I'm at wit's end: Catastrophic crashing

[cmkennedy]

I'm finally coming to you all for help. For the last couple months I've been trying to wrap my head around a most irritating problem with Oblivion, and only with Oblivion. As I'm playing, no matter where I am, the screen will randomly lock up, fill with graphic noise, and the sound will begin to loop. I have no way out of it except for restarting my computer with the reset button.

 

Here's what I've done to troubleshoot so far, and each time the problem has returned:

• Purchased a new HDD, right in the middle of the Thailand flood crisis (was about 120% more expensive than it used to be, ouch). Cloned my old HDD, same problem.
• Uninstalled Oblivion completely. Removed all data from "My Games" folder. Reinstalled, patched the game to 1.2.0416 and played with no mods. Crashed before the end of the tutorial.
• Replaced the graphics card, used Driver Sweeper to remove all traces of previous drivers. Reinstalled latest NVIDIA drivers. Uninstalled Oblivion, then reinstalled in the same fashion as above (no mods, just the official patch). Still crashed before the end of the tutorial.
• Formatted the HDD. Installed brand-new Windows XP 64-bit. Installed the game with the official patch only. Still crashed before the end of the tutorial.
• Bought a new copy of Oblivion, this time the GOTY edition. Uninstalled previous Oblivion, installed new GOTY Oblivion. Experienced no crashes before the end of the tutorial, thought I was in the clear. Re-modded everything and started playing again. Crashed in the exact same fashion after about 30 minutes of playing.

I've run Microsoft's on-boot Memory Diagnostic from CD, letting it do its job for several hours and it came up with no problems. I plan to do the same thing again tonight, but leaving it on all night long to see if there's something it didn't pick up before.

 

I've been playing the game for many months – at least since mid-2011, and experienced nothing like this until around late December. It first started after I had done a defragment on the HDD I replaced in step 1, which is what led me to believe it was the HDD causing problems. Not the case, obviously. Nothing else changed on my computer at this time. No new hardware, no new drivers, nothing. I play other games on this PC and have never had this problem in any other game.

 

I've attached a photo taken by my camera phone (since I can't screenshot it) so you can see what I see. This particular shot was taken after I'd tried playing in Windowed mode, to see if that made a difference. As you can see it didn't. Crashed the whole screen, not just the Oblivion window.

Edited February 5, 2012 by cmkennedy

[cmkennedy]

Just a quick update: Ran Memory Diagnostic all night, about 8.5 hours. It had gone through four extended tests and come up with no errors.

[Raffer]

Have you tested your CPU? Try to reduce its clock speed and see if Oblivion still crashes.

[cmkennedy]

Have you tested your CPU? Try to reduce its clock speed and see if Oblivion still crashes.

Just gave that a shot. Reduced clock ratio from 9 to 6, taking my E8400 from 3.0 to 2.0. Crashed after about 10 minutes.

 

I'm not overclocking normally. The only tweaks I've made were to manually set RAM timing to match the manufacturer's specs.

 

For kicks, here's my setup:

 

E8400 Wolfdale 3.0GHz C2D

PNY GTX 260 (running 285.58 drivers; also tried 280.26 to no avail)

GIGABYTE GA-P35-DS3L mobo (link)

Corsair CMPSU-550VX (link)

Windows XP 64-bit SP2

Mushkin Enhanced Silverline 4GB Dual-Channel DDR2 800 (link)

Edited February 5, 2012 by cmkennedy

[Striker879]

I made some basic assumptions regarding your machine's other components and ran it through eXtreme Power Supply Calculator. Using a reasonable Capacitor Aging of 20% your machine (using my assumed specs for things like HHD, Fans etc.) came in with what should be ample headroom. When I bumped the Capacitor Aging up to 50% (the max in the calculator) your machine still showed headroom, but it was getting closer to maxed than I like in my own machines. The fact that your powersupply uses a single 12v rail is good, providing it's not getting close to it's limits.

 

I suggest you try running the calculator and see what you get (without my possibly wrong assumptions regarding your other attached hardware).

