NVidia OpenGL driver error code 3

mikey186

More TO Explore
I don't know what happened but just a few minutes ago, while I was importing items in CMP, and I have seen this wierd brown-ish screen for a few seconds and back. This didn't happened the first time, today was the first time that brownish screen appeared out of nowhere nearly 2 times. Then an error appeared that said "Nvidia drivers Geforce 8400 GS have stopped working and has been restored" then a second error that said "The nvidia OpenGL driver encountered an error and needs to close" and error code 3 it says right here. I was SO lucky I didn't get a BSOD.

Any idea why I'm getting this? I've researched everywhere to find the problem but no answer found. Help??? I'm running an nVidia Geforce 8400 GS with 512 VRAM
 
Now seconds later, I can't get to device manager because of the same screen again, and then I get this

WSK4wjn.jpg
 
One cause of this is power saving settings in Nvida settings, change power management mode to prefer maximum performance.

Of course it could also be that an 8400GS is a bit past it now, 7 year old, 64 bit DDR2 entry level card, surprised it would even work with TS12 if that's what you are using it with.
 
You think it's safe? I have an ASUS Desktop brought Dec 2011 with NO graphics card preinstalled. Just some Intel graphics thing. And it's a DDR3 with 512 VRAM. And yes it even works with TS12 just with no DX11 support.

Yes it is old, but it's the cheapest card as I can have. And unfortunately, most modern cards are over $300, NO WAY JOSE.
 
You think it's safe? I have an ASUS Desktop brought Dec 2011 with NO graphics card preinstalled. Just some Intel graphics thing. And it's a DDR3 with 512 VRAM. And yes it even works with TS12 just with no DX11 support.

Yes it is old, but it's the cheapest card as I can have. And unfortunately, most modern cards are over $300, NO WAY JOSE.

Check the fan and or heatsink hasn't come loose or the fan hasn't got blocked up with dust, could be its shutting down due to overheating.

Actually there are plenty of newer and better cards a lot cheaper than $300 dollars, over here you can get a GTX650 for £80 which is around $130 and more than adequate for TS12 an entry level card equivalent to the one you are using would be a lot less based on UK prices, around $57 US for an EVGA 1GB DDR3 GTX620.
TS12 only needs DX9c, DX11 isn't fully compatible anyway so you often have to add D9C even with a DX11 card.
 
Check the fan and or heatsink hasn't come loose or the fan hasn't got blocked up with dust, could be its shutting down due to overheating.

Actually there are plenty of newer and better cards a lot cheaper than $300 dollars, over here you can get a GTX650 for £80 which is around $130 and more than adequate for TS12 an entry level card equivalent to the one you are using would be a lot less based on UK prices, around $57 US for an EVGA 1GB DDR3 GTX620.

TS12 only needs DX9c, DX11 isn't fully compatible anyway so you often have to add D9C even with a DX11 card.

Yea I guess from the overheating that caused it. Some say it was clocking issue or I don't even know. I was lucky I did shut down, waited 15 min and turned it back on, no problems thank god.

And yea, i don't know where is the best place to buy a cheap card, you do have a point with the "EVGA 1GB DDR3 GTX620", very high-end. But on my mistakes, I actually brought the same card with the same VRAM (I thought it was a 1GB), but it had a fan and a HDMI output. I was thinking of newegg, but I'm not too sure because of customer sastifaction.
 
Go to New Egg online... www.newegg.com and you can order a card there. You can also pickup these older cards at BestBuy, although their prices aren't the best.

