Nvlddmkm stopped responding is getting VERY OLD...

[pgruebele]

I just spoke to EVGA tech support and they say that 4 DIMMS always causes problems with anything overclocked. My RAM is spec'd for 1200Mhz 5-5-5-15 but nevertheless EVGA considers anything over 800Mhz overclocking (rightfully so). I'm still testing so I can't say that the TDR problems are 100% related to 4 DIMMS, but in any case I am returning my 4 DIMMS in favor of 2 DIMMS since this should generally make it easier to run my ram at spec'd speeds and stably.

I can say that this TDR error I am getting is 100% a hardware issue because:

1. It starts to happen more and more the longer the PC is on.
2. Once it start to happen, rebooting has little effect. The only thing that helps is powering down for a while.

This would not be the case if there were a driver bug. This is with 163.11 beta drivers....

High speed ram can be hard to find, but newegg has "MEM 2x1G|PNY DII1173 D22GX93XLP-5" 1175Mhz 5-5-5-15 2GB kits for $266, so I am returning my Kingston HyperX PC2 9600 512MBx4 DIMMS. Although spec'd slighly lower, 2 DIMMS will actually allow me to run at higher speeds...

Philip

[cabnz]

I'm using 4 DIMMS but I never over clock so maybe it's not the problem

[ekovalsky]

Four DIMMS does put more stress on the memory controller, which may need a bump in voltage to stay stable at settings that are fine with two DIMMs. Also it is not really possible to run at 1T command rate with four DIMMs, unless you drop the clock speed way down and give the NB near maximum voltage. This is particularly true with the latest 680i BIOS (P29 & P30).

Still, I think the issues are within the drivers rather than hardware. My conclusion is very simple -- with 158.45 everything works fine and stable even with four DIMMs, two 8800ultra in SLI, and major overclocking. With the 16x.xx drivers I can reliably trigger the error doing mundane things with Vista Aero with the system stock or underclocked. Turning off desktop composition (disabling Aero) makes the errors go away, even with overclocking. All this suggests that the added DX10 SLI functionality as the culprit. It is not just a bad interaction with SLI & 4GB RAM, since the 158.45 drivers work fine and they do support SLI in DX9.

[Slammin]

Running 4 DIMMs at 1T here.

Do the registry fix and turn off DeskTop Composistion (Aero).

Works for me, but I can make the problem reappear by running my ram too fast. As it stands though, running my ram at 3-3-3-3 1T and below 670Mhz keeps everything fast and stable.

By the way, I've flashed my cards to 648 core, 999 mem and they run a lot cooler with Desktop Comp disabled too, on the order of 7c cooler at idle!

[cabnz]

What is the best way to disable desktop compensation?

[cabnz]

Oh, and whats the registry fix?

[ekovalsky]

You can disable desktop composition via:

My Computer > Right click to properties > Advanced System Settings > Performance Settings > Visual Effects > clear the 3rd check box down

By the way I only have the nvlddmkm errors under the 16x.xx drivers when SLI is enabled; if you select "do not use SLI technology" in the Nvidia control panel the errors disappear. So, I again infer that DX 10 SLI is broken on the drivers released after 158.45, at least with four DIMMs on a 680i reference board. Maybe the non-reference boards i.e. Asus, Abit, etc. don't have this problem ?

[AgentSmith2k]

I just had to say that I've had this problem for quite a while and was driving me nuts. Every single day I would get it no matter what I was doing, even when I’m not doing anything. Some days it would happen soo many times at once that my PC would either reboot, freeze or BSOD.

For the past week though I haven’t had it not once(nana2). I don't know if this would be any help to anyone else but for a while I have been using 4Gigs of ram (1Gig per slot) but have changed to my old CellShock 2Gig. No matter what I’m doing or how many apps to stress the system I’m running (would usually lead to Nvlddmkm error) I get no problems, apart from running out of memory.

Maybe someone with spare memory sticks can try this as well.

[AgentSmith2k]

I've also tryed the 162.22 and 163.11 drivers, both work fine.

[ekovalsky]

I've also tryed the 162.22 and 163.11 drivers, both work fine.

