nvlddmkm.sys BSOD issue Windows 7 64 bit

[Pentium4]

I built a gaming computer for my friend in September and it started blue screening on December. He lives a long ways away but I am visiting the area for a few days and would like to get this issue fixed before I leave.

Everything was working fun and when he was playing a game, the computer blue screened, and would not boot into regular mode but boots into safe mode no problem. Specs:

Intel DG31PR with all Japanese capacitors
Core 2 Quad Q6600 @ stock 2.40GHz
4GB DDR2 @ 800MHz
EVGA GTX 570 with fresh thermal paste
Seagate 1TB hard drive
Recapped Allied AL-D500EXP

Nothing is overheating. I don't think it's the PSU. Before deploying the rig, I had it on my kill-a-watt meter and the machine never pulled more than 220W DC from the wall. It uses a 40A rectifier on the 12V and when under that load the 12V rail always stayed above 12.15V

I've tried three different versions of nvidia drivers. Tried unchecking everything during install except the video driver itself, and check "perform clean installation" I've done an sfc /scannow, chkdsk, turned off TDR, turned off UAC, unchecked all startup items. Still blue screens after "Staring Windows" always with nvlddmkm.sys as the culprit.

Has anyone experienced this before? This issue is very frustrating and I'd love any advice or ideas

[pfrcom]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Has anyone experienced this before? This issue is very frustrating and I'd love any advice or ideas

Yes, I've experienced it with W7 32-bit, trying to use a Leadtek PX8400 GS TDH Extreme

Card's fine with XP, but won't work with anything later, including Vista

I asked Leadtek support if it could be a firmware issue - to my surprise they responded - said firmware was latest, card is OK for W7, and mine must have a hardware issue

I wasted hours finding & trying various drivers, all to no avail

Problem just went away when I substituted a different 8400GS card of another make

All I can suggest is try a different graphics card

[stj]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

try a linux live-disk first - just to see what happens.
or plug in a harddrive with a different win version on it.

[Per Hansson]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Yea, if the card was working fine in the system and this started happening out of the blue it sounds like hardware failure.
Try another VGA card.

[kaboom]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

I had something similar with my main XP PC about two weeks ago.

First was loss of wifi after a long defrag. Thought bad RAM but passed memtest. Backed up/reinstalled- but wifi still acted up. Actually, while in device manager, I tried the wifi drivers five times, each with a different error. I pulled the card and examined it- didn't like how the fingers looked so I cleaned them. Finally! So I thought..

Just after I finished restoring everything back over, wlan did the same as before.

While not leading to BSODs, I had "high CPU" usage w/o any processes claiming such, and lousy wireless network performance. Acted like a bad router, but other comps connected to same router and out to internet.

Usually, any browser would randomly show, in its status bar, "downloading proxy script." Funny, no such script existed, and this started suddenly. Anytime the "script" business showed up was when network activity ceased and "high CPU" showed. Again, no process claimed the load, but each core (C2D E8400) got "loaded" to 25%.

Could usually ping and not much else after this point.

Sometimes, it was fine, but within 15 minutes to an hour, network activity ceased, always after that "proxy script" thing.

I got tired of this, took a "software break" and cleaned the case- oddly, the wifi card started working. And quit again. Tried resetting TCP/IP stack and winsock several times, with the fix holding for shorter periods each time. Did a cursory check for malware; didn't expect or find any; remember nothing changed before all this started, except maybe a "heatsoak" during that long defrag.

Since pulling/cleaning the card made a difference for the better, I repeated this. After several cleanings and reseats/removes, it's been OK. This was a real PITA... almost thought BGAs were going bad somewhere. It appears that noise on the PCI bus (from wifi card's oxidised lands) was literally interfering with who-knows-what.

For the last several days, it's been fine. Before the final repair, it certainly would've quit again, many times.

Initially, I pulled cleaned the RAM, video and capture cards, too, when I first had trouble. But it all came down to that stupid wifi card.

What fun...


Pentium4, pull that card and clean the fingers with CRC 2-26 or Deoxit. Spray some on a low-lint paper towel, and clean that troublemaker! If the driver, via the chipset/memory controller/NB, can't talk to the card (open contacts, noise), or hits the wrong address on the card because of an open line or noise-induced "bit flipping," I'd expect nothing less than a crash-n-burn. I really think post-RoHS plating is grossly inferior to the old stuff.


Don't condem your card just yet- I almost smashed my wifi card out of frustration, but things turned out for the better.

Try another VGA card.

