BleepingComputer.com: Geforce 9800GTX nvlddmkm.sys error (PLEASE HELP!)

I have a 9800GTX and Vista for a while now, and I have always had problems with it. In the start it was nvlddmkm.sys error, complete with miscolored pixels and brown lines popping up everywhere. Then it was fine for a long while, until I started playing Far cry 2, and got the nvlddmkm.sys error and artifacts in games. Well, I could still play some games fine, so I shut down my computer and went to sleep. The next day I got the error on start up, and the miscolored pixels back again. Also, the screen blinks black a few times before I get the message.
WHAT IS WRONG WITH MY COMPUTER? It's driving me insane, it's a high end computer, so I spent A LOT of money on it, and it doesn't even work without error?
I have thought of two things:
1. Changing to xp (I have the cd)
2. Getting a more powerful psu (My PSU may be too weak, 550W PSU for a 9800GTX? Enough?)

Before doing anything, I need to know if it will work. A PSU is expensive, and I'm not using my hard earned money on it unless I now it will clear the problem. I have tried the latest drivers, yes. I uninstalled the old with Drive cleaner too. Tried older drivers, just got worse. Tried staring at my computer screen, didn'
t work.
My temperatures on the card are 37-38 idle, not under stress.
I am willing to do almost everything to get this unstable piece of junk to run again.

This post has been edited by Auroron: 08 November 2008 - 06:11 AM

550-watts should be plenty for a 9800GTX, but perhaps it is failing. Do you know how many amps are on the +12V rails? If you don't know, you can check by reading the label on the power supply. If there are two or more +12V rails, just add up there total amounts of amps labeled as "a". I think the recommended for a 9800GTX is at least 25-26.

If the power supply seems to be fine, and has enough amps, then maybe it's your ram. Try running memtest86 . You can burn the iso image to a disc and then boot your computer off of it. It's good to let it run for a few hours. This will test to see if your ram has any errors. I remember having constant computer lock ups while gaming and it was because one of my sticks of ram was bad.

I took a look around google on this issue, and I really can't find a sure answer. Every website I went to someone had different solutions and said the others didn't work. It may be more of a software issue though.

i'm too lazy right now to open by cabinet, but speedfan tells me that the +12v is 11,97v right now. I have no idea what that means. I also have a -12v, which is on 16,97v.
I have heard from different people that 550W is not enough, so I am confused here, to say at least. This problem needs to be sorted out quick :S'
I have also heard that it's a vista comptability issue, and the problem occurs very, very rarely on XP, so switching to XP may be last resort. Any idea how I do it? A guide or something?


Mod Edit: Edited to remove unnecessary quote. ~tg

This post has been edited by tg1911: 08 November 2008 - 09:33 PM

I'm running a 9800GTX in Vista 64-bit and I have no issues. Or to be more specific, I'm having no nvddmkm.sys issues. Problems related to that file are often driver-related, anytime a game or app crashes in a way that affects the video card, that file is often pointed to. Sometimes, the driver crashes, causing the screen to go black, and then recovers. For some reason, anytime that happens with me it affects stability causing more nviddmkm crashes until a new driver comes out. Most 9800GTX's require a 475 to 500 watt power supply, although I have seen card manufacturer XFX suggest a 630 watt power supply or higher for non-SLI use. If you want a power supply suggestion, I would get the CORSAIR CMPSU-650TX 650W, it's at a good price, gets excellent reviews, and comes with mail-in rebate.

This post has been edited by DJBPace07: 08 November 2008 - 08:36 PM

So, I think I will get a new PSU. But a question, raising fan speeds, will it increase PSU usage? Because of heat problems (My two front fans are not working, so I had heat problems), I run most of my fans at around 80-100%. Noisy, yes, but at least it's not overheating.
A corsair 750W will cost me around 160 USD, the 650w will cost me around 141 USD, so I might take one of those. Also, on the 13th, the new World of Warcraft expansion comes out, so I need to have it fixed by then

This post has been edited by Auroron: 09 November 2008 - 11:02 AM

Is there something preventing replacement of the fans which don't function?

It's not like doing so is difficult or expensive...I just superglue my fans to the appropriate positions.

Considering the price versus potential for trouble...I don't understand any hesitation.

Louis

The Corsair 660W costs $99 USD, the 750W is $120 USD. However, actual price varies based on where you live. It doesn't sound like the GPU's are getting too hot, 37 degrees is actually quite good for a 9800GTX. A PC needs to have a fan running on the front to force air into the case and one on the back to force it out. Heat problems can occur if the front or rear fans stop working, it disrupts normal air flow. Raising fan speeds should not cause a significant drain on the PSU. Replacing the fans would be a more long-term and beneficial option.

