175.16 - NVLDDMKM.dll error still happening... anyone else?

Status
Not open for further replies.
R

RyanLucier

Gawd
Joined
Nov 23, 2006
Messages
667
UPDATE:
The 8800gtx's were simply INCOMPATIBLE with the x58 system. This same behaviour was exhibited by 7 different 8800gtx cards running in SLI.

The 8800gtx sli crashing problem was fixed on the 680i system however by swapping out the cards for another pair of 8800gtx's. One, or both, of the original cards went bad or was bad all along.

so if you've got an x58 and you're having problems with 8800gtx's... RMA'ing it won't fix the problem. The 8800gtx's have issues with the x58. This behaviour was also confirmed on another system by another person, I had them run the same test with a very similar system and it had the same behavior, crashing etc.
/UPDATE

THREAD INFO:
This thread is documenting the 175.16 and NVLDDMKM.DLL error many people are experiencing under Windows Vista, as well as the things I've tried to solve it for my particular case. Anyone feel free to say you're experiences as well, primarily those with problems, so we can figure out if there's any common threads here.

DESCRIPTION OF PROBLEM:
Using windows Vista, while playing any DX10 games, the screen flickers a few times, and then the inevitable crash follows, resulting in a blue screen of death and the culprit being NVLDDMKM.sys. I experience crashes with the system overclocked and not overclocked at all. And in windows XP, it's stable as can be.

OBSERVATIONS:
1. This error ONLY happens in Vista, not XP, with the exact same system.

2. So far, only experiencing this problem with DX10 games. DX9 games, such as Warcraft 3 or Supreme Commander: Forged Alliance work fine.

3. According to Hasoos, a forum member, if you are playing a game where your frames per second drops under 10 fps, it could trigger this failure. That also means if you underclock your card, and it was already struggling, you could just make the problem worse.

4. According to: http://www.hardforum.com/showthread.php?p=1032579964#post1032579964

Apparently, pirated versions of vista do not have this problem... That leads me to believe that the "crippling" that results from having a pirate version turns off some features that are actually the cause of all these NVLDDMKM.DLL issues. Perhaps coincidence, perhaps something more.

5. Something further supporting Observation 4 is the fact that I've noticed that information pop-ups seem to be the culprit of some of the crashing. Playing world in conflict, at the main menu screen, if I hit the volume up or down button on my keyboard, it results in the driver stop responding and restarting - this is also precisely when a popup occurs on my screen, telling me the volume would be raised (I don't actually see the popup, but it WOULD have shown up if it didn't crash.). It doesn't hard crash in a BSOD though.

Something to ponder, especially since the crash hasn't happened in DX9 games when I raise and lower the volume to the best of my knowledge.

THINGS TRIED TO SOLVE THE PROBLEM:
1. Installed SP1 and all available updates off windows update.

2. Ran a "barebones" vista install with ONLY the 680i 9.64 nforce motherboard drivers and nvidia 175.16 graphics card drivers installed, vista SP1 and all windows updates installed. Crash still happened.

3. Installed nvidia beta 175.75 drivers - "insta-crash" in hellgate london disappears. However, world in conflict develops a "rainbow color" effect, where everything is all trippy, like you ran the entire screen through an invert filter in photoshop or something.

4. Removed all overclocks on system.

5. Disabled SLI.

6. Turned off UAC.

7. Turned off Aero.

8. Heavily underclocked 8800gtx's, does not rectify insta-crash. Haven't endurance tested playing games with heavily underclocked cards, but I imagine it will crash eventually too.

9. Regarding northbridge cooling & cooling in general: I removed the northbridge and replaced the thermal pad with AS5 the day I received it. 2 x 120mm intake on the front, 2x80mm exhaust on the back, as well as 3x120mm fans mounted to the radiator, external to the case so they're not dumping heat inside the case. CPU waterblock is Swiftech Apogee GT. The GPU's idle around 62 and 70C in Vista doing random web surfing etc. and in games never exceed 85C.

10. Renaming NVLDDMKM.SYS in c:\windows\system32\drivers to nvlddmkm.sys.old and expanded the nvlddmkm.sy_ included with the nvidia driver package into the same directory (effectively replacing the old version) and the lockup still occurs.

11. Running the system with a APC 1300VA Uninterruptable Power Backup, so the voltage coming into the computer is "clean." This is the model:
http://www.apc.com/resource/include/techspec_index.cfm?base_sku=BR1300LCD&total_watts=200

12. Bumped up voltage of ram to 2.1v. According to the SPD, the memory is spec'd to run 400mhz 4-4-4-15-2T @ 2.0v, and I gave it an extra .1v.

13. Removed soundblaster X-Fi Elite pro from the system, didn't help.

=============================
As always, if anyone has any ideas, fire 'em up lets figure out WTF is wrong with this problem for everyone experiencing it. Hopefully by documenting things attempted to fixing it and putting it in one neat thread we can get some sort of ideas going.
 