That would often work, but it could mask the problem- removing/reseating any card would tend to make it "play nice," even if putting the original back. It was like dealing with a cranky NES with those stupid, cheap contacts. Most PCs are not in a corrosive environment, yet the "goldfingers" have really been acting up. Had two other systems recently where bad RAM connections gave trouble. Yet again, 2-26 saved the day...

We'll take the easy ones when we can get 'em!

[keeney123]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Sounds like malware to me.

[kaboom]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Keeney, if you're replying to me, scanned/uploaded running system procs and many installers/.exes to virustotal; all clean, and no problems after cleaning that card.

Neither Hitman nor avast (old avast, before the sellout) never found anything either.


Dirty contacts and bad connections do happen. I seriously doubt cleaning those lands would've improved a malware situation.

[Pentium4]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Yes, I've experienced it with W7 32-bit, trying to use a Leadtek PX8400 GS TDH Extreme

Card's fine with XP, but won't work with anything later, including Vista

I asked Leadtek support if it could be a firmware issue - to my surprise they responded - said firmware was latest, card is OK for W7, and mine must have a hardware issue

I wasted hours finding & trying various drivers, all to no avail

Problem just went away when I substituted a different 8400GS card of another make

All I can suggest is try a different graphics card

Yea, if the card was working fine in the system and this started happening out of the blue it sounds like hardware failure.
Try another VGA card.

Unfortunately I didn't bring any, so that isn't an option right now

try a linux live-disk first - just to see what happens.
or plug in a harddrive with a different win version on it.

What's a good live install I can use that is under 700MB? All I have available is a blank CD.

I had something similar with my main XP PC about two weeks ago.

First was loss of wifi after a long defrag. Thought bad RAM but passed memtest. Backed up/reinstalled- but wifi still acted up. Actually, while in device manager, I tried the wifi drivers five times, each with a different error. I pulled the card and examined it- didn't like how the fingers looked so I cleaned them. Finally! So I thought..

Just after I finished restoring everything back over, wlan did the same as before.

While not leading to BSODs, I had "high CPU" usage w/o any processes claiming such, and lousy wireless network performance. Acted like a bad router, but other comps connected to same router and out to internet.

Usually, any browser would randomly show, in its status bar, "downloading proxy script." Funny, no such script existed, and this started suddenly. Anytime the "script" business showed up was when network activity ceased and "high CPU" showed. Again, no process claimed the load, but each core (C2D E8400) got "loaded" to 25%.

Could usually ping and not much else after this point.

Sometimes, it was fine, but within 15 minutes to an hour, network activity ceased, always after that "proxy script" thing.

I got tired of this, took a "software break" and cleaned the case- oddly, the wifi card started working. And quit again. Tried resetting TCP/IP stack and winsock several times, with the fix holding for shorter periods each time. Did a cursory check for malware; didn't expect or find any; remember nothing changed before all this started, except maybe a "heatsoak" during that long defrag.

Since pulling/cleaning the card made a difference for the better, I repeated this. After several cleanings and reseats/removes, it's been OK. This was a real PITA... almost thought BGAs were going bad somewhere. It appears that noise on the PCI bus (from wifi card's oxidised lands) was literally interfering with who-knows-what.

For the last several days, it's been fine. Before the final repair, it certainly would've quit again, many times.

Initially, I pulled cleaned the RAM, video and capture cards, too, when I first had trouble. But it all came down to that stupid wifi card.

What fun...


Pentium4, pull that card and clean the fingers with CRC 2-26 or Deoxit. Spray some on a low-lint paper towel, and clean that troublemaker! If the driver, via the chipset/memory controller/NB, can't talk to the card (open contacts, noise), or hits the wrong address on the card because of an open line or noise-induced "bit flipping," I'd expect nothing less than a crash-n-burn. I really think post-RoHS plating is grossly inferior to the old stuff.


Don't condem your card just yet- I almost smashed my wifi card out of frustration, but things turned out for the better.




That would often work, but it could mask the problem- removing/reseating any card would tend to make it "play nice," even if putting the original back. It was like dealing with a cranky NES with those stupid, cheap contacts. Most PCs are not in a corrosive environment, yet the "goldfingers" have really been acting up. Had two other systems recently where bad RAM connections gave trouble. Yet again, 2-26 saved the day...

We'll take the easy ones when we can get 'em!

Wow, that is such a bizarre issue! Never would have guessed that that was what was causing the issue. Nice job figuring it out, I probably would have just smashed the wireless card

Thanks for sharing Paul, I love this forum because it brings out a lot of difference ideas. Unfortunately, it didn't solve the issue. However, it did need to be done. The contacts were dirty, and there was a large hair wedged in the Pcie slot

Sounds like malware to me.