Thanks for the tip. I found a couple of affordable cards, I did found this http://www.newegg.com/Product/Product.aspx?Item=N82E16814127610&cm_re=2gb-_-14-127-610-_-Product but I'm not too sure with the 3 star rating (average).

my specs are in this German site (I think you know what those meant) http://www.sysprofile.de/id101510
 
Thanks for the tip. I found a couple of affordable cards, I did found this http://www.newegg.com/Product/Product.aspx?Item=N82E16814127610&cm_re=2gb-_-14-127-610-_-Product but I'm not too sure with the 3 star rating (average).

my specs are in this German site (I think you know what those meant) http://www.sysprofile.de/id101510

Not a bad system, except for the video card. :)

The card you spec'd out is an ATI (AMD) Radeon which is a different brand. You might want to stick with the NVidia chip if you want and easy upgrade, otherwise you have to search for a file cleanup utility. NVidia and AMD drivers don't mix well! :)

Here's some you might be interested in....

http://www.newegg.com/Product/Produ...escription=nvidia+geforce+620&N=-1&isNodeId=1


John
 
That will work. The GtX6xxx series are not bad cards. They're kind of low end by today's standards, but for most daily use they're probably fine. People used them before for TS12 so there is no reason why you couldn't use it now.

Now having said this... you might want to check the hardware requirements for T: ANE. That will make use of other things that may or may not be available to you in the video card. The best place to check is the hardware requirements thread over in the T: ANE specific portion of the forums.

John
 
Thanks for the tip. Now that it stopped overheating, I'm still nervous of a graphic card overheat. And yet I'm STILL looking for a new one, and I don't know how long am I supposed to have my computer 'naked', since like 3 days ago.
 
In general it's best to run the computer with the side on as it does help with the airflow by forcing the cooler air across the components. You might want to put the side back on and see if things overheat again. If they do, I suggest looking for other causes such as plugged air vents and fans, or even a stopped or stuck fan. If this is the latter case, you will have to replace the fan which is easily done. With plugged fans, you'll need to remove the dirt and dust clumps. This is easily done using canned air which you can get at an office supply store. What you don't want to do though is blast the air at the fans as that can break their bearings or bushings. So, before you do that gently place a pencil or drinking straw into the fan blades to keep the fan from spinning before you blast the air across it to blow the dirt out. Remember, never, ever, ever, use a household vacuum cleaner! This can cause a blast of static electricity to move across the components and zap everything.

With the newer 900-series NVidia cards out now, the other models, which are still perfectly fine, will drop dramatically in price. Keeping this in mind, you might even be able to afford a nice 700-series card by the time T: ANE comes out.

John
 
In general it's best to run the computer with the side on as it does help with the airflow by forcing the cooler air across the components. You might want to put the side back on and see if things overheat again. If they do, I suggest looking for other causes such as plugged air vents and fans, or even a stopped or stuck fan. If this is the latter case, you will have to replace the fan which is easily done. With plugged fans, you'll need to remove the dirt and dust clumps. This is easily done using canned air which you can get at an office supply store. What you don't want to do though is blast the air at the fans as that can break their bearings or bushings. So, before you do that gently place a pencil or drinking straw into the fan blades to keep the fan from spinning before you blast the air across it to blow the dirt out. Remember, never, ever, ever, use a household vacuum cleaner! This can cause a blast of static electricity to move across the components and zap everything.

With the newer 900-series NVidia cards out now, the other models, which are still perfectly fine, will drop dramatically in price. Keeping this in mind, you might even be able to afford a nice 700-series card by the time T: ANE comes out.

John
Thanks for the tip! The only thing it was dusted was just a processor CPU fan. NOT the card so I dusted it off thank goodness. But I think my guess is a clocking issue when it comes to render something. Not like rendering a converter or anything, I actually render 3D via/LuxRender and start the PC in the morning, launch the render, and leave it on as I go out the door letting the renderer do its job until I get back and just stop it. Simple. Oh and it was NON-opencl, just so u know.

But I guess it may be overheating or by a clocking issue. In terms of a GT700 series, maybe I SHOULD get a GDDR5 driver in which now I know in a video that the CPU DOES NOT affect the driver itself. After all, it actually stands for "Graphics Digital Data Rate". If I'm right, then I need to find a affordable GDDR5 (least between $30 to $80). That should (hopefully) run TS10 and TS12, and as well as the upcoming T:ANE.
 
Thanks for the tip! The only thing it was dusted was just a processor CPU fan. NOT the card so I dusted it off thank goodness. But I think my guess is a clocking issue when it comes to render something. Not like rendering a converter or anything, I actually render 3D via/LuxRender and start the PC in the morning, launch the render, and leave it on as I go out the door letting the renderer do its job until I get back and just stop it. Simple. Oh and it was NON-opencl, just so u know.

