Did I get a bunk GT????

[M Diddy]

Ok guys. I got my EVGA 8800GT SC 3 days ago. I was able to bag a 14220 in 3DM06 with it clocked to 730/1800/1000. With the fan set to 60% in Rivatuner, temps would get no higher then 66C. 3DM06 would go through with no problems or errors. I then began playing some games, Crysis Demo, HL2: Deathmatch, etc. and started to get some driver errors every few minutes. Game would play fine and then all of a sudden stutter and stop. I'd then get the "Display driver nvlddmkm stopped responding and has successfully recovered." error. So, I decided to clock down a bit.

I first took it down to SSC speeds (700/1728/1000) thinking the card should be able to pull these off fine. I load the settings, play a few minutes of a game, and the error was right back. Doing a little research, I came upon JohnnyGuru's post on HardForums about the RAM clocks on these cards so I decided to take the RAM down to 975 Mhz. No dice. Took it down to 950, and still got errors. Keep in mind 950 is the stock speed for the SC. I started to drop the core clock down, and even with it at 675Mhz and the shader at 1680, I was still getting the driver errors.

Just to see, I then took the card back to stock speeds. After playing a few hours, I did not get one hangup or error. Ok, so the card will run at stock but after seeing all the OC's people are getting both on here, EVGA forums, and XS, this card should be capable to at least game at SSC speeds without errors. Thing is, I'm not sure if it's a problem with the card or the simple fact that the drivers are still so young for this thing. With my old EVGA 8800GTS 640 SC I had no issues at all in 4 months of use while OC'd so I'm thinking drivers aren't the issue here.

Any help or insight you guys could give me would be much appreciated. Thanks guys!

 

[aja]

If your motherboard has a BIOS option to increase slot power, then do so

That has helped me in the past

 

[DumpALump]

Could be either or both.

I'm also thinking what might have happened is they were testing for stability at KO or SSC speeds and it was not stable. However, it was then stable at SC speeds so they just sold it as so.

Might have also already fried the ram by running it so high? One guy said he killed his card within 24 hours of owning the card. He had a SSC.

Its not exactly Bunk as it is only guaranteed up to SC speeds. Anything after that is just luck of the draw. Bunk would be it wouldn't be stable at stock speeds.

 

[aja]

Bunk would be it wouldn't be stable at stock speeds without any previous attempts at overclocking.

 

[SuperDave1685]

2 x 1GB G Skill HZ PC6400 1135 Mhz @ 5-5-5-12 2.25V

I think therein lies your problem... Are you actively cooling those RAM sticks? For that speed, I have a strong suspicion that 2.25v isn't cutting it...At 1135. you're really pushing those sticks hard without a lot of volts.... I'd back your RAM down first to stock 400Mhz and then try to overclock your card again and test it then...

Also, I take it you're using Winblows Vista.... Head on over to the "xxxxxx.sys has stopped responding thread" and read over what some people have done to get that error to go away. It as to go with Driver communication timeouts if I recall... Are yo using 32-bit or 64-bit version of Vista?

 

[Hipcrostino]

I'd then get the "Display driver nvlddmkm stopped responding and has successfully recovered." error.

Nothing to do with the graphics card mate. Its a nvidia driver issues in vista.

see http://www.ocforums.com/showthread.php?t=518749

 

[xilix]

I think therein lies your problem...

Ya think? That's way too high dude. You're just getting greedy with that ram. Back it off a bit and take out all the variables before you decide the card is crapped out. The rule of thumb is that if you suspect a part to be defective, that you put your clocks back to stock and then test that part while the rest of the system is at stock speeds. That way you know there aren't any variables messing with your results.

 

[aja]

For that speed, I have a strong suspicion that 2.25v isn't cutting it...At 1135. you're really pushing those sticks hard without a lot of volts.... I'd back your RAM down first to stock 400Mhz and then try to overclock your card again and test it then...

RAM that runs stock at 1200MHz (or thereabouts) typically runs at 2.35V

 

[M Diddy]

Guys, I'm not thinking it's the RAM for two reasons...