There was minimal malware on the system. Didn't seem to make a difference.

[ChaosLegionnaire]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

GTX 570

+1 its a dying video card. it is well known from around the web that the fermi cards (gf 400 and 500 series) do not last long. their failure patterns are almost like the wd green drives. they often fail just after the warranty expired or just before. it is not worth fixing/reflowing it unless the card has sentimental value. prolly better to get a mid-end gtx 600, 700 or 900 series card to replace it.

prefail symptoms of the bga starting to get detached from the video board or substrate can also appear as noise, intermittent connection as if its caused by dust when it is not. i had two gf 8800gt cards run at x1 mode and refuse to run at x16 about a week before they finally failed with a screenful of like an "abstract art painting".

i also have a slightly faulty radeon 9600 non-pro. it can install and boot the last catalyst 10.2 drivers for the radeon 9000 series but with minor artifacts in 3d mode. and when i tried the omega drivers, winxp simply refused to boot with an ati2mtag stop msg. im guessing it runs a quick intense test on the hardware on driver load. if there's a fault, it quickly returns a bluescreen, so thats why its highly likely that the gpu is toast.

while u're at it, u can also try upgrading your friend's cpu to a q9650. u can get one cheap for around 70-80 bucks on fleabay/amazon or even better a modded 771 xeon for lga775 at an even cheaper cost of around 50-60 bucks. a g31 board should have no problems taking a 1333 fsb cpu. the vrm shouldnt be a problem also as the q6600 and q9650 both have a tdp of 95w.

What's a good live install I can use that is under 700MB? All I have available is a blank CD.

thats not good... because all of the recent ubuntu flavours have their install images exceed 700mb and so require a dvd.

after doing some digging, your only solution is the lubuntu 14.04.1 alternate amd64 (NOT mac) image. why 64 bit? why an ubuntu flavour? well u need 64 bit ubuntu or opensuse to run geeks3d's gputest program. its the only comprehensive gpu test program for linux that can run several different tests on the hardware.

however, be warned that i'm not sure if the alternate image cuts down on the "extra bits" like the gpu driver bits such that testing the hardware becomes impossible or non-functional for running the required tests...

[kaboom]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Wow, that is such a bizarre issue! Never would have guessed that that was what was causing the issue. Nice job figuring it out, I probably would have just smashed the wireless card

alright- this stuff is supposed to work for us, not the other way around!

I've been noticing that the lands can even look clean and still give trouble. Not like an N64 cart, which can look nasty and still play. RoHS strikes again...

Thanks for sharing Paul, I love this forum because it brings out a lot of difference ideas. Unfortunately, it didn't solve the issue. However, it did need to be done. The contacts were dirty, and there was a large hair wedged in the Pcie slot

There are many different ways to troubleshoot and reach the same goal; part of the experience is knowing what road to take with a given set of problems. Yes, I used a cliche.

There was minimal malware on the system. Didn't seem to make a difference.

That's what I wanted to rule out for my problem too. Really, nothing changed; trouble just started after that long defrag. It could've just as soon acted up after a long memtest. The card moved ever so slightly with temperature, attempting "contact" via an oxidised land. Or, the entire land finally oxidised over- again, this RoHS junk can be bad and not look it.

OTOH, all the caps are still good- SEPCs in the VRM and post-defect HMs elsewhere. FC, PW and HE in the power supply- a Hipro no less...

+1 its a dying video card. it is well known from around the web that the fermi cards (gf 400 and 500 series) do not last long.
...
prefail symptoms of the bga starting to get detached from the video board or substrate can also appear as noise, intermittent connection as if its caused by dust when it is not. i had two gf 8800gt cards run at x1 mode and refuse to run at x16 about a week before they finally failed with a screenful of like an "abstract art painting".

Losing the pairs for the PCIe lanes, one at a time- and who knows what else.

i also have a slightly faulty radeon 9600 non-pro. it can install and boot the last catalyst 10.2 drivers for the radeon 9000 series but with minor artifacts in 3d mode. and when i tried the omega drivers, winxp simply refused to boot with an ati2mtag stop msg. im guessing it runs a quick intense test on the hardware on driver load. if there's a fault, it quickly returns a bluescreen, so thats why its highly likely that the gpu is toast.

I never liked some of those lousy heatsinks modern VGA cards use. Combine that with RoHS BGAs, high temps and cycling, and it's all over. Not if, but when.

while u're at it, u can also try upgrading your friend's cpu to a q9650. u can get one cheap for around 70-80 bucks on fleabay/amazon or even better a modded 771 xeon for lga775 at an even cheaper cost of around 50-60 bucks. a g31 board should have no problems taking a 1333 fsb cpu. the vrm shouldnt be a problem also as the q6600 and q9650 both have a tdp of 95w.