But I guess it may be overheating or by a clocking issue. In terms of a GT700 series, maybe I SHOULD get a GDDR5 driver in which now I know in a video that the CPU DOES NOT affect the driver itself. After all, it actually stands for "Graphics Digital Data Rate". If I'm right, then I need to find a affordable GDDR5 (least between $30 to $80). That should (hopefully) run TS10 and TS12, and as well as the upcoming T:ANE.

The GPU is probably just getting old. As time goes on components become flaky, some die suddenly or some will become temperature sensitive. By this I mean they will work fine while cool, but as soon as they warm up, they fail. This is due to the little bonds inside cracking and flexing with the temperature changes. In my ancient work life, I was an electronics technician who troubleshot right down to the chip level, and more importantly, it was on video circuitry in video terminals and graphics boards. I learned a ton about how they work and then some by fixing them. Even though I did this over 25 years ago, the theory behind the current video cards is still the same, just they work a lot faster.

There is nothing wrong with what you're doing. In fact this has been done for years by 3d modelers. Way back in the mid to late 1990s, I worked for a company that made training programs. We had a PC running 3ds4 for DOS, and we would let it render overnight since it took 2-3 minutes per frame on the animation. Watching something like that run is like watching paint drip, so we'd test it briefly, head out the door, and collect the .tiff images the next day for processing on the //Fast AG system. At one time, we setup a render farm using all the machines in the office. The boss had a hissy fit over it, but in the end he liked the idea because the data was processed a lot faster.

Don't worry about the GDDR5 stuff. This is a 'selling' feature by AMD to show they have something different from NVidia. What you are more interested in is GPU cores and pathways. NVidia also supports Open/CL and Open/GL quite well and always has. The programs such as 3dsMax make use of the CUDA chip for extra processing and use all the pipelines. I say this from experience since I always used ATI cards until I ran into major compatibility issues with Open/GL and Open/CL. The thing with this stuff too is graphics RAM today is not like it was decades ago. The old VRAM was very, very fast, and I think worked faster than the stuff we use today because of how it was made. The problem is the old VRAM was really, really expensive, and more so than regular RAM ever was. This is why the video cards are relatively inexpensive today. Instead of using expensive VRAM, they use regular DDR5 RAM, or the graphics equivalent.

The thing is if you were to do strictly 3d modeling and nothing else, I'd suggest a CAD-type graphics card. These cards are way beyond your budget, and even well beyond mine too. But because you also play games you need the best of both worlds with a decent game card capable of doing graphics stuff. At the moment, I just replaced my GTX680 with a GTX780Ti. I know it's expensive, but I don't replace the hardware unless it's necessary. Sadly, my '680 had problems and will be replaced under warranty. I'll put the replacement away as a spare as I use the new card going forward. But anyway, the 600-series is getting quite old now in a computer part's lifetime terms, and the 700-series cards are now moving down as the 900s are coming out. NVidia just came out with the 900s and skipped over the 800s. Don't ask me why they did it.

If anything, here's an example of what you can get if you watch the sales at New Egg.... I think you have to sign up for the newsletter/advert though like I did.

http://promotions.newegg.com/NEemai...D-_-Weekend&et_cid=11549&et_rid=67996#Weekend

This came in today.

John
 
Last edited:
The GPU is probably just getting old. As time goes on components become flaky, some die suddenly or some will become temperature sensitive. By this I mean they will work fine while cool, but as soon as they warm up, they fail. This is due to the little bonds inside cracking and flexing with the temperature changes. In my ancient work life, I was an electronics technician who troubleshot right down to the chip level, and more importantly, it was on video circuitry in video terminals and graphics boards. I learned a ton about how they work and then some by fixing them. Even though I did this over 25 years ago, the theory behind the current video cards is still the same, just they work a lot faster.