1. I never had the issue pop up at all when running my GTS 640 for the past 4 months OC'd. I don't think my RAM OC would just go bad the day I install my 8800GT.

2. I've tested the RAM at this speed for over 8 hours running Memtest86+ overnight and it came up with no errors. Now, I know it's not the end-all, be-all definition of stable, but that, coupled with the fact I've tested the CPU OC using Prime95 plus the fact I've never had one Blue screen in months of using the PC, I don't think the RAM is the culprit here.

I've done the remove driver, run DC Plat, reinstall thing. I've also followed the steps Jen listed out in the Sticky here. None of these things seems to help.

As far as Overclocking the card, there should be no reason for me not to be able to hit SSC speeds. I believe the card should at least be able to do that. Hell, while I know OCing is grounds for a broken card not be covered under warrenty, I specifically went with the EVGA because they encourage overclocking the cards and cover any failures even if you do.

 

[M Diddy]

And Dave, I am running Vista 32-Bit BTW...

 

[aja]

I wonder if they have started to bin these cards then. Maybe they have been doing it from the beginning

 

[M Diddy]

I wonder if they have started to bin these cards then. Maybe they have been doing it from the beginning

I had asked that before buying my card and the majority of people on here stated there was no binning going on. I really think there might be though given the problems many people are having when OCing these things. Just check the EVGA forums and you'll see I'm far from alone.

 

[thebearot]

Somehow I ran into something that is very similar to Mr.Diddy as well.
Just got my eVGA 8800gt SC a couple days ago and I was busy oc'ing my q6600. (I got a feeling that a load of people upgarded their system in nov. with the same setup)

The game that I was playing with is Jericho. First I noticed the "display not responding" problem with nvlddmkm.sys. I simply restarted the system when the monitor goes into powersaving mode since I've had this problem with my previous system (6600gt) as well.

Now to the overclocks. I first ran this card at 702/1002 and system froze about 30+minutes of playing. (Noticed the artifects) Then I started to lower the core/ram speed and eventually to the stock level but problem STILL persisted. Then I had to goto sleep.

Could it be the oc'ed system that caused the problem instead?

This is quite interesting as we almost had the exact same problem.
Gonna bookmark this page and follow the fixes from the pros.

 

[deathman20]

I'll bring up what was said once again since it seemed to be over looked...

Nothing to do with the graphics card mate. Its a nvidia driver issues in vista.

see http://www.ocforums.com/showthread.php?t=518749

I have the same issue as well it happens after multiple installs of nVidia drivers on the OS. At least for me it doesn't cause a crash but slight pause in the game for a few seconds then back to normal throught the rest of the time the system is up. At least thats been the case for me. That error is a drive problem, surely it might mean the card is clocked too high but at least in the case of this GT card most likely its just an issue with the driver itself.

 

[M Diddy]

Well guys, here's a possible fix. I installed the newest DirectX Update and was able to play well over an hour tonight with no hangups or crashes whatsoever. Don't know if that's the fix for me, but I was able to OC the card a little further and play that way with no problems. Worth a shot to anyone having issues with their card.

 

[Maverick0984]

Test your OC with ATI Tool and scan for artifacts. If it doesn't last at least 10 minutes, you need to lower your OC. Only touch your OC in Rivatuner though. You just use ATI Tool to test the OC, even though it has the ability to change it anyway. (Don't want conflicts)

3dMark06 isn't nearly enough load/stress to fully test it.

 

[Agallion]

thats a vista driver issue. Have had a few clients with it, as far as i know theres a few "solutions" but none are a sure shot.

 

[aja]

Test your OC with ATI Tool and scan for artifacts. If it doesn't last at least 10 minutes, you need to lower your OC. Only touch your OC in Rivatuner though. You just use ATI Tool to test the OC, even though it has the ability to change it anyway. (Don't want conflicts)

I thought ATI tool wasn't all that great

 

[tommcd64]

What driver version are you using?

 

[Hipcrostino]

I'll bring up what was said once again since it seemed to be over looked...

Thanks for trying deathman.