Nice rig. Looks like you are not running SLI which is what is causing the errors with the 16x.xx drivers, at least on my setup. With one GPU installed, or with both installed but SLI disabled, I don't have the errors. Once SLI is enabled they are present.

[AgentSmith2k]

I'm going to be purchasing OCZ Reaper 4GB PC2-9200 soon so I will be able to find out if it was specific to that memory.

At the moment I'm just happy of my error free time.

[vmsein]

I'd like to chime in here to once again completely agree with ekovalsky's conclusion that the evidence would suggest that this is a driver level problem further exacerbated by enabling DX10 support within the driver in SLi mode. The problem goes away when SLi is not enabled in the control panel, and the problem goes away with the 158.45 driver with SLi enabled. Correct me if I'm wrong, but DX10 support was introduced in drivers that came after 158.45?

[HYBRID]

well.....after using XDC Cleaner and installing 163.11 on my rig a week ago Ive been lucky, and the NV driver stopped error has not reared its ugly head.
However from time to time I get the runtime error on Enemy Territories atleast this one doesnt completely bork things.

I also notice that when I get either the runtime or the NV stopped working driver error showup after reboot I have no internet access I have to reboot a second time wierd.

[Revs]

I've only had it once with the newest WHQL drivers, but about once an hour on average with the old ones (non-whql). It's always on the desktop and always when SLI is on. It also never happens when the PC is just stood or in games.

So I would agree with ekovalsky and vmsein. It seems to be a problem with Vista using the nVidia SLI drivers for the desktop stuff.

[AgentSmith2k]

Well i've been testing the OCZ Reaper (4gig 9200) and havn't had a single error. I't may have just been a combination of the drivers and the Geil memory but its hard to say.

[EverGreen]

Yup driver 162.22 WHQL did also works here fine now, no more nvlddmkm problems in the main driver.
I did installed several driver versions and did look inside the nvlddmkm driver, noticed that some parts been missing to run all the graphic cards in full 3D hardware acceleration mode, besides i could see fine when i did try to run no-limits coaster who gives me all the time some nfo that my 3D hardware had failures.
Everest told me that a to much old driver did been used for 3D hardware acceleration 97.64 beta driver did do the lead, even when installed other or new ones it tells me all the time that still that beta driver been there.
So the whole nvlddmkm bug is nothing more than a false setup who not overwritten this 3D hardware accelerator standard.
When im looking inside this nvlddmkm driver i noticed with this 162.22 WHQL that there been some instructions missing, nothing is depended on over or underclocking the software, not your mem banks or any other lies by guys who think they have the sollution for it.

Same apart for the guys from nvidia, stop with your pretting lies to tell me that i need to empty my pockets for new HW, all pretty dumb lies to make every one fools.

I didn't have any more BSOD's now who were killing my PC all the time down...

[AgentSmith2k]

Why would you think i'm lying?

I'm just letting others know what solved it for me i'm not saying "If you wan't to get rid of the Nvlddmkm error go out and buy new hardware". Maybe it would be better if everyone with a suggestion or possible solution should just keep it for them selves.

That damn error f****d me off so much so i know what people with the same error feel like doing.

The ONLY hardware/software that changed was MY memory and the error stopped.

Oh and the 162.22 gave me the Nvlddmkm error on my Geil memory.

[Slammin]

Yup driver 162.22 WHQL did also works here fine now, no more nvlddmkm problems in the main driver.
I did installed several driver versions and did look inside the nvlddmkm driver, noticed that some parts been missing to run all the graphic cards in full 3D hardware acceleration mode, besides i could see fine when i did try to run no-limits coaster who gives me all the time some nfo that my 3D hardware had failures.
Everest told me that a to much old driver did been used for 3D hardware acceleration 97.64 beta driver did do the lead, even when installed other or new ones it tells me all the time that still that beta driver been there.
So the whole nvlddmkm bug is nothing more than a false setup who not overwritten this 3D hardware accelerator standard.
When im looking inside this nvlddmkm driver i noticed with this 162.22 WHQL that there been some instructions missing, nothing is depended on over or underclocking the software, not your mem banks or any other lies by guys who think they have the sollution for it.

