9800GTX corrupts and crashes 3d gfx


Recommended Posts

Hi

I have a rather serious issue. I just bought an upgrade - 9800GTX by BFG, Asus P5K-E WIFI/AP, Intel Q6600 and 4gb [2gbx2] of corsair ram. The powersupply is an XClio 750W PSU and I'm running Vista 64-bit with SP1

Now the problem is that when I am playing games (tried this on AOE3 and COD4) I am getting corrupted graphics every now and then. Corruptions include glitches as well as for example crates or lands coloured coloured with a red or blue tone. This has also occured when I tried to benchmark the GPU (as well as put some stress on it to see whether this happens in these two games or is common) in 3Dmark 2006 and the same happened. Another thing that occured during the benchmark was that after a few tests execute I get a DirectX error (device failed to initialize).

I have tried different drivers (BFG's own, latest from NVidia as well as the omega modified ones) but the issue is still occuring.

Anyone maybe knows what's causing this and a possible fix? I have also tried to use a different GPU (an 8600GT) but the sympthoms where different. Glitches were still occuring as well driver restarts while playing games.

This evening I will be swapping memory and possibly do a format and change to a 32-bit OS and see if this issue occurs again (maybe its specific to x64 version of nvidia's driver, or some other problem). I just wanted to see whether you guys may have any ideas as how to solve this.

PS: I have already contacted OCuk and I'm ready to RMA the gfx card, so if you think that even though I find a driver issue, the 9800gtx shouldn't be doing that weird corruption stuff, I will send it out and wait for a replacement.

Link to comment
https://www.neowin.net/forum/topic/637050-9800gtx-corrupts-and-crashes-3d-gfx/
Share on other sites

Hi

I have a rather serious issue. I just bought an upgrade - 9800GTX by BFG, Asus P5K-E WIFI/AP, Intel Q6600 and 4gb [2gbx2] of corsair ram. The powersupply is an XClio 750W PSU and I'm running Vista 64-bit with SP1

Now the problem is that when I am playing games (tried this on AOE3 and COD4) I am getting corrupted graphics every now and then. Corruptions include glitches as well as for example crates or lands coloured coloured with a red or blue tone. This has also occured when I tried to benchmark the GPU (as well as put some stress on it to see whether this happens in these two games or is common) in 3Dmark 2006 and the same happened. Another thing that occured during the benchmark was that after a few tests execute I get a DirectX error (device failed to initialize).

I have tried different drivers (BFG's own, latest from NVidia as well as the omega modified ones) but the issue is still occuring.

Anyone maybe knows what's causing this and a possible fix? I have also tried to use a different GPU (an 8600GT) but the sympthoms where different. Glitches were still occuring as well driver restarts while playing games.

This evening I will be swapping memory and possibly do a format and change to a 32-bit OS and see if this issue occurs again (maybe its specific to x64 version of nvidia's driver, or some other problem). I just wanted to see whether you guys may have any ideas as how to solve this.

PS: I have already contacted OCuk and I'm ready to RMA the gfx card, so if you think that even though I find a driver issue, the 9800gtx shouldn't be doing that weird corruption stuff, I will send it out and wait for a replacement.

Download RivaTuner

http://downloads.guru3d.com/RivaTuner-v2.0...wnload-163.html

And have a look at your temperatures, they could be a tad high, what case do you have also?

Download RivaTuner

http://downloads.guru3d.com/RivaTuner-v2.0...wnload-163.html

And have a look at your temperatures, they could be a tad high, what case do you have also?

Case is a LianLi V1100, but one of its side fans is pushing air from the outside directly over the GPU's side-blowing fun (that cools that heatsink). I have checked the temperatures with ntune, and they fluctuate between 56 degrees and 70 degrees (I presume this is normal for a highend card, right?)

Case is a LianLi V1100, but one of its side fans is pushing air from the outside directly over the GPU's side-blowing fun (that cools that heatsink). I have checked the temperatures with ntune, and they fluctuate between 56 degrees and 70 degrees (I presume this is normal for a highend card, right?)

70C is a tad hot for a graphics card at load but this temperature may be normal for a 9800GTX, you could try temporarily underclocking your graphics card slightly and seeing if it helps.

