :cool:

Sorry about that. Updating the OP with the info you guys need. It's definitely not my system itself (have pretty beefy specs).
 
Oh, something else I should mention. When I first got the card and was having these crashes with COD:AW, I put the card through 9 hours of MSI Kombustor (the full test where it even loads the memory) with ambient temps about 90F. The core temps reached 89C, but it never crashed once, so I thought the card was stable.
 
My first MSI 980Ti woudln't work outa the box properly either. It didn't perform like a 980Ti, it ran like at 980 standard speeds, no matter what I did, even after a full system rebuild. I sent it back
 
89c is a little too hot for comfort.

89C at a ridiculous ambient temp and using MSI Kombustor. This will get the card up to temps that you will never see in real life. Highest I've personally seen in real life was 83C.

Ok guys, ran the full gamut here.

Unigine Valley benchmarks ran with card at complete defaults. No crashes for 2 hours. I then started Far Cry 4. Here's what I did.

1. Increased core voltage. This made no difference, as the card always set it back to 1.162V. Game still crashed.
2. Backed down core clock and memory clock as far as I could. Game still crashed.
3. Put artificial temp limit to 70C to see if it was a heat issue. No matter, game still crashed.
4. Raised temp limit and power limits to absolute max. Game still crashed.

These crashes are completely random in nature. What they do is kick me back to the desktop, but without any error, and the sound continues.

So, sounds like it's not the card? Next up, I'm going to start disabling things in the game option menus, like Soft Shadows, because I've heard those were terribly buggy.
 
Turn off Hairworks entirely for Witcher 3 and Fary Cry 4 is still a buggy POS, even after all the patchs. Run Witcher 3 more. Honestly sounds like a bad card, RMA it
 
My card was bad and it crashed within minutes of running Valley. If he's able to run Valley for 2 hours, I'm not sure that it's the card. Try a different driver and reinstall the game.
 
My card was bad and it crashed within minutes of running Valley. If he's able to run Valley for 2 hours, I'm not sure that it's the card. Try a different driver and reinstall the game.

Valley stresses memory more then the core from what i've heard. Far Cry 4 is a notoriously buggy and shitty game, play something more stable
 
My card was bad and it crashed within minutes of running Valley. If he's able to run Valley for 2 hours, I'm not sure that it's the card. Try a different driver and reinstall the game.

Valley runs fine for more than an hour. MSI Kombustor, 9 hours with no crashes.

Turn off Hairworks entirely for Witcher 3 and Fary Cry 4 is still a buggy POS, even after all the patchs. Run Witcher 3 more. Honestly sounds like a bad card, RMA it

Valley stresses memory more then the core from what i've heard. Far Cry 4 is a notoriously buggy and shitty game, play something more stable

I don't play Witcher. Crysis 3, Borderlands TPS, and Metro LL (from what I can remember) all stable. Far Cry 4 and COD:AW are both crashfests.
 
Have you tried completely removing all traces of the drivers and installing fresh, like you'd do switching brands of GPU?
 
Have you tried completely removing all traces of the drivers and installing fresh, like you'd do switching brands of GPU?

Sure did. Did that when I first installed my GTX 980 Ti. Down to using the "set devmgr_show_nonpresent_devices=1" command line.
 
If that is all well, then I chalk it up to far cry 4 being and unstable pos as it would not run stable on my overclocked 980 either when all my other games were fine. I finally Uninstalled it with no desire to finish it
 
Sure did. Did that when I first installed my GTX 980 Ti. Down to using the "set devmgr_show_nonpresent_devices=1" command line.

Had to ask sometimes the obvious gets overlooked :).

I don't have FC4 but looking over forums I've found some settings that seem to be working for people (at 1080p, if you're higher you might lower some of these but try these first anyway):
vsync - on

quality settings
motion blur off
Textures - Ultra
Shadows - High
Post FX -Ultra
Geometry - Very High
Vegetation- Ultra
Environment - Ultra
water - Ultra
ambient occlusion: SSBC
SMAA for Antialiasing

god rays - off
fur - off
trees relief -off

advanced settings
triple buffer - off
Vsync mode - normal

NVidia control panel:
16 X ansiotropic filter enabled


I do seem to be seeing various comments by people who upgraded to 980ti having similar problems to you in FC4, some titan owners as well.
 