YEAH i had that same problem...which made my 750i/9600gt bug worse..dropped back down to the 174's and that error stopped
 
My sig rig is running all of the latest games, CoD4, Crysis, Assassin's Creed, Gears Of War, perfectly stable. I'm on the 175.16's as well.

What games are you having problems with?
 
I've seen this on and off for over a year, and most recently, just this week. It's a Vista version of the NV4_dsp.dll from 98/XP, but way more annoying.

For what it's worth, this explains exactly why the error happens, and a possible fix. I say possible because sometimes it works, sometimes it doesn't.

http://www.microsoft.com/whdc/device/display/wddm_timeout.mspx

Good luck. I'd be happy to discuss some other fixes in PM, or you can send an email to my Puget addy.

http://nvidia.custhelp.com/cgi-bin/nvidia.cfg/php/enduser/std_adp.php?p_faqid=22
 
Heather, if you dont mind, Id rather contact you in the forums so others can possibly solve their problems if they have the same problem as me. I read through the links you provided, but the first didn't actually say any possible solutions at all, unless you were referring to increasing the timeouts? Could you explain a bit further what you were suggesting me to try? As for the second link, everything is kosher on that front as well. Bios is configured correctly etc.

The problem occurs in any game that uses DX10 so far from what I've seen. The more intense the graphical settings, the more likely a crash. In fact it all started with me wondering how much of a slide show crysis would be when run at 2560x1600 with maximum details - yes I realize not playable at all, I was just curious. Anyways, it resulted in a very fast crash. But it seems any DX10 game will cause the crash. The very same games which function just fine in XP will not function in Vista, with the only real change being a switch from DX9 to DX10.

Things I've noticed/tried:
1. Enabling/disabling SLI makes no difference (although when disabled, sometimes it doesn't BSOD, it just exits out of the game and says the driver recovered, whereas SLI has always BSOD'd).

2. Overclocking makes no difference, if my cpu/ram/gpu are all set to stock it still happens.

3. Turning off UAC makes no difference.

4. I tried a fresh vista install with nothing installed besides the graphics driver and nforce drivers and there is no difference, it still crashes, even without sound drivers and mouse drivers etc. installed

5. The 175.75 beta drivers seem to eliminate the error too, might I add. Hellgate London doesn't insta-crash like it does with 175.16 drivers, even with SLI enabled and maximum details enabled. But these are supposedly internal beta's, and I'd much rather use WHQL drivers rather than untested beta's.

6. SP1 and all windows updates installed, no difference.

any ideas?
 
i have a feeling that its largely caused by the 2560x1600 resolution, I'm going to do a bit more research on that...
 
RyanLucier said:
Heather, if you dont mind, Id rather contact you in the forums so others can possibly solve their problems if they have the same problem as me. I read through the links you provided, but the first didn't actually say any possible solutions at all, unless you were referring to increasing the timeouts?
Click to expand...

I was--sometimes that does resolve the issue. Sometimes it doesn't. It depends entirely on when it happens, I think. If it happens in game, increasing the the TDR does help, if it happens randomly while at the desktop, it doesn't do a blasted thing other than cause other things to crash. I solved the same issue about 22 hours ago by changing the TDR settings in the BIOS for a customer who was crashing to desktop in EQ2.

Since we're going to try to hash this out on the forums, can you post a copy of your dxdiag file? If it's DX10 related, that will give me something to take back to my industry contacts.

**Oh, and some people have reported that disabling UAC has fixed it for them. Others have taken that NVLDDMKM.dll file, renamed it to NVLDDMKM.old, and reinstalled the drivers...and that fixed it. In the end, it seems to largely be a crapshoot for those still unfortunate enough to still have this stupid problem almost 2 years later.

***There were hotfixes out, too, but I have no idea if those were included in SP1 or not. 940105, 938194, and 938979 were the numbers.
 
Thanks heather, and by the way, I wanted to point out the fact that I truly appreciate you helping me even though I'm not even a customer of yours. Kudos to Puget!!!

Now in regards to your ideas:

1. Ok, I'll try messing with the timeouts, although to be honest this isn't something I'd think should be necessary, or even attempted really. I will also dig up a copy of the dxdiag file for you guys to take a look at and post it here later on, can't do it at the moment as I'm on a different PC.

2. UAC is disabled, and I also tried deleting NVLDDMKM.DLL and reinstalling it with the one from the 175.16 drivers. Didn't help.

3. According to:http://technet2.microsoft.com/Windo...7038-4e82-a32c-4bc10ffe56ab1033.mspx?mfr=true

It says the hotfixes you mentioned were included.

Thanks Heather.
 