[Raffer]

I do not think it is the power supply. An old power supply failing to deliver the required power because of aged capacitors usually restarts the computer when it is under heavy load. It will do this with every game, not just Oblivion.

 

If you haven't done before, try it with RAM timings to automatic setting. If this does not help, can you test the mainboard? It might be something with the graphics bus and the northbridge controlling it.

[cmkennedy]

I do not think it is the power supply. An old power supply failing to deliver the required power because of aged capacitors usually restarts the computer when it is under heavy load. It will do this with every game, not just Oblivion.

 

If you haven't done before, try it with RAM timings to automatic setting. If this does not help, can you test the mainboard? It might be something with the graphics bus and the northbridge controlling it.

Turns out on Auto it picks up the correct timing, which I'd thought it didn't. Switched to Auto, removed the +0.2V overvoltage I had set up. Crashed after about 10 minutes.

 

Are there any drivers that Oblivion is known to have issues with? I've heard Realtek has some incompatibilities with some games, but I've been using a Realtek on-board sound driver on this system for years and have never had these problems. Not sure how I could go about testing the mainboard. Might need suggestions for that.

 

EDIT: I did run the PSU calculator mentioned above. It recommended at least a 520W PSU with 50% degradation. I think I'm good to go in that regard.

Edited February 5, 2012 by cmkennedy

[Striker879]

I could be wrong, but I believe the capacitors the calculator is referring to are the motherboard power circuit. And I agree that trying automatic RAM timing settings may turn up a culprit. I use the same CPU in an ASUS motherboard and Corsair RAM. When setting the machine up after the build I went through a lot of RAM timing tweaking to arrive at my current settings (was troubleshooting a random Win XP crash).

[Striker879]

On a couple of previous occasions Realtek onboard sound and sound codecs turned out to be at the root of some crashing issues, but in those cases no graphics glitches like your's were reported. Have you tried updating the Realtek drivers from their website? My laptop (non-gaming) uses Realtek, and I've updated the drivers a few times. Used to be you needed to uninstall the old ones first, but more recent versions don't require that step (if I recall correctly). Here's a link to their download site.

[cmkennedy]

On a couple of previous occasions Realtek onboard sound and sound codecs turned out to be at the root of some crashing issues, but in those cases no graphics glitches like your's were reported. Have you tried updating the Realtek drivers from their website? My laptop (non-gaming) uses Realtek, and I've updated the drivers a few times. Used to be you needed to uninstall the old ones first, but more recent versions don't require that step (if I recall correctly). Here's a link to their download site.

Updated to the newest release (2.67), as I had been running 2.65. Still crashed.

 

Decided on one more troubleshooting technique that for some reason I never tried before: Set all graphics way down. 640x480 resolution, low details on everything, no shadow processing, 50% view distance for everything (except grass, 0%) and played until I got sick of looking at it, which was about 30-45 minutes. I guess I'm gonna have to start stepping the settings up gradually until I reach an upper limit.

 

So...what could this be? GPU overheating? I think I need to run Speedfan in the background and set an alert for high GPU temps. My motherboard has a built-in alarm for the CPU, and I haven't heard it go off since I bought my Scythes back in April 2010. Replaced 1 120mm fan in front (lower, intake), added another in the back (midway, exhaust), and added an 80mm to the case cover for the CPU (intake). So does this seem to mean what I think it means?

 

Also note that I've experienced the same problem with two different vid cards, one of which was broken and had its fan speed permanently set to 100% (an 8800GT that I replaced with the current GTX 260). I'd try a different PCI-Ex x16 slot, but I only have one. :(

 

Related: How do I uninstall QTP3? Deactivate the omod in OBMM and then clean it? Or should I just reinstall? I'm running QTP3, the unofficial patch, and the QTP3-UOP compatibility omod.

 

EDIT: Found this: http://devnull.sweetdanger.net/qtp3faq.html, which says deactivate then clean. I did use an OMOD. I guess I'll do that for all three, then reinstall the UOP.

Edited February 5, 2012 by cmkennedy