I have a Gigabyte OC reference 980ti and its rock fckn stable!

I would RMA that card or your drivers are crud and you need better drivers.

I am literally using the absolute latest drivers from nVidia with no problems.
 
Well what do you know. Installed my old GTX 570 and no crashes.

I believe I am at wits end with this MSI card. I will be getting an RMA for a refund for it tomorrow. I also am going to create a thread about reference 980 Ti vs. custom ones, because it looks like I'm going to have a preference about the type that dumps the heat OUTSIDE the case instead of inside, as is the case with all the custom ones so far.
 
Update: I found a MicroCenter close to my house that had the Asus GTX 980 Ti I just linked, so I picked it up this morning and installed it.

Far Cry 4 STILL crashes. Just got a CTD when playing after 3 hours.

I'm going to try at least one other game to see how it behaves.
 
I havent used DSR on my 980ti since I got it about a week ago because I forgot about it.
I've been trying it for the last day and it has become crashy, even at stock speeds.
I can hardly start a race in Project Cars and I get an NVidia driver crash, GPU use is around 50% so its not taxing the card.
This is with 353.30 WHQL.

Previously, I could run wildly overclocked with lots of AA with no problem.
fyi in case you are using DSR.
 
I havent used DSR on my 980ti since I got it about a week ago because I forgot about it.
I've been trying it for the last day and it has become crashy, even at stock speeds.
I can hardly start a race in Project Cars and I get an NVidia driver crash, GPU use is around 50% so its not taxing the card.
This is with 353.30 WHQL.

Previously, I could run wildly overclocked with lots of AA with no problem.
fyi in case you are using DSR.

353.30 are terrible, I had every game produce driver crashes, use 353.38 . Haven't had a single crash yet.
 
353.30 are terrible, I had every game produce driver crashes, use 353.38 . Haven't had a single crash yet.

I went looking for them last night on Nvidias site and under the beta search, I selected search for all and it gave me 353.06 and 353.30 WHQLs only.
No betas or any other drivers.

Thanks for the reminder of the version number, I'll go find them now.
 
353.30 are terrible, I had every game produce driver crashes, use 353.38 . Haven't had a single crash yet.

Im also using 353.38, they seem to be best for Ti's


I went looking for them last night on Nvidias site and under the beta search, I selected search for all and it gave me 353.06 and 353.30 WHQLs only.
No betas or any other drivers.

Thanks for the reminder of the version number, I'll go find them now.

http://nvidia.custhelp.com/app/answers/detail/a_id/3694/~/geforce-hotfix-driver-353.38
 
353.38 took a lot longer to install.
After my PC wasnt responding well and NVidia control panel wouldnt open.
I remembered the last time this happened was with an older driver on my 980 card (non ti) and the solution was to use only one display when DSR is enabled.
So I disabled multi display with Windows settings and yep, same old bug is in 353.38. It was fixed in 353.30.
Guess I'll keep display cloning disabled for now, it made it easy for me to switch displays on/off and be working without having to change anything.

Plus side is no crashes now :)
Cheers.
 
Alrighty, so since my other thread got closed, I'll update this one.

Took some extra time with the reference card. Bottom line:

I'm returning it and keeping my MSI. It seems the MSI is not defective after all. The reference crashes just as much as it does in Far Cry 4. So it's definitely a game/driver issue and not the card. One of the reasons for me keeping the MSI is because there is ZERO coil whine. After you've heard silence, any kind of coil whine is going to grate your nerves.

The MSI definitely dissipates heat much faster than the reference card. It can also keep GPU usage higher and in a straight line much better than the reference can. This was definitely intriguing to me. Also, it reaches 83-84C and stays there much more easily than the reference.
 