There is nothing wrong with what you're doing. In fact this has been done for years by 3d modelers. Way back in the mid to late 1990s, I worked for a company that made training programs. We had a PC running 3ds4 for DOS, and we would let it render overnight since it took 2-3 minutes per frame on the animation. Watching something like that run is like watching paint drip, so we'd test it briefly, head out the door, and collect the .tiff images the next day for processing on the //Fast AG system. At one time, we setup a render farm using all the machines in the office. The boss had a hissy fit over it, but in the end he liked the idea because the data was processed a lot faster.

Don't worry about the GDDR5 stuff. This is a 'selling' feature by AMD to show they have something different from NVidia. What you are more interested in is GPU cores and pathways. NVidia also supports Open/CL and Open/GL quite well and always has. The programs such as 3dsMax make use of the CUDA chip for extra processing and use all the pipelines. I say this from experience since I always used ATI cards until I ran into major compatibility issues with Open/GL and Open/CL. The thing with this stuff too is graphics RAM today is not like it was decades ago. The old VRAM was very, very fast, and I think worked faster than the stuff we use today because of how it was made. The problem is the old VRAM was really, really expensive, and more so than regular RAM ever was. This is why the video cards are relatively inexpensive today. Instead of using expensive VRAM, they use regular DDR5 RAM, or the graphics equivalent.

The thing is if you were to do strictly 3d modeling and nothing else, I'd suggest a CAD-type graphics card. These cards are way beyond your budget, and even well beyond mine too. But because you also play games you need the best of both worlds with a decent game card capable of doing graphics stuff. At the moment, I just replaced my GTX680 with a GTX780Ti. I know it's expensive, but I don't replace the hardware unless it's necessary. Sadly, my '680 had problems and will be replaced under warranty. I'll put the replacement away as a spare as I use the new card going forward. But anyway, the 600-series is getting quite old now in a computer part's lifetime terms, and the 700-series cards are now moving down as the 900s are coming out. NVidia just came out with the 900s and skipped over the 800s. Don't ask me why they did it.

If anything, here's an example of what you can get if you watch the sales at New Egg.... I think you have to sign up for the newsletter/advert though like I did.

http://promotions.newegg.com/NEemai...D-_-Weekend&et_cid=11549&et_rid=67996#Weekend

This came in today.

John

Huh, you may be right, after I 'accidentally' brought the identical model from Vegas, the new 8400GS had a HDMI output and a fan. my old 8400GS does not have any of those. But you do have a point that some say it was over 7 yrs old.....or whenever when that was first released.

And thanks for Newegg and amazon, I'm stuck to down 2 cards that are close to afford: the GT610 (http://amzn.to/1ooXMnQ) and the GT640 (http://bit.ly/1u5tLPt). Both are the exact to my needs, but I have to make sure that it runs good and has a good quality and speed and compatible to the games I go with. I can't just buy it and just run it. What if the programs I run are NOT compatible with the card, it'll be a waste.
 
Huh, you may be right, after I 'accidentally' brought the identical model from Vegas, the new 8400GS had a HDMI output and a fan. my old 8400GS does not have any of those. But you do have a point that some say it was over 7 yrs old.....or whenever when that was first released.

And thanks for Newegg and amazon, I'm stuck to down 2 cards that are close to afford: the GT610 (http://amzn.to/1ooXMnQ) and the GT640 (http://bit.ly/1u5tLPt). Both are the exact to my needs, but I have to make sure that it runs good and has a good quality and speed and compatible to the games I go with. I can't just buy it and just run it. What if the programs I run are NOT compatible with the card, it'll be a waste.

The GT640 is a nice card and most likely the better of the two, without peeking at the specs at the moment.

You shouldn't have a problem running anything you are running now. If they run on the 8400s, they should run okay on a GT640. You'd have a bigger issue if you swapped to ATI (AMD) because their drivers do some weird things at times especially with graphics applications.

That's interesting how one version of an identical part name had the HDMI port and fan. Perhaps as time went on the board designer wanted to update the card with the HDMI port, and the fan to cool the GPU. Speaking of HDMI ports, you can get adapters for those to convert from HDMI to standard VGA, I think. If you can get one of those, it will buy you time until you can get a 640, which will be even less expensive as time goes on because the newer models are coming out now.

