OpenGL crashes on nVidia

[Shadow Fox]

Looking for help and assistance on a possibly GPU related issue... Several applications and games I run that use the OpenGL environment are consistently crashing on my system.

I've been getting these errors on my GTX 960, on a Windows 7, Windows 8.1, and more frequently on Windows 10.

I've been able to benchmark the card, and overclock it without much hiccup, and the card in most games and applications seems to be working without issue. But as soon as it loads OpenGl based programs, problems happen.

• Driver version has not seemed to have changed anything.

• Installing / not installing Geforce Experience has changed nothing.

• DDU for a fresh install has done nothing.

• Changing Power Management Mode from Adaptive to Performance has changed nothing.

Screenies of the errors / error information...

[Shadow Fox]

No one has any ides or solutions?

This sucks! >,<

[SoulForge]

Shadow,

The only thing I could find referencing this is issues with power management.

1. Open your NVIDIA Control Panel and select “Manage 3D settings”
2. Then, select the “Power Management mode” and change the option from “Adaptive” to “Prefer maximum performance”
3. Save your changes, and your are done!

Having 2 980 Ti's in SLI and preferring OpenGL in my games/applications over DX I don't see this issue.

[Shadow Fox]

Shadow,

The only thing I could find referencing this is issues with power management.

1. Open your NVIDIA Control Panel and select “Manage 3D settings”
2. Then, select the “Power Management mode” and change the option from “Adaptive” to “Prefer maximum performance”
3. Save your changes, and your are done!

Having 2 980 Ti's in SLI and preferring OpenGL in my games/applications over DX I don't see this issue.

Bah, unfortunately, that hasn't fixed anything. >,<!!

At a loss on what to do next, and Google result showing it happening fairly often on different cards (Some Current Gen. Titans, to cards as far back as the GTX 500's) and driver version; Some solutions working for people, and not for others...

Card doesn't seem to be defective with everything of been able to stress on it; Just OpenGL will crash and lock up / crash the application without crashing Display driver, or even my GPU clocking software...

>,<!!
Which sucks, because a lot of games are using the OpenGL API: https://en.wikipedia.org/wiki/List_of_OpenGL_programs

I'm not even sure what exactly is triggering it; It's so sporadic and random.

[Evildeffy]

Bah, unfortunately, that hasn't fixed anything. >,<!!

At a loss on what to do next, and Google result showing it happening fairly often on different cards (Some Current Gen. Titans, to cards as far back as the GTX 500's) and driver version; Some solutions working for people, and not for others...

Card doesn't seem to be defective with everything of been able to stress on it; Just OpenGL will crash and lock up / crash the application without crashing Display driver, or even my GPU clocking software...

>,<!!
Which sucks, because a lot of games are using the OpenGL API: https://en.wikipedia.org/wiki/List_of_OpenGL_programs

I'm not even sure what exactly is triggering it; It's so sporadic and random.

Try clocking down your cards to either reference levels or sub reference levels.
Do this by using stuff like MSI Afterburner.

See if there's a newer BIOS version out for your GFX card (though be careful with this as it WILL brick your card if it messes up unless you have a dual BIOS switch)

Or go back some 2 (or more) years with driver versions.. you'd be surprised how often older drivers tend to impact stability for nVidia cards.

That's really the last thing I can think of barring a complete reinstall of Windows.

It's quite possible that your PSU or GFX card may be dying as well.. since your stability issues have been reported with either of these but they are very hard to confirm, especially if you don't have HW lying about to test.

[Shadow Fox]

Try clocking down your cards to either reference levels or sub reference levels.
Do this by using stuff like MSI Afterburner.

See if there's a newer BIOS version out for your GFX card (though be careful with this as it WILL brick your card if it messes up unless you have a dual BIOS switch)

Or go back some 2 (or more) years with driver versions.. you'd be surprised how often older drivers tend to impact stability for nVidia cards.

That's really the last thing I can think of barring a complete reinstall of Windows.

It's quite possible that your PSU or GFX card may be dying as well.. since your stability issues have been reported with either of these but they are very hard to confirm, especially if you don't have HW lying about to test.

Was having the issue with this card, with a previous PSU. Different Motherboard, even. Almost all the parts on this build are new (Within the last year.), GPU with OpenGL is the only thing having hiccups.

Windows versions haven't made a difference; OS installed on HDD, or either one of my SSD's haven't changed anything.

Driver Updates for my card only go back about a year for Windows 10; Using the earlier driver hasn't changed a thing.

I have not tried a different BIOS update.. .Not sure if my card has a switch for multiple Bios's.