Additionally what is the highest amperage that your PSU supplies on the 12v rail?

70C is a tad hot for a graphics card at load but this temperature may be normal for a 9800GTX, you could try temporarily underclocking your graphics card slightly and seeing if it helps.

Additionally what is the highest amperage that your PSU supplies on the 12v rail?

I will try under clocking it, but still, its sold as an OC graphics card ....

As for the amperage, the PSU has 4 12V rails all rated at 18A.

I will try under clocking it, but still, its sold as an OC graphics card ....

As for the amperage, the PSU has 4 12V rails all rated at 18A.

Does your PSU have a "Turbo" mode of some sort? I.e it can combine the 4 12v rails to give a total amperage of 35A or some value.

If not then you'll need to buy a new PSU, as 18A will NOT properly power a 9800GTX, especially one that is pre-OCed.

(I'm not sure, but i think a 9800GTX requires a total of 30A on the 12v rail to function properly.)

70C is not high at all for a high end graphics cards. These cards are power hungy beasts that are built for high temperatures. My 8800GTXs will reach as high as 80C and that is considered perfectly normal. Do you have another machine you can test the graphics board on to see if it gives the same results? If not, I would consider an RMA on the card as soon as you can.

Btw, I forgot to mention, in some instances and I have noticed this only with the 9800GTX, the corruption is persistent even after the game is closed, thus ending up with several glitches on the desktop as well...

Last time I had this happen was on a 6800Ultra *those was the days ;) * was defective memory circuits on the card, I would RMA it personally.

Last time I had this happen was on a 6800Ultra *those was the days ;) * was defective memory circuits on the card, I would RMA it personally.

Hmm, RMA it is then.

Btw, regarding the PSU with a 30A lane on the +12v, any suggestions on which brand have these? Also, the GFX takes two inputs for extra juice, don't they help with power?

Yeah I'd always use all power inputs on the GPU mate. For example I run now a 8800GTS Alpha DOG XXX from XFX I've OC'd it even higher and I run it from a Seasonic S12 600 it has 2x12v rails each with 30ampre on them.

All of it runs stable .

For recommendations I would say try the Corsair HX620 or perhaps if you really want the juice the new 1000 version.

Yeah I'd always use all power inputs on the GPU mate. For example I run now a 8800GTS Alpha DOG XXX from XFX I've OC'd it even higher and I run it from a Seasonic S12 600 it has 2x12v rails each with 30ampre on them.

All of it runs stable .

For recommendations I would say try the Corsair HX620 or perhaps if you really want the juice the new 1000 version.

Yeah I did connect them as well, w/o them it does not switch on. Either way, I am probably RMA the gpu eitherway and gonna get this PSU: http://www.hipergroup.com/English/products/hpu-4m880.html, it has 30A on lane 3 and 4. I guess it should be enough right?

Also, if it was a power issue, wouldn't the 8600GT work fine, or maybe the problem is caused by a mixture of hardware as well as driver issues?

70c is pretty good for that card, I have a 8800GTS G92 OC based off the same G92 core and it idles at 74c lol. Partially because I like the silence factor (All my case fans are Noctua units running at 800rpm unless required, then it spins up to 1200rpm-1300rpm under full load motherboard controlled) and the 8800GTS' fan is at 480rpm. Generally speaking, these things can handle 95c no probs at all. No worries about that temp.

Okay, small update. I bought the Tagan PSU and after letting 3D mark running for an hour or so at full settings as well as playing some games, the corruptions did not come up. The only thing that occured was that twice the display driver was reset causing 3D mark to stop the tests. Other than that, i have noticed that while playing, the games run smoother (this was mostly noticable when in AOE3, the main menu felt sluggish. Now the main menu runs very smooth).

I am gonna clean up the graphics drivers and install the latest WHQL drivers from nvidia and see how it goes :)

Thanks alot...

Help me please, i also have problems with my Nvidia 9800GTX. Whenever i play Crysis and some other online games, there come a time when the game freezes and i have to go to the task manager, then switch again to the game so that it can continue. failing to do so will crash and reboot my system. what my real concern is to know if my PSU (Corsair 520W) is sufficient enough to power up the Video Card. If yes, then what could be the problem/s causing the inconsistencies? thanks in advance!!!

well buddy, I can tell you that the issues I had have re-appeared after I formatted and installed Vista 32-bit (in hope of MAYBE solving the random driver restarts I was getting when I was running x64 vista). I am half way fearing that it is software related or maybe some hardware incompatibility....