John
 
Well I think the difference with that, is that it has 1GB VRAM and with a CUDA core 3x than my card (not sure if those are really important for gaming). I think that TS12's CMP might've been a factor that caused my card to go AWOL. One example is that, when I imported the CDPs on the import screen, it created a LOT a memory usage, and I bet some memory might've 'leaked' out of the card, caused the overheat (maybe). Not sure if it's a programming issue, but TS12 I have already have it to it's latest SP1, nothing more. TS10 however (the one I mostly use all the time), It didn't heat up the CPU (or the driver).

I was so lucky I didn't come up with ATi. And yep, with more NEW modern cards coming into the market these days, they've implemented a HDMI and had a VGA port STILL. It's nearly 2015 and yet, they STILL have VGA ports.
 
Well I think the difference with that, is that it has 1GB VRAM and with a CUDA core 3x than my card (not sure if those are really important for gaming). I think that TS12's CMP might've been a factor that caused my card to go AWOL. One example is that, when I imported the CDPs on the import screen, it created a LOT a memory usage, and I bet some memory might've 'leaked' out of the card, caused the overheat (maybe). Not sure if it's a programming issue, but TS12 I have already have it to it's latest SP1, nothing more. TS10 however (the one I mostly use all the time), It didn't heat up the CPU (or the driver).

I was so lucky I didn't come up with ATi. And yep, with more NEW modern cards coming into the market these days, they've implemented a HDMI and had a VGA port STILL. It's nearly 2015 and yet, they STILL have VGA ports.

I doubt anything "leaked" from your video memory it's more likely that the card was pushed hard by graphics intensive applications including Trainz. Remember anything that pushes components to work hard, will cause them to heat up. Heat is the enemy of computer components as it stresses out the internal parts of the chips. With thermal conditions being in the 100s of degrees, yes 64C is quite warm - not quite the boiling point of water, which is 100C, but still warm enough to damage components if they're not cooled off with a fan or some other means. If there is a constant cycle of heating and cooling, the parts will eventually fail internally as the bonds break from their surface inside the chips. Sometimes this happens catastrophically as in an instant failure, while other times it happens over time making the parts fail later on. These late failures are called latent failures meaning they happen long after the initial incident occurred. These are also the most difficult to troubleshoot, and also the most difficult to put in insurance claims on should there have been an incident that caused a failure that was covered under insurance. A good example would be a lost power ground connection on the power lines. I had that happen and while the initial claim was paid by the insurance, equipment that blew up, it did not cover equipment that died much later on due to the same incident as it's difficult to prove the events were related. I could do that, but it's an awful process of documenting the incident and correlating the failure.

Latent failures also show up as temperature sensitivity in components, meaning you over taxed the system one time too many. It may continue to work, however, there are now "triggers" where the system will start to fail. These are also the move awful things to fix because they can work intermittently so you think you fixed the problem only to have that happen or another problem occur later. The thermal condition means that the boards will work fine when they are first powered up, or sometimes not, but will fail later on as they warm up, and or cool down depending upon their sensitivity. As a technician we used cans of Freon to freeze the chips to test for cold sensitivity, or sometimes to freeze a failed warm chip and have it work again. We also used a heat gun to warm the circuit and force failures so it was easier for us to troubleshoot the failing components. God, I loved that job... sitting there with a scope and DVMs mounted to the circuit test unit and board as I chased around signals on the scope using schematics and my brain to work out the logic and failures.

Anyhow, CUDA is a processing chip on your NVidia card and works with the data-processing of video information, and comes on all their chips today. The chip is used by many games and can be enabled or disabled via the NVidia control panel. 3D programs, such as 3dsMax and others make use of the CUDA even more as the chip can push the data through the processor and GPU very quickly compared to just your computer's CPU. This is why the newer video cards are so much faster than they used to be. They are also used by some engineering software to process large amounts of data because they are so much faster than even the fastest CPUs. NVidia has a line of special processor boards that are made up of banks of GPUs. They have no video ports and work with special software to run huge calculations on data for physics and artificial intelligence projects. Pretty cool, I say, but not any good for us to use.

John
 
Back
Top