No sutch shit anymore. In my case it was a cooling issue dispite the best aircooling there is for it, TR HR03 Plus and/or to High OC on the memory on my old 8800GTX. After put it under water with a FC-block it was gone on the level where it was before. Never seen it with the Gx2.
 
Well, I tried underclocking my 8800gtx's quite heavily, and the crashes still happened, instantaneously for some programs as well, such as crysis or hellgate london. I can reproduce the crashes without fail... My GPU's never exceed ~78C, and the cpu never exceeds ~50C. I'm quite positive it's a driver issue.

In addition, my CPU has a triple radiator / swiftech Apogee GT watercooled setup, and this is in a coolermaster stacker with multiple 120mm fans bringing air in. There is plenty of airflow, and lots of open space in the case, it isn't cramped. Long story short, in my case, I don't believe it is a cooling issue. I also want to repeat that this system works just fine in Windows XP, no games crash. I also use this machine for adobe CS3 design work (i.e photoshop CS3, flash CS3, dreamweaver CS3) and it's solid as a rock.

Furthermore, I have a silverstone zeus 850w power supply and a APC 1300watt power backup, so the power coming into the system is clean and sufficient.
 
My pleasure, Ryan. Part of the reason that I work in this industry is because I like the people here, and if ever I can help, you better believe I'll try.

I'm quite positive that it's a driver-related error as well, and it's incredibly frustrating every time I see it manifest itself. It has so many different variables at play, and each time, there seems to be a different way to fix it. There's no "across the board" fix. Microsoft acknowedges there's no real fix for it other than the one I linked to...and that's about all the help you can get.

Who manufactured your video cards (ie XFX, BFG?)? I might be able to get you some help from that quarter. Nothing says lovin' like sticking the manufacturers on the spot for these problems. :) It may well be that you have a bad card...not always the answer, but occasionally, I've seen a card replacement fix it. It's almost like the BIOS on the card creates the problem with the driver in some cases, though I can't get official confirmation on that.
 
RyanLucier said:
Wow... A year and a half after vista's release, and I'm still getting nvlddmkm.dll crashes. Seems to only happen in DX10 games, and the higher the graphical settings, the more likely a crash.

The screen flickers a few times, and then the inevitable crash follows. I experience crashes with the system overclocked and not overclocked at all. And in windows XP, it's stable as can be.

Anyone else experiencing this with 175.16 drivers?
Click to expand...


Hi:

I have had and in certain scenarios continue to experience this problem. In my opinion it's a heat problem with the motherboard. The heatpipe arrangement just isn't enough to keep the temperatures low enough, and when the northbridge gets hot enough bizarre problems creep up. Turning the PC off for an hour and letting it cool completely off solves the problem.

I noticed you have a 680 nvidia motherboard, it was so bad on that motherboard that the PC was nearly unusable. I went and bought a new video card and a new motherboard (780i / 8800GT). It still does it, but only when it's thoroughly heat soaked.

I won't buy another nVidia motherboard.

Just my personal experience and observations - truth may vary. :D
 
Yeah, the since they've introduced passive heating on the northbridge as a standard, I've seen all kinds of weird instabilities pop up. The bad thing is that there's no real way to get an accurate temperature because of sensor placement--I end up using an IR gun to try to nail down temps in house. The problem was alleviated some with the introduction of the optional northbridge fan that comes with the motherboard--supposedly for non-air cooled systems. Sometimes installing that over the NB heatsink will help, even if you're not running liquid cooling.

As to whether this is part of that, I have my doubts. The northbridge is the conduit between CPU and memory...it shouldn't make you throw a nvblahblahblah.sys error.
 
I had that problem with my orriginal 8800GTX. I did everything to try and fix it. I ended up sending it back to EVGA to get replaced. I bought a 8800GTS (G92) to replace while i waited........no errors with that. I received the new 8800GTX back and said hey i'm going to try it again....................no problems........
 
Yeah, I got that once on my new rig, really annoying. Right after I installed the 175.16 driver too (I was hoping they were solid since it was non-beta release). Pretty sure it was Crysis (big surprise there). I think its just the driver, I'm not sure.

One thing that may help, adjusting the fan on the GPU. I've been using the EVGA Precision, but there are other apps (like RivaTuner). For some reason the auto-fan wasn't kicking in at all. It didn't sound like it was ever going higher than 50-60%. I set it up to like 90-100% when gaming. I've heard that helps with the stability (I'm also running a factory-overclocked card). Saw my GPU temps drop from around 60C load to around 45C load (according to EVGA). Haven't seen any BSODs since, but I don't want to jinx my luck.
 
With this latest nvidia driver I have been getting crashes in city of heroes. The screen flickers/flashes a few times, then the system locks with a loud buzz. Display driver every time.

I tried using driver cleaner and a clean install too. I'll have to try out the beta driver.
 