This post has been edited by DJBPace07: 09 November 2008 - 06:11 PM

DJBPace07, on Nov 9 2008, 06:10 PM, said:

Yes, I live in Norway, and an online shop sells it for less.
Also, i'll take a look at the fans. If I get them working, I can probably decrease my fan speeds a little too.
So no ideas what is causing my error? Is it the psu? The problem came back when I played a very graphic intensive game. I had it in window mode for a while, and the error came back.
And also, the shop that I got my computer from (They built it for me, but I chose the parts. Saves me time) have a 3 year warranty rule, so I can RMA any part of my computer. You think it's maybe a faulty card? It didn't take long after I got my computer that I got the error (While watching videos on the internet, so no gaming o.O)
The support line for the shop also tells me it's probably a faulty GPU, and tells me to send in the card. I know thats probably a standard answer, but it might be?

Thanks for all the help
Auroron

If the builder is willing to RMA the card at no charge to you, even in shipping and handling, you can try RMAing the card. You have nothing to lose except time. If you have a spare GPU, you could use that while you wait for your 9800GTX to be sent back to you. Hardware issues can be difficult to diagnose, a manufacturing defect may manifest itself in a number of ways. The problem is that many hardware problems can often appear as software issues. I had a PC that would freeze, every time, when I would play a certain type of game or CPU intensive application. The motherboard was replaced, as was the graphics card, and still the problem remained. I think the CPU had something wrong with it, but the point is the freeze could have been caused by software too. Only through elimination was I able to discover that it was probably hardware related. The nvlddmkm.sys error is often caused by a driver crash, but if the GPU has a problem with its hardware, it too could throw a nviddmkm.sys error.

Hi There,

I've the exact same problem, so I was wondering, any news here?
This are my specs;

(CPU1) Intel® Core™2 Quad CPU Q9550 @ 2.83GHz @ 2833MHz
(MOBO) ASUSTeK Computer INC. P5Q-PRO
(RAM) 8GB
(VGA1) NVIDIA GeForce 9800 GTX+ DK , 1680x1050x32, 59Hz
(OS) Microsoft® Windows Vista™ Ultimate (SP1) 64-bits

Monitor: Samsung SyncMaster T220 /22"-2ms
RAM Specs: Corsair Dominator TWIN2X8500C5D 4096MB, DDR2,2x2GB(KIT),DHX,E.P.P, CL5-5-5-15-2T --2x
PSU Specs: Corsair Powersupply 650W Black, ATX/EPS, 120mm Fan, 8xSATA, SLI
CPU Cooler: Cooler Master Eclipse CPU-Cooler, Socket 775/754/939/940/AM2/F
Primary HDD: Samsung SpinPoint F1 1TerraByte SATA2 32MB 7200RPM

I've tried almost everything, from performance test to ASUS smartdoctor, AI Suite, etc. Temperatures always good (GPU around 55, CPUs around 42), no problems with stress tests etc. Drivers from the CDs, drivers from the web, drivers through windows update..

I've got the 'pixel' error also, or sometimes the game just goes to windows, I see the popup in the systray with 'Monitordriver restored' or something (the nvlddmkm error) and it continues, or indeed the pixels just go wild and a restart is required to solve it.

The 9800gtx+ is the overclocked version, so I didnt do anything myself. Tried running on Windows 7 (the public beta) but the same error occurs (pixels).

Screenshot shows it

Thanks in advance guys

Welcome to bleepingcomputer! In the future, feel free to start your own topic, and then you could refer back to an older one.

First off, only use the drivers from the nvidia website. Uninstall all other graphics drivers that you have installed, then install the latest from the nvidia website.

Next, try running memtest. http://memtest86.com/ . This will scan your ram for any errors. You need to burn it as an iso image to a disc, and then boot off of it (make sure your boot order is set to cd-rom before anything else). It's best to let it run over night.

Thanks sterling14, I will do next time, didn't want to start spamming a new (duplicate content) topic as my first post.
Alright, I've installed the latest nvidia drivers, and burned the .ISO from memtest and let it run last night.

Had to use version 3.4 since the 3.5 multiple-CPU expermental didnt work, SPM timeout or something.. Anyway v3.4 gave me the message 'All passes complete, no errors found, press esc to exit'. In the tables above there where only zeros under errors so I guess the test worked fine.

I've send a email to my supplier, and its in limited warranty. So if I send the card back, and it isnt defect, it costs me around 50€ (at start). I thought the card was almost to expensive for its kind already lol, so want to make sure it is a faulty card and not my setup or anything before I send it.. Thanks again!

Well, I'm out of ideas. Too bad you couldn't try a different video card in it, just to be sure. That fee of 50€ sounds pretty harsh if it's not broken.

These are the only things I can really think of (in order from most likely to least likely):
1.Video Card
2. Power Supply
3. Motherboard.

This post has been edited by Sterling14: 04 March 2009 - 03:43 PM

Yeah I dont have a additional PCI-E card, other cards are all AGP so can't test that.
All the hardware is purchased at the same time, november 2008, so its a bit harsh..

I've send a email to Asus, they told me to ship it to the supplier because it was most likely a false GPU or a error in the RAMDAC.. I think I/we have tried to get enough info about the problem and that RMA'ing it is the only solution.. Will ship it out 2morrow, and will let you know how it went, so perhaps we can close this topic.