I'm thinking at this point it is just one of the many problems nVidia has been having lately with their long list of ignored & known problems with no ETA's on fixes. :S

Don't think I can return the card, may be too long for that. But it should still be in Warranty... Got it last Spring / Early summer, I think.

EVGA GTX 960 FTW Edition ACX 2.0 2Gb.
The GPU tool I've been currently using has been off of Steam; EVGA PrecisionX 16 x64, Version 5.3.11


My GTX 660 seems to be solid with OpenGL with no issues on my system; But it is a significant difference in performance on my system.

This was after running Fallout 4 on Ultra, Battlefield 4 on Ultra, WoW on Ultra, Tomb Raider (Benchmark) on High, Elite Dangerous on Ultra, and Rainbow 6 Siege (Benchmark) on Medium-High.

[Evildeffy]

Was having the issue with this card, with a previous PSU. Different Motherboard, even. Almost all the parts on this build are new (Within the last year.), GPU with OpenGL is the only thing having hiccups.

Windows versions haven't made a difference; OS installed on HDD, or either one of my SSD's haven't changed anything.

Driver Updates for my card only go back about a year for Windows 10; Using the earlier driver hasn't changed a thing.

I have not tried a different BIOS update.. .Not sure if my card has a switch for multiple Bios's.

I'm thinking at this point it is just one of the many problems nVidia has been having lately with their long list of ignored & known problems with no ETA's on fixes. :S

Don't think I can return the card, may be too long for that. But it should still be in Warranty... Got it last Spring / Early summer, I think.

EVGA GTX 960 FTW Edition ACX 2.0 2Gb.
The GPU tool I've been currently using has been off of Steam; EVGA PrecisionX 16 x64, Version 5.3.11


My GTX 660 seems to be solid with OpenGL with no issues on my system; But it is a significant difference in performance on my system.

This was after running Fallout 4 on Ultra, WoW on Ultra, Tomb Raider (Benchmark) on High, Elite Dangerous on Ultra, and Rainbow 6 Siege (Benchmark) on Medium-High.

<pic removed for space>

Well as I said try DOWNclocking your card to reference levels or just lower.

The info you posted now should've been included earlier as well as we require every little detail but it's cool now.
With the statements you've made if downclocking/new VBIOS doesn't work well then you just have a crappy card and should exchange it for another.
I am going by this due to the fact that you state your old GTX 660 works perfectly but your GTX 960 craps out like a turd in the same computer.
(NOTE: Your card does NOT have a multi-BIOS switch)

Try downclocking to reference levels (found on the nVidia website) first and/or up the voltage a slight bit and if that doesn't work I'd just return the card.
Honestly VBIOS update is not worth it at this point to possibly fuck up your card/warranty.

[Shadow Fox]

Well as I said try DOWNclocking your card to reference levels or just lower...

...Try downclocking to reference levels (found on the nVidia website) first and/or up the voltage a slight bit

Going to reboot here in a second to reset my HWmonitor and some sensors...


Is this the reference sheet you're talking about? Link.

Google page was trying to sell me GPU cards on Amazon and Newegg when I had typed "nVidia GTX 960 Reference card" >,>

[Evildeffy]

Going to reboot here in a second to reset my HWmonitor and some sensors...


Is this the reference sheet you're talking about? Link.

Google page was trying to sell me GPU cards on Amazon and Newegg when I had typed "nVidia GTX 960 Reference card" >,>

Yes that's the reference clocks of a GTX 960.
Some factory overclocked models work perfectly fine in some rendering paths but if insufficiently tested for another (D3D vs. OpenGL) might cause issues such as this.
This has happened to other graphics cards and manufacturers before so that's why I'm thinking this.

This is assuming that your OpenGL stuff is of course not corrupted in some way.

[Shadow Fox]

This is assuming that your OpenGL stuff is of course not corrupted in some way.

I'm not even sure how to being testing for that. :S

Dota 2, SecondLife, Google Earth, Minecraft would be the only OpenGL Engines / API's I have currently have access to, and have installed on my system.

[Evildeffy]

I'm not even sure how to being testing for that. :S

Dota 2, SecondLife, Google Earth, Minecraft would be the only OpenGL Engines / API's I have currently have access to, and have installed on my system.

Only way to test it is to run a integrity check on DOTA2 via STEAM f.ex.
But as I said if you exchange your GTX 960 for the GTX 660 RIGHT NOW and these things function perfectly then ... it won't be the software.

[Shadow Fox]

I'm able to lower my clock rate.

Can't seem to lower my voltage to the card. Can raise it, though. (In intervals of 6mV at a time.)

If I'm reading it right; HWMonitor says 1.018 V (1200 mV) without touching the voltage.