Same apart for the guys from nvidia, stop with your pretting lies to tell me that i need to empty my pockets for new HW, all pretty dumb lies to make every one fools.

I didn't have any more BSOD's now who were killing my PC all the time down...



Amazing.

Even with your broken English, your arrogance and ignorance speaks volumes.

We are all here offering suggestions that may help others. Several suggestions, such as lowering CPU overclock, memory overclock, disabling Sli, etc., have in fact worked for some, and not so well for others.

Amazing.

[EverGreen]

Amazing.

Even with your broken English, your arrogance and ignorance speaks volumes.

We are all here offering suggestions that may help others. Several suggestions, such as lowering CPU overclock, memory overclock, disabling Sli, etc., have in fact worked for some, and not so well for others.

Amazing.

LMAO it isn't personally to you pointed, but it seems you been not alone on this planet, perhaps forgotten to be polite in personal mind, that there been more peeps round the globe whom been on this forum, who have another English grammer than you will do.

But besides this im not here for your bad reply bout my bad english or arrogance and ignorance, im just here to figure out several problems to solv this bugs and i found several bugs problems inside this nvlddmkm main driver.
Perhaps you forget also that im a few years been around on this board and did read every topic very carefull before i written something back, but there still been some.

Besides this team of Vista did launch some nice Vista Compatibility & Reliability Patch (Vista 32&64-bits) on Guru3d

32 bits - http://downloads.guru3d.com/download.php?det=1686
64 bits - http://downloads.guru3d.com/download.php?det=1687

After testing see no more the annoying nvlddmkm main bug driver problems, my HW acccelerator is found again and lol it seems works fine again.
Not all problems for all nvidia card users will be solved at once but the first steps been taken after 6-7 month's to really solv this major problem.

Same for the latest DX 9 en 10.1 Aug release some lilltle problems been also solved, without changing or tweaking up or down your system.

Running several rigs now without any BSOD's, my P35 with 4 gb and a 2 SLi Gigabyte 8800 GTX freezes still from time to time.

Perhaps i need a full job at nvidia for helping this problem out of the world.

[pgruebele]

These TDR errors are incredibly annoying and make you want pull the hair out of your head!

I finally got rid of ALL TDR errors by changing my memory. I was using 4x512MB DIMMS and switched to 2*1GB DIMMS. I am even running the system with higher FSB/memory speed, so the memory definitely seemed to solve the problem. My EVGA 680i SLI board just had trouble with 4 DIMMS, although I suspect that one of the 4 DIMMS had a problem so 4 DIMMS may not actually what caused the problem.

So, I can say that in my case, hardware definitely appeared to be the problem. I have been running with 163.11 drivers the whole time.

My advice to people with this problem is to uninstall old drivers, clean the system, reinstall at least 162.22, and overclock their system less. If this does not work, see if you can narrow down the problem by checking/swapping individual hardware components, if possible.

Hope this helps some of the poo folks suffering from this.

Philip

[pgruebele]

I also installed "32 bits - http://downloads.guru3d.com/download.php?det=1686" although I got it from another source. However, the big improment came from changing the memory, not the MS patches.

One of these patches does supposedly fix some Vista memory corruption problems as well as some video card driver compatibility issues, so perhaps some people that are getting these TDR errors could improve the situation by installing these 2 KB patches.

On another forum someone said that these patches did not fix their TDR problems, but you never know. These patches certainly did not cause any problems on my machine aside from requiring multiple reboots, but under X64 they no longer allow permamently disabling the driver signing feature.

Philip

[-n7-]

FWIW, since using the 163.11 drivers & now the current 162.22 drivers i have not seen the "driver stopped" error once.

[ikjadoon]

Ditto, n7.

~Ibrahim~

[sclaws]

Wish I could say the same...tried 163.11 and 162.22, both would eventually cause the error. I'm still on 158.45 hoping for another release soon :(

[vmsein]

The TDR errors seem to affect those of us that are trying to run SLi configurations the most. I do not get the TDR errors if I disable SLi with any of the recent driver releases.