Is my graphics card giving up?

[ConnyG]

I February I build a new computer:
- Asus P5E
- Intel E6750
- Corsair TWIN2X2048-6400C4 (2 set for a total of 4 GB)
- Leadtek GF8800 GTS 512 MB
- Vista Ultimate 64-bit

It started out working just fine but after a few weeks it would hang. Vistas memory diagnostics said my memory was bad. After reading Asus support forums I flashed the BIOS to 0605 and then the memory checked out just fine. Last week I also ran MemTest86+ both for DOS and Windows and the memory passed so I feel resonably confident that the memory is ok.

Again it was mostly stable for a few weeks but last week all h*ll broke loose. I fired up World of Warcraft on wednesday evening to do some raiding and in the process I knocked some stuff off my desk which fell and hit my graphics card cable/DVI-VGA adapter connected to the motherboard.

This caused my computer to hang with a blank screen and now it even refused to POST, gave me a black screen, froze on the boot menu (boot windows failsafe or normal, etc) and on several occasions I got into Windows and was able to start WoW only to have it hang immidiately. It took me 20 minutes before it started working normally and then I could play for a solid 4 hours.

Yesterday problems were back. I could log into WoW only to have it hang immidately. I therefore proceeded to remove all my components from the chassi and assembled everything on my desk.

I noticed my memory was very hot, almost that it burnt me when I put my finger on it so I put a fan facing the modules to cool them. Temps are ok, about 30 for CPU and MB (used to be 31 for CPU and 37 for MB when it was installed in my case, about 80 for the GPU when gaming).

The positive thing is that now the computer will POST and boot correctly even after a hang, the negative is that I cannot start any games at all. WoW = hang. Overlord = hang. 3DMark06 = hang. I even reinstalled Vista last night and just put the Nvidia driver it. Now it hangs when Vista does the D3D parts of the "Peformance Rating".

Reading on Asus forums it seems a lot of people have been having problems with this motherboard so I am still not 100% sure what is causing the problem.

Do you think the graphics card is broken and needs to be returned? (I'm gonna go buy a cheap card today to check if it works stable)

Could possible overheating of the RAM have caused the problems with POSTing?

Could a broken graphics card have caused problem with POSTing?

Could there have been some issues with grounding or something electrical connecting which may have caused instability or zapped the graphics card? (or caused the problems with POSTing)

 

[ConnyG]

From my point of view:

- if the memory is faulty, incompatible with the motherboard or overheating I should not be able to run MemTest86+ successfully (standard tests, one full run done successfully)

- I have not tested the CPU yet, but I have no reason to suspect it at the moment

- running Windows seems to work fine now, but anything 3D hangs (games, Vista performance rating test)

I cannot explain the POST issues or freezing on the boot menu. The motherboard once gave me "BIOS corrupt, blah blah blah", on at least two occasions failed to find any boot devices and if it fails to POST it will reset the BIOS settings (this is default by design though). In these issues Vista should not be the problem though. These issues are why I originally suspected only the motherboard.

Right now my money is on the graphics card, motherboard possibly due to something zapping any of them.

 

[error8]

The videocard is probably damaged because of your "stuff" hitting the graphics cable. I would say to open your case, take the card out and inserting it in again. Also, you could try using another cable, maybe the shock of the hit damaged it somehow and switch to the second DVI port.

The worst thing is that something on the card was damaged and I'm thinking at the gpu mostly. If this is the case I don't think the warranty will cover it.

 

[ConnyG]

It has crashed/hung several times before that incident and the graphics card was firmly secured in the chassi and so was the cable/adapter to the card. I'd be very surprised if this damaged the GPU (it's not like I dropped an ACME 1 TON weight off my desk). I have tried re-seating the card. A bad cable would not explain why the computer totally hangs. As mentioned it appears to be 100% stable as long as I don't run any 3D-stuff/games.

I will try the different PCIE-x16 slots and DVI-ports today, as well as another graphics card, to rule out the motherboard.

 

[Denithor]

I would suggest trying the other DVI port to see if that makes any difference.

 

[chizow]

The hanging in 3D apps points to the RAM going bad as those symptoms are identical to the problems I had when my RAM died on 3 different occasions last year. Let me guess, those Corsair are 2.1-2.2V default kits? My advice would be to RMA those sticks and replace them with 1.8-1.9V kits. Run much cooler and less prone to die due to heat or voltage spikes, IME.

