Official GIGABYTE Forum
Questions about GIGABYTE products => Graphic cards => Topic started by: Strahd on August 05, 2010, 11:52:16 am
-
Greetings,
it was about a week ago that I bought a new computer and I've been plagued with the "nvlddmkm stopped responding and has recovered" problem.
My system specs are as follows:
i7 930
Giel Value RAM 6x 2gb
Gigabyte GeForce GTX 470
Ga-X58A-UD3R
LC ARKANGEL 850W PSU
WD 500gb Black
WD 1500gb Green
Plextor BluRay Burner
24" Samsung screen (DVI) & 32" Samsung LCD TV (HDMI)
All of that beeing run by Windows 7 Pro 64bit
I have tried everything I can think of but I keep getting the "nvlddmkm stopped responding and has recovered" problem. When it happens twice, usually inside an hour, the machine reboots. I've ran every test i know of (memtest, prime95) and everything goes through but for the 3D Mark Vantage or any 3D app/game. So that led me to conclude that there's a problem with the graphics card so I tried all of the different drivers on Nvidia site for it as well as those on the windows update and the ones from the DVD that came with the card. So far those on windows update seem to be the most stable ones, there are weird things still happening such as screens going blank, crashing of apps but at least the machine does not reboot. Mind you every driver replacement has been done after a full format and windows reinstallation (I've never really trusted the uninstall proggys when it came to drivers). I've also been monitoring the temperature of the graphics board with SpeedFan and it does not go above 88 degrees cel.
Is there anything else I can do or try safe for the board replacement (I was waiting for a friend to bring a spare so I can try a different one)?
-
Well I think you've hit the nail on the head with your diagnosis but without a better driver there's not a lot you can do. Possibly the card is at fault but it is more likely the software. You did the correct thing by formatting each time( most of us wouldn't have bothered) because at least you know the installation is clean. The only thing I could recommend is to try clearing the CMOS first and then loading the BIOS optimised defaults before doing the format and reinstall. I really don't think it would make a difference but when you have tried everything else.
-
You don't say which BIOS version you are using but sometimes a newer version might address this type of issue.
If you are running the latest BIOS for your board, it might be worthwhile contacting Gigabyte Support at: http://www.gigabyte.com/support-downloads/technical-support.aspx?ck=2 just to see if they are aware of a problem with this card/mobo combination.
-
@Dark Mantis well I figured, since I have a new system and clean install anyway, it's prolly the same (time wise) trying to remove the drivers or reinstalling the windows and I did clear the CMOS but only the first reinstall.
@absic115
If I'm not mistaken I do have the latest BIOS on the mobo (not counting any beta versions) but will recheck when I get home. I have no idea about the BIOS on the Graphics Card though.
It's the weirdest and most persistent problem I had in the 14 years of assembling the blasted machines together...
-
I would definitley check on the firmware for the card. There have been a few cases recently where the firmware needed updating to get the cards to run properly.
-
One of the main reasons I switched from Nvidia to ATI were the problems I was encountering from Nvidia drivers and they were very similar to what you are describing now and I never did manage to resolve them! ?? :P
-
OK. Motherboard BIOS has been updated to the Fb11 Beta, I was on the only other BIOS FA.
I have no idea where to get new/different BIOS for the graphics card though.
This be the its proper ID GV-N470D5-13I-B with the BIOS version 70.00.21.00.03
Going to test now if the new mobo bios did anything.
EDIT: Nope no good. Problem remains!
-
OK so the problem still persists even after replacing the graphics card. As I couldn't get a different card to test this I RMAed the 470 and got a GT 240 as a replacement while I wait for the other one to come back.
After another windows reinstall and what not I still have the same problem. Oh and the new card is Gainward not Gigabyte.
So after all this I started leaning forward to a broken motherboard.
It's interesting though everything works, every single burn-in test I can think of went through no problem until I try 3D.
Any additional thoughts on this from anyone? Could it be a faulty PSU maybe?
-
Personally I dont like the PSU very much for several reasons the main one being that the four 12volt rails are only rated at 18amps each to start with and for a 850w PSU that's not great. One 12v rail performs better with most new technology and so it might be worth swapping it out. See if you can just borrow one to test it first.
-
Well I tried 2 different PSUs and still nothing. I'm picking up a new motherboard today. Will keep you posted on how that goes.