I just got this problem twice with Rainbow Six Vegas 2 (which is running in DX9, I believe.) Happened in the same place both times, and I am using the 175.16 drivers. Nvidia needs to get its shit together...
 
Hmm.. now it makes sense... all of a sudden, Bioshock, UTIII, and STALKER all decided not to even run... looks like this driver might be to blame. I'll try reverting.
 
i'm glad someone brought up the point about the northbridge passive cooling. the first time i felt the northbridge heatsink it almost burned my damn fingerprint off. on every system i've had in the last 3 years i've put one of those $25 thermaltake heatsinks on and have never dealt with heat related crashing/instability. i'm not sure what the mb manufacturers are thinking by not putting a $3 fan on to prevent 100's of possible rmas.
 
If it makes you feel any better, I think that the newer tech is probably going to not be northbridge/southbridge dependent at all.
 
The crash has occurred for me on Vista Ultimate SP1 x64 @ 1920x1200 on most games. I found out that the best way to combat this is to turn off all internet apps like email & IM to avoid random popup balloons & windows. This slimmed down the amount of crashes by a lot for me. Seems that the Desktop Window Manager has a tough time with Nvidia's drivers and dies in a heartbeat when a separate event occurs underneath a game.
 
Hey guys, thanks for the continuing responses/help!

Heather: My graphics cards are EVGA 8800gtx 768mb cards, stock clocked at 576/900 and I'm running them at 620/1000 (again though, underclocking has no effect). I can get you more info like serial #'s etc. if needed.

Ok, some observations/questions:
1) Regarding northbridge cooling: I removed the northbridge and replaced the thermal pad with AS5 the day I received it. Also did the same for 2 other sets of heatsinks on the boards, I forget now what they were cooling but it was recommended on many sites/forums at the time, including EVGA forums and the [H]. I forget what the heatsinks were specifically cooling, I want to say mosfets but I don't think that was it. That aside, I never had an issue with that or NB temps reaching the astronomical highs I've heard others reach. I also have the fan installed on the NB as well. I really, really, do not feel cooling is an issue with my system. 2 x 120mm intake on the front, 2x80mm exhaust on the back, as well as 3x120mm fans mounted to the radiator, external to the case so they're not dumping heat inside the case.

2) The "flicker then lockup" issue I talked about, I can instantaneously reproduce it in hellgate london. It happens at the exact same time immediately after changing resolution / DX10 etc. This isnt' something where my cards are "heating up" I'm talking from a stand still startup and immediately launching the game, crash is guaranteed 100%. Never fails.

3) These cards have never had a single problem with any XP games whatsoever. Highly doubt they're at fault. But I'd be willing to entertain the thought if EVGA wanted to ship 2 cards for me to try, I'd be happy to try.

4) Just for the sport of it, I reinstalled vista tonight and created a hard drive image of my entire system ready to roll with all drivers etc. except the graphics drivers. Going to try experimenting with drivers and seeing how things compare with a completely clean install, one driver against another.

As always, if anyone has any ideas, fire 'em up lets figure out WTF is wrong with this scenario for everyone!!!
 
i think i could live with hellgate london crashin... that game won hand down worse game of the year
 
lol, to be honest I played an hour of hellgate crash-free in XP so far, and its been relegated to vista crash tester. ;) It's quite crap. But still, it's crap that shouldnt crash flat out in vista.
 
Alright, Ryan, I sent a call out to contacts at eVGA. I'll let you know when they get back with me, and I'll loop you into the e-mails with them.
 
Ive been getting driver stop respondign with new drivers again...this time in COD4. SOmetimes it does it sometimes it doesnt but pissing me the *&() off again. Heard only in Vista this happens. someone needs to get their act together on this bc lots of peeps r having the problem and still know1 knows how to fix it.
 
Well, turns out WiC crashes too. So far, DX10 = crash, Hellgate london, crysis, and World in Conflict all crash...
 
Ok, I've made the main thread all nice and tidy, hopefully we can organize the information and come up with some sort of solution.
 
I get these crashes in TF 2 with an 8800GT 512mb OC from eVGA. latest drivers
 
I no longer get the crashes on my Ultras, decided to go down in drivers and currently on 169.25.
 
I've found out that I can greatly reduce the lock-ups/crashes/BSoDs of games in Vista SP1 x64.

1) Turn off all web-based & notification-featured apps.
2) Disable desktop composition & visual themes for each game through the compatibility options through game executable properties.

I've been testing this for months & it seems to actually work out really well. This has been tested with almost every driver version up to the one stated in the main topic.

I'm curious to know if this works out for anyone else. Thanks.
 
One thing I noticed on the technote listed above that some folks seem to have missed. It mentions that if you are playing a game where your frames per second drops under 10 fps, it could trigger this failure. That also means if you underclock your card, and it was already struggling, you could just make the problem worse.
 
Status
Not open for further replies.
Back
Top