 

[error8]

Originally posted by: ConnyG
It has crashed/hung several times before that incident and the graphics card was firmly secured in the chassi and so was the cable/adapter to the card. I'd be very surprised if this damaged the GPU (it's not like I dropped an ACME 1 TON weight off my desk). I have tried re-seating the card. A bad cable would not explain why the computer totally hangs. As mentioned it appears to be 100% stable as long as I don't run any 3D-stuff/games.

I will try the different PCIE-x16 slots and DVI-ports today, as well as another graphics card, to rule out the motherboard.

If it appears to be 100% stable as long as you don't run any games and 3d applications is pretty clear that the card is faulty. I don't actually now what you've dropped on to the cards cable because you haven't really mentioned, but it should have been something big, because a coup of coffee or some news papers wouldn't hard lock your computer. Stuff can mean a lot of things and yes an ACME 1 TON is possible, who knows what you keep on your desk.

 

[ConnyG]

Originally posted by: chizow
The hanging in 3D apps points to the RAM going bad as those symptoms are identical to the problems I had when my RAM died on 3 different occasions last year. Let me guess, those Corsair are 2.1-2.2V default kits? My advice would be to RMA those sticks and replace them with 1.8-1.9V kits. Run much cooler and less prone to die due to heat or voltage spikes, IME.

Hmm... then how come I can run MemTest86+ without any problems? (that should stress the memory)

The Corsairs are 2.1V. Voltage is set to auto in BIOS.

 

[ConnyG]

Now this is just odd. As I mentioned I removed everything out of the chassi and assembled it on my desk yesterday and it still did not work.

Last night I did a fresh install of Vista on another harddrive and installed the Nvidia driver and it would crash when updating my "Performance rating" on the D3D Texture Load test.

Today I fired it up and it did the same. I tried the other DVI port and the same.

So I switched to the other PCI-E x16 slot and guess what, it ran flawlessly. Tried both DVI ports and it worked perfectly.

Then I switched back to the original PCI-E x16 slot and now it works like a charm.

I unplugged this other drive and plugged in my original OS and so far I have been able to start Overlord without it hanging.

Could it have been a poor connection and simply re-seating the card made it work?

 

[ConnyG]

Oh, and "stuff" was a body fat monitor which weighs exactly 252 grams. Not too heavy but half a meter of fall gives it some velocity.

 

[chizow]

Originally posted by: ConnyG

Originally posted by: chizow
The hanging in 3D apps points to the RAM going bad as those symptoms are identical to the problems I had when my RAM died on 3 different occasions last year. Let me guess, those Corsair are 2.1-2.2V default kits? My advice would be to RMA those sticks and replace them with 1.8-1.9V kits. Run much cooler and less prone to die due to heat or voltage spikes, IME.

Hmm... then how come I can run MemTest86+ without any problems? (that should stress the memory)

The Corsairs are 2.1V. Voltage is set to auto in BIOS.

Well, it might not be the RAM per se, could also be the # of dimms, chipset, etc. Have you tried with only 1 stick at a time installed? Memtest86 isn't really that great for stressing anything, it just reads/writes blocks of data and tells you if there were any errors. I've had Memtest tell me sticks were bad when they weren't (you can often simulate this with bad memory timings) and I've also had Memtest report no errors once, then report tons of errors without changing anything.

But ya like I said I've seen this problem 2x in both XP and Vista and both times it ended up being the RAM. Exactly the same symptoms that would make you think the video card was definitely the culprit, like lock ups at game splash screens, right? Swapped out my board and went to 1.8-1.9V default RAM kits and have not had the problem since. But ya I'd try only 1 stick at a time and see if the problem persists. Memtest can help you try and isolate a bad stick since you're running 4. If all 4 run ok, then try to run them in pairs, single and dual channel. If any of the 4 are bad individually then try running with the 3 good ones.

 

[chizow]

Originally posted by: ConnyG
Could it have been a poor connection and simply re-seating the card made it work?

Absolutely, although I think with the video card you would've run into problems much sooner. I know I recently had a problem like this with an IDE DVD-ROM drive where the cable wasn't fully plugged in. Drive would be seen in Windows but as soon as I tried to access it, it'd drop off after spinning for about a minute. So annoying, finally checked the cable, fully seated it and all was well again.