[Evildeffy]

I'm able to lower my clock rate.

Can't seem to lower my voltage to the card. Can raise it, though. (In intervals of 6mV at a time.)

If I'm reading it right; HWMonitor says 1.018 V (1200 mV) without touching the voltage.

Clock speeds of the core and RAM first.
If it still occurs after that you raise the voltage supplied bit by bit.

If it still occurs after that I recommend going to the store, dropping it in their face, explain the situation and ask for a new card.
(Note: Never say you added more volts to the GPU..... EVER! Just underclocked it)

[Shadow Fox]

Only way to test it is to run a integrity check on DOTA2 via STEAM f.ex.
But as I said if you exchange your GTX 960 for the GTX 660 RIGHT NOW and these things function perfectly then ... it won't be the software.

Switching a 660 into the system at this moment will be a bit difficult; Was meant to be past tense. Currently do not have it it readily available.

My other Card, is a R9 270X, however, it's red-screening and does need to be RMA'd; Just haven't made the time to do so with MSI. (And it's also less performance then this 960)

- - - Updated - - -

Clock speeds of the core and RAM first.
If it still occurs after that you raise the voltage supplied bit by bit.

If it still occurs after that I recommend going to the store, dropping it in their face, explain the situation and ask for a new card.
(Note: Never say you added more volts to the GPU..... EVER! Just underclocked it)

30days is the return policy for refund to EVGA's website; Checking to see if I'm able to on Newegg where I bought the card (I'm betting not, though.), back on... 2015/March/25.

So my option is to get it RMA'd under the 3-year warranty it's on, or deal with the OpenGL Crashes... Or get a new card. (And I'd rather wait until the new cards are released later in the year.)

[Evildeffy]

Switching a 660 into the system at this moment will be a bit difficult; Was meant to be past tense. Currently do not have it it readily available.

My other Card, is a R9 270X, however, it's red-screening and does need to be RMA'd; Just haven't made the time to do so with MSI. (And it's also less performance then this 960)

Is there anyone near you could ninja a GTX 960 or any other nVidia card from?

30days is the return policy for refund to EVGA's website; Checking to see if I'm able to on Newegg where I bought the card (I'm betting not, though.), back on... 2015/March/25.

So my option is to get it RMA'd under the 3-year warranty it's on, or deal with the OpenGL Crashes... Or get a new card. (And I'd rather wait until the new cards are released later in the year.)

Well... not dealing with it sounds not like an option I'd pick but it's up to yourself to determine that.
RMA is still better than nothing if it's the culprit.

[Shadow Fox]

Is there anyone near you could ninja a GTX 960 or any other nVidia card from?


Well... not dealing with it sounds not like an option I'd pick but it's up to yourself to determine that.
RMA is still better than nothing if it's the culprit.

I'm probably going to go the RMA route until A: Fix, or B: the new cards are out... I don't have the cash to sink into a $200-$400 purchase.

[Evildeffy]

I'm probably going to go the RMA route until A: Fix, or B: the new cards are out... I don't have the cash to sink into a $200-$400 purchase.

Hence my suggestion of borrowing someone's card for like 5 minutes to determine if another card doesn't freak out.
Any would do but preferably the same type of card or range of cards would be best.

Anyhow ... off to bed, post progress here.

[Shadow Fox]

ATI R9 270x card sent in for RMA from MSI; Looks like I'm going to be getting it first.

EVGA has a more detailed RMA system then MSI where I have to troubleshoot with their tech support before a RMA can be validated.

Their first suggestion was:

We are sorry to hear that you are having issues with your GPU.
You can try enabling debug mode in the Nvidia control panel. This might help with the issue.
1. Open the Nvidia control panel.
2. Go to help at the top.
3. Enable debug mode.

Debug Mode not only didn't help with the crashes; But the fans on my card stopped turning and heated my GPU up to 62°c.

[Evildeffy]

I take it the clock speed adjustment didn't work and neither did the core voltage increases?

[Shadow Fox]

I take it the clock speed adjustment didn't work and neither did the core voltage increases?

Did not.
EVGA says that nVidia Debug Mode will put the card at nVidia's reference levels when my GPU software is not running.

They pushed the request to a their Advanced RMA program to send me a card ahead of time, and send this one back. Free of charge in Shipping and Replacement unless the current card is physically damaged, then they are going to charge me full-price for the new GPU...

They seem to think that it is a nVidia issue, not an EVGA issue, but they are at least honouring their warranty policies unlike some others...

I would much rather have a refund for this card to push for a purchase in a newer one, but I'm well past that refund date.

As for why Debug mode started to hike up my Temperatures and not turn on the fans, I have no idea...