Lockup after few minutes in 3d

[VirtualGhost]

Hello,
My system has been running well for years, and I have not done upgrades of late. However last week I started to get strange lock-ups while loading a game of Fallout New Vegas. Because the game is rather buggy, I first dismissed it as game related issue.

Game would load just fine, but it would crash after moving the character a bit and turn the screens to rainbow coloured nonsense locking up computer while doing so. It would happen with starting a new game too.

Only way to get past that seems to be to do a cold boot (turning off power entirely) and game would run just fine. Normal reset wont do, problem will continue with reset.

After few days I noted it affects other games too, with exact same symptoms so I decided this must be something beyond a single game.

The problem happens first time I switch computer on that day, so it has been unused for hours. I do not think it can be overheating as I can play the game hours without trouble and the crash only happens in first few minutes or not at all.

Does anyone have any ideas? This is just weird in my book, what could be failing? Is it software or hardware? Only one clear indication is that it happens ONLY when using 3d accelerated game.

Windows 7 Professional 64-bit

Memory: 4096MB RAM
Available OS Memory: 3008MB
RAMPage File: 2198MB used, 3814MB available
DirectX Version: DirectX 11
DX Setup Parameters: Not found
User DPI Setting: Using System DPI
System DPI Setting: 96 DPI (100 percent)
DWM DPI Scaling: Disabled
DxDiag Version: 6.01.7600.16385 32bit Unicode


Card name: NVIDIA GeForce GTX 280
Manufacturer: NVIDIA
Chip type: GeForce GTX 280
DAC type: Integrated RAMDAC
Device Key: Enum\PCI\VEN_10DE&DEV_05E1&SUBSYS_82901043&REV_A1
Display Memory: 2257 MB
Dedicated Memory: 1009 MB
Shared Memory: 1247 MB
Current Mode: 1680 x 1050 (32 bit) (60Hz)

I would really appreciate any ideas.

 

[wa1]

The first thing you want to check is the card temps, use GPU-Z or HW Monitor to check all components temps...
http://www.cpuid.com/downloads/hwmonitor/1.17-setup.exe
and
http://www.techpowerup.com/gpuz/

 

[VirtualGhost]

GPU-Z gives GPU idle is 75 C which seems too high, but when checking at the same time with Core temp (http://www.alcpu.com/CoreTemp/) with Core0 and Core1 both are about 53 C which is normal, to my understanding.

What conclusion should I make? If it is temperature, shouldn't it crash always? There is no gaming issues if the start up lock-up doesn't happen, that is why I did not think it can be heating issue. Is this typical to have it overheat at first boot, and somehow even up after short cold boot and work perfectly? I never had this sort of heating issue before.

 

[wa1]

How old this card? (GTX280)
Make sure it's free from dust and you might want to re-apply the thermal paste, in case the old paste is dried...
I trusted GPU-Z more than Core Temp if you ask me...
What about your PSU? brand? model?

 

[VirtualGhost]

Processor: Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz (2 CPUs), ~2.4GHz
Power Supply is Thermal Take Toughpower 750W.

Graphic card is almost 2 years old, actually its 10 days from warranty to expire, now that I look at receipt. Its by Asus. I have two monitor setup to both of the DVI ports.

I vacuumed it a bit, but so far no huge difference. it seemed to start from 56 C and is slowly climbing up to 70 C or so over 15 minutes of idle.

I guess for better clean up, I should take the thing out of the case. I do not have experience cleaning these cards that come in plastic case to begin with, they open up easily or I have to break them open (thus probably nullifying warranty).

I have no overclocked this, I have just used it on standard settings.

 

[benski]

70c+ at idle is definatly a problem. Try MSI afterburner, it has some graphs that will show you what your temps are over a period of time and if your fan speed is reving up as it should when the card gets hot. Get a can of compressed air and blow the card out real good, dust tends to get packed in the fins on the cooler and vacuming alone probably wont' get it out if it hasn't been cleaned in the 2 years you've had it.

 

[VirtualGhost]

Does it matter that I do not have MSI card? GPU-Z seems to present average score as well. I guess the goal is to get it to 50 or so?

Yeah, it has been properly cleaned in two years. Do I need to try to open the card's plastic case for the dusting, does it come off without breaking warranty?

 

[VirtualGhost]

Some additional information. I have been keeping my case open until I get the chance to visit a store to get the card a proper cleaning. I had GPU-Z to make a log while I played, and this is the data I got up to the crash. At some point of gameplay the temperature was up to 82 C, but at crash point it was 75 C. Idle is now around 50 C, which I guess is normal.

 

[VirtualGhost]

As an update, cleaning up did not have an affect. It had some dust, but notably so.

The problem occurs every other boot, systematically. I can for instance boot the windows up, then shut it down, take the power off and when I load windows again nothing will crash (had 6 hour sessions or such). Its rather bizarre.

If it is hardware, could it be this systematic? Shouldn't it fail more randomly or every time.