I have searched a little bit around the internet and found some other forums with users having issues similar to mine and they were blaming SP1 (I have not run my system without SP1 since I am using a slip-streamed DVD) ...

[Edit: attached a screenie of the resulting "corruption"]

post-12667-1211394439_thumb.jpg

Edited by g0wg

Aww mess... sorry... but let me share another experience with 9800GTX...

The VC i have right now is the replacement of my previous one (same specs) but had worse problem (somewhat close to your problem but still worse). The older one displays white horizontal lines in checkered formation whenever i install any of its available drivers...

I dont know the real cause of the problem but i guess it lacks power, the light inside the VC should be green right? my first one lit in red.

So what i did is i RMA'ed it. I think you should do the same.

I will post my version of "inconsistencies" with my VC whenever i get a chance to picture it.

(Re-waking up my thread)

Is there some issue with the nvidia 8 and 9 series gfx, their drivers and vista sp1? I mean, I replaced the GPU with a temporary 8series one and now I am getting "nvlddmkm not responding" warnings (followed by a driver restart).

I have done a search for the error on goog, and I am finding occurences of this problem everywhere and going back more then a year, yet, can't find a viable solution. I am now blaming the motherboard for these issues (i don't think the processor can cause such issues, or it can?), maybe incorrect voltages or something ...

  • 4 months later...

In all seriousness - swap to ATI. nVidia is... faltering. I used them for the longest time, and in single card setups they are slightly better still... but ATI has it made with Crossfire. Where as nVidia's SLI gives you 2 x 1 cards total capacity = 1.5x increase, ATI's Crossfire gives you 2x 1 cards capacity = 1.9x - 2.2x increase in capacity. This is because in SLI the cards "talk" via the FSB if I'm not mistaken. In Crossfire, they use the little crossover connector thing... forget what it's called.

In all seriousness - swap to ATI. nVidia is... faltering. I used them for the longest time, and in single card setups they are slightly better still... but ATI has it made with Crossfire. Where as nVidia's SLI gives you 2 x 1 cards total capacity = 1.5x increase, ATI's Crossfire gives you 2x 1 cards capacity = 1.9x - 2.2x increase in capacity. This is because in SLI the cards "talk" via the FSB if I'm not mistaken. In Crossfire, they use the little crossover connector thing... forget what it's called.

In all seriousness, no, nVidia isn't faltering.

  • 2 weeks later...
Hi

I have a rather serious issue. I just bought an upgrade - 9800GTX by BFG, Asus P5K-E WIFI/AP, Intel Q6600 and 4gb [2gbx2] of corsair ram. The powersupply is an XClio 750W PSU and I'm running Vista 64-bit with SP1

Now the problem is that when I am playing games (tried this on AOE3 and COD4) I am getting corrupted graphics every now and then. Corruptions include glitches as well as for example crates or lands coloured coloured with a red or blue tone. This has also occured when I tried to benchmark the GPU (as well as put some stress on it to see whether this happens in these two games or is common) in 3Dmark 2006 and the same happened. Another thing that occured during the benchmark was that after a few tests execute I get a DirectX error (device failed to initialize).

I have tried different drivers (BFG's own, latest from NVidia as well as the omega modified ones) but the issue is still occuring.

Anyone maybe knows what's causing this and a possible fix? I have also tried to use a different GPU (an 8600GT) but the sympthoms where different. Glitches were still occuring as well driver restarts while playing games.

This evening I will be swapping memory and possibly do a format and change to a 32-bit OS and see if this issue occurs again (maybe its specific to x64 version of nvidia's driver, or some other problem). I just wanted to see whether you guys may have any ideas as how to solve this.

PS: I have already contacted OCuk and I'm ready to RMA the gfx card, so if you think that even though I find a driver issue, the 9800gtx shouldn't be doing that weird corruption stuff, I will send it out and wait for a replacement.

hi

i am having the same problem and i need to know if you fixed it or not

thanx

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.