-
Yes but without meaning any offence what quality were the PSUs that you swapped out? I mean it is no good changing the one you had for another inferior make or model. Anyway we will look forward to hearing how you get on with the new bits.
-
One was also LC Power the other one was Enermax LIBERTY ECO II 720W
-
One was also LC Power the other one was Enermax LIBERTY ECO II 720W
I am not familiar with the LC Power series but the Enermax even though it is a quality PSU still has the output split into three 12v rails which I personally don't find as stable.
-
Ok then recommend me one and I'll see if I can get my hands on it :)
-
Try and test your system with one of the bigger Corsair HX series say 850 or 1000.
-
Well looks like I sorted it out. But let me fill you in on everything :)
After replacing the PSU and further testing with another 2 PSUs I now switched back to the original LCs Arkangel as it was obvious it was not the PSU (was unable to get the Corsair to test though). After the PSU fiasco I opted for motherboard replacement and got the GA-X58A-UD7 rev. 1 and akasa X2 CPU cooler in hopes of decreasing CPU temps while I'm at it.
As the CPU temps did go down a degree or 2 the original problem still remained. I would still get the nvidia driver failure. At this point I also have my original GTX 470 back "on" the system (can't really say in the system as the entire rig is sitting on my desk right now :P).
At this point I was at a total loss but figured I should try this anyway. I removed half of my RAM even despite the memtest saying everything was fine. And right now I'm in my fourth day of Ultima Online, Starcraft II and 3D Mark tests without a single crash. Previously UOL alone would produce a crash within 5 minutes.
I did note that RAM is now significantly cooler (do not know the exact degrees, only the touch of my finger sensors :P)
I also noticed the default settings for RAM in BIOS are 8 8 8 20 while RAM is clocked at 9 9 9 24 (this goes for both motherboards!). However there is no difference if I change those values manually. I'll mess with the ram a bit more and see if I can figure something out.
On a side note my roommate bought the exact same configuration (ram, mobo, proc.) at the exact same time I did and he has 0 problems with his rig.
-
Well I am glad that you seemed to have found the answer. It is sometoimes a long, hard slog but it is worth it in the end. As a matter of interest what GPU was you roommate running?
-
He is running GTX 260, Gainward variant I think
-
Less power hungry but apart from that nothing special. Have you tried swapping the memory out to see if the other set works ok?
-
No not yet. He's on a work related project at the moment and can't spare the computer. But it is the next planned step ASAP.
-
What about trying the other set on your computer and see if they work?
-
Right sorry misunderstood you there and no not yet. I'm just backing up something so I can't do it right now, but I should be able to try it in about an hour.
-
I think it would be worth knowing if the problem was faulty memory module/s or just that the full slots were the cause.
-
After about n hour and a half of UOL everything still works as it should. Will report back in a day or so after some more benchmarking tests.
-
Thanks will look forward to it. ;)
-
OK After extensive benchmarkings and burunin testst and what-not I can not seem to crash the driver/system no matter what configuration of the RAM chips I put in. If I have not made a mistake I have, since the last post, tested every possible combination of the 6 2gb chips with 3 at the time.
Yesterday I've put all 6 chips back in and have disabled various turbos and power saving modes in motherboard BIOS. so far everything works fine. Will check now over the weekend if i can crash it.
-
Hopefully without the power saving modes enabled it will be more stable. I will keep my fingers crossed ;)
-
Sorry it took so long to reply but i got sick in the mean time and were not able to do anything until last week.
So after removing 3 of the 2gb ram sticks everything now seems to be working fine and stable. So I guess I'll just use 6gb and be done with this.
Thanks for your help.
-
Sorry to hear that you were ill but hopefully are feeling better now. It seems to be the solution to your problem and now you can enjkoy usiong your computer rather than rebuilding it every five minutes ;)
-
Well some of us enjoy rebuilding them too ;)
Thanks Mantis for your help :)
-
Well me too actually. I find the building and hardware side more interesting than actually using them. I guess that't why I keep customising mine. Absic is the same, he's always pulling it apart to test something or other and then rebuilding. Part of the fun.
-
Hey, I never said it was fun to keep pulling things out and putting them back! :o
But, I have to admit that I do enjoy it a little! ;)
-
I think a lot of it is just the satisfaction of building something out of a lot of different parts and getting it to work successfully(hopefully). I also enjoy specifying the parts first.I can spend my budget ten times over and not part with a penny. 8)