Be sure to check BIOS before swapping the CPU.

[ChaosLegionnaire]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

I've been noticing that the lands can even look clean and still give trouble.

for me, i noticed that the goldfingers look clean but when u wipe them, there is a shit ton of dirt/dust seen on the tissue/cloth after wiping.

I never liked some of those lousy heatsinks modern VGA cards use. Combine that with RoHS BGAs, high temps and cycling, and it's all over. Not if, but when...

thats why i always replace the stock cooler and thermal paste of my video cards with better ones. changing the stock paste alone lowers load temps by at least 5°C or more!

[pfrcom]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

@Kaboom, THANKS

Frustrated by Leadtek PX8400 GS TDH Extreme above, I bought a Deoxit K-DG100L-2DB kit

Leadtek's gold contacts looked immaculate, had previously been wiped with IPA, and didn't appear much different after I'd applied red cleaner followed by gold conditioner

However, trying the Leadtek with a Windows 7 install, went through without an issue

Same for 3 different motherboards, using Intel, nVidia and Via chipsets

Economic stupidity considering price of Deoxit, but when I think of the time I wasted on that Leadtek ...

For any Aussies interested, K-DG100L-2DB currently au$29-95 at Jaycar versus us$39-95 in its Caig manufacturer's store

[RJARRRPCGP]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

the vrm shouldnt be a problem also as the q6600 and q9650 both have a tdp of 95w.

I hate to burst your bubble, but a Q6600 requires an OC to keep up with many CPUs of today and they easily use roughly 250W at 3.3!

Bus errors are real common with Q6600 when OC'ing as well, "STOP:0x00000124" BSOD and "bus/interconnect error" error messages typically rears their ugly head while running a blend mode Prime95. FSB termination voltage is an important voltage tweak with Q6600! I get grief even with a P45 ROG motherboard!

They easily use more power than an FX while getting decimated by an FX. LOL.

[ddscentral]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

I hate to burst your bubble, but a Q6600 requires an OC to keep up with many CPUs of today and they easily use roughly 250W at 3.3!

Bus errors are real common with Q6600 when OC'ing as well, "STOP:0x00000124" BSOD and "bus/interconnect error" error messages typically rears their ugly head while running a blend mode Prime95. FSB termination voltage is an important voltage tweak with Q6600! I get grief even with a P45 ROG motherboard!

They easily use more power than an FX while getting decimated by an FX. LOL.

It also depends on the stepping. Q6600 B3 produce way more heat than G0.
Also Q6600 is a Core 2 CPU - quite old by modern standards.

[RJARRRPCGP]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

It also depends on the stepping. Q6600 B3 produce way more heat than G0.
Also Q6600 is a Core 2 CPU - quite old by modern standards.

Even a G0 is similar once at or around 3.3 in my case... Others may get to 3.5 before using the same amount of watts...

[Pentium4]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Are you comparing the ancient Core architecture procs to the new FX line of CPUs? Of course FX will be better...but Core destroyed everything AMD did at the time it was current. A Core 2 Quad will still play games really well.

[shovenose]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

A Core 2 Quad will still play games really well.

http://cpubenchmark.net/compare.php?...cmp%5B%5D=1038

No, no it won't. A Q6600 will severely cripple a gaming rig for modern games... How can I prove this? Well, GTA V uses about 40-60% of my CPU during gaming, and if an i5-4590 gets a CPU mark of 7214 and a Q6600 gets 2988 then it would be 100% usage and still not keep up. If overclocked really far then MAYBE it could BARELY keep up with what a GTX 570 could do.

[RJARRRPCGP]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Well, I was just showing how long of a way that CPUs went!

Now try getting 4.0 out of a Q6600! I bet it sucks up 300W+ for the CPU alone! And that's probably with a cherry-picked batch...

[ChaosLegionnaire]

Re: nvlddmkm.sys BSOD issue Windows 7 64 bit

Now try getting 4.0 out of a Q6600! I bet it sucks up 300W+ for the CPU alone! And that's probably with a cherry-picked batch...

ummm u are talking about extreme oc-ing. no one does that for daily use only benching. even the dual core conroes struggle to reach 3.9ghz on high-end air cooling much less the quads. a high typical overclock for a quad conroe kentsfield is around 3.6ghz at most. 4ghz on a kentsfield for daily use? not a chance. only possible on a yorkfield.