A couple of things I've noticed. I'm using the same driver on my 980ti as my 770's 353.30. The sound kept going off the default every boot (my Panny ax800u) and I kept having to reset it, until the last couple of reboots where it finally seems to have stuck well hmmm. Alt-Tabbing out of Civ5 cause a system lockup, the worst that ever happened (rarely) with the 770's was the game would crash but could always ctrl-alt-del. I'm thinking the 980ti (and Titian) being fairly new have some issues still that may hopefully get worked out in the drivers. One possible thing I did that might cause the above problem was activate yCbCr444 but it makes a huge difference in readability in Civ5 at 4K otherwise settings are pretty similar to the 770's, just faster.

Anyway did you try using a known good set of game settings for your resolution like the ones I posted above for 1080p? At least that way you know the game settings hopefully aren't a problem or adding to the problem. Also what's your basic hardware specs? Might be interesting to see how they compare to other people with similar problems. Also have you tried killing all background processes that the game doesn't need...maybe FC4 + 980ti are particularly sensitive to something?
 
Anyway did you try using a known good set of game settings for your resolution like the ones I posted above for 1080p? At least that way you know the game settings hopefully aren't a problem or adding to the problem. Also what's your basic hardware specs? Might be interesting to see how they compare to other people with similar problems. Also have you tried killing all background processes that the game doesn't need...maybe FC4 + 980ti are particularly sensitive to something?

Unfortunately, I don't know of a "good set" of game settings. All I did was get my 980 Ti, then the game, then cranked up all the settings.

Basic hardware specs are in the OP.

Yes, I also tried killing background processes. It helps some, but again, it's really hit and miss.
 
Look in post 21 I put in a set of settings I found that are working for several people, who have not experienced crashes (or so they claim), at least it's a starting point. What resolution are you using? And Motherboard. Also I'd assume your system drivers and BIOS are all up to date. If FC4 works in off-line mode try that killing steam and all its processes.
 
Look in post 21 I put in a set of settings I found that are working for several people, who have not experienced crashes (or so they claim), at least it's a starting point. What resolution are you using? And Motherboard. Also I'd assume your system drivers and BIOS are all up to date. If FC4 works in off-line mode try that killing steam and all its processes.

Yeah, I might give those a shot.

Forgot to include the motherboard. It's an Asus P8Z68-Deluxe GEN 3. All drivers up-to-date.

Here are my Far Cry 4 settings.

2560x1440
All options set to Ultra
AA set to SMAA (TXAA gives weird shimmering issues)
All Gameworks stuff other than HBAO+ set to Ultra or turned off, because the game crashes quicker with them on
All Vsync set to OFF, in game as well as in the nVidia control panel
All settings in nVidia control panel set to High Quality (no performance or quality)

It's important to note that COD:AW behaves the same as Far Cry 4. Crashing randomly. I'm sure it's not just Far Cry 4.
 
So still trying to figure out all the crashes, and I want to move to a stable newer game that doesn't crash. COD:AW and Far Cry 4 are like the worst you can test with, so that's out the window.

How is Dying Light for stability?
 
Have you tried running prime? Just for shits and grins, never hurts.

I had to update the bios on my motherboard to get my ti running. When I redid my oc it was mildly unstable and causing issues.

Better yet, check event viewer for whea errors.
 
Have you tried running prime? Just for shits and grins, never hurts.

I had to update the bios on my motherboard to get my ti running. When I redid my oc it was mildly unstable and causing issues.

Better yet, check event viewer for whea errors.

Yes. Ran that a looonnnnnngggg while ago. Rig is rock stable, so not that.

Update: Bought and played about 3+ hours of Dying Light. No crashes at all. So must be those two shitty games (Far Cry 4 and COD:AW).
 
Have you checked event viewer?

Always. I watch it like a hawk. No event errors to speak of....other than when FC4 was crashing left and right. Of course, those event logs didn't give any useful information.
 
Back
Top