my Geforce 250GTS seems to have a problem concerning the Memory Controller, at least that's how I interpret the facts concerning that piece of crap but since I'm not absolutely sure, I'd like to hear other opinions or possible solutions other than buying a new card.
So first of all the object in question is as stated above the GeForce 250GTS 1024MB (Yeah I know it's probably not the best card to have but hell, it's cheap and I never had a problem with Sparkle cards until now)
It is installed on a Asus M3A32-Deluxe WiFi and is accompanied by a AMD Phenom 9850 and 4 GB of RAM (On an unrelated note: Anyone got any idea how to downclock the RAM on said Mainboard? I got 1066 RAM and I can only use one DIMM per channel since Asus in all their wisdom forgot to mention in the description of the mainboard in question, that 1066 RAM is only supported if there is only one DIMM per channel, but since I have 4 DIMM a 2gb I'd like to use them all) all powered by a Be-Quiet Darkpower 750W PSU.
To finish the description part, I'm running Windows 7 Professional, the Mainboard BIOS is updated to the newest version and there seems to be no newer version of the Graphics Card Bios.
So let's come to the fun part, the actual Error.
Whenever I play Lord of the Rings Online (That's about the only game I have installed at the moment, I recently re-installed Windows to have a clean slate for troubleshooting) on the highest graphics settings it is only a matter of time (and luck, doesn't happen every time) until the monitor freezes, both displays turn black and after about 1-3 seconds the image reappears together with the lovely error in the taskbar
"Graphics Drivers stopped responding and have been restored" (or something like that, firstly the error quickly disappears and secondly I had to translate it since I'm running a German version of Windows).
Before reinstalling Windows the error also occurred while playing "The Witcher" there it happend everytime and occurred after 15min (or even before that)
In LotRO the error can be prevented by running the game on lower graphics settings (I only tested the extremes, Ultra High produces the error, very low runs fine but looks crap )
What I tried to fix it:
Reinstalling Drivers - Failed
Updating Mainboard BIOS - Failed
Using previous versions of the Driver - Failed (though I must admit, I'm not quite sure which version I tested instead of the newest one)
Setting the power of the PCI-E Port to 175 (Default is 75, seemed kinda low) - Failed
After adjusting the power as stated above, I watched the card using GPU-Z since I didn't want it to overheat, Error still occured.
Max Temp was 68°C, GPU Fan was at about 50% (Since the temperature always seemed normal to me I haven't tried adjusting the fanspeed until now) as was the GPU Memory Control.
After the error the GPU Load dropped to 0 (was at about 97% before the error but reached 99% whilst playing before) which can be explained by the fact, that LotRO whenever confronted with this error only shows a black window (Since I use 2 displays, I play in windowed mode, which is handy if you want to control your cards temp on the other screen) which doesn't need lot of GPU resources.
Up until now I wouldn't even think of a broken card but - as the title states - it struck me as pretty odd, that the memory controller went up to 100% after the crash. Although that might be a bug in GPU-Z created by the graphics driver, in which case I would be at a total loss.
id agree with nforce. also how hot is your card getting? is it getting a good airflow? is dusty inside the plastic cover where the fins are?
I have the exact same card, never had a problem with it. mine is built by BFG tho. It runs Crysis and FarCry2 on ALMOST max graphics. I had to turn a couple things down one notch. They run full graphics but it gets a little sketchy during action sequences. (and that's in 1920x1080p)
The card should be getting enough air, there is no dust involved and the card never reaches 70°C
Also I was able to play Dragon Age on full Details in 1920x1080p without problems (This was while the error occured in Lotro or the Witcher)
Hey Tolon, I have had the same issue with memory controller load since i bult my pc five months ago. The problem is after the black screen appears for 2 secs ,memory controller load locks at 100% and remains constant until i restart my pc. The screen goes black while playing cod4, cs source or using apec dc++ for about 10 mins to 1 hour. I haven't experienced black screen issue yet playing games which use little pc resources like cs 1.6 or warcraft 3.
I forgot to mention that sometimes my pc is booting with black screen(i can't see post screen and windows environment). I have to restart pc when i'm encountering this.
I haven't solved this issue after:
Plugging a dvi-i cable of LCD monitor in another GPU's female connector.
Plugging grapchics card in another PCI-E socket.
Reinstalling graphics driver using driver sweeper.
My system specs:
Graphics Card: MSI GTX275 TWIN FROZR OC
Processor Core i7 920 2.66 Mhz
Memory: Corsair TR3X6G1600C9 3x2GB 1.65v
Hard Drive: samsung HD103UJ
Power Supply: Chieftec CFT-650-14CS
Windows 7 x64
Same here GTX 260. After the nvlddmkm error, memory contoller load locks in at 100% until re-boot.
Max GPU temp 49c. Seasonic 650 power- tested. All amps etc. good.
Returned card once, new card same results as old card. Tried all versions of GPU-Z that show memory controller load, all give same results.
The good thing is the nvlddmkm.sys error will not occur again until computer is re-booted.
It would seem though, that a constant 100% load can't be a good thing.
Don't think I would get 2 bad vidio cards. If I could afford it I would try a differant MB.