Alienware M17x R3 - Dedicated Video Card Dead!

So, I've been experiencing blue screens since April, at maybe twice a month in frequency, up until August.  From August until now, I was getting BSODs 3-5 times a month.  I couldn't figure out why because the error messages were obscure.  I've been trying different video card driver version as well, with no good results.

I finally downloaded WinDbg, a debugging tool, and used it to read the minidump files that the system compiles after a BSOD, and saw this:

Problem signature:

  Problem Event Name:        BlueScreen

  OS Version:        6.1.7601.2.1.0.768.3

  Locale ID:        1033

Additional information about the problem:

  BCCode:        116

  BCP1:        FFFFFA800CCFE4E0

  BCP2:        FFFFF88002C1501C

  BCP3:        FFFFFFFFC000009A

  BCP4:        0000000000000004

  OS Version:        6_1_7601

  Service Pack:        1_0

  Product:        768_1

Files that help describe the problem:

  C:\Windows\Minidump\091815-12682-01.dmp

  C:\Users\ron\AppData\Local\Temp\WER-30763-0.sysdata.xml

This error could be generated by a multitude of issues, such as motherboard problems, software problems, operating system problems, or hardware issues.

I opened up my system and cleaned it of dust (there wasn't all that much, considering I've not opened it to clean it -- ever).  I also removed the heat sinks from the CPU and graphics card, cleaned off old thermal paste (Dell just dumped a ton of it on each part without care).  I then re-applied new thermal paste and also added 4GB of RAM since I had it open.

I tried gaming again, after removing all the Nvidia drivers and reinstalling them, but was still getting BSODs.  What's funny is that the drivers disappear and I have to reapply them every time I get a BSOD, but still have to use a driver uninstaller to do it.

In looking at my logs for the month of October, I've experienced 8 BSODs so far and it's only halfway through the month.  Also, the logs say "Video hardware error", specifically:

A problem with your video hardware caused Windows to stop working correctly.

So, I guess my Nvidia GeForce 580M finally died.  Apparently, people have been having problems with them since 2012.  I guess I got lucky...I'm only now experiencing issues, 3 years after people began to notice they fail.  I shouldn't complain, I guess.

Now I need to determine which would be a good replacement card.  I've seen people using 680Ms in their M17x R3s without issue.  Some people have even used 780Ms and 880Ms, but those are pricey...I'm not ready to spend $800 on a video card.  I can either get another 580M and hope it gives me another 3 years or I can get a 680M (can find them new for $320-$360 on eBay, with warranties).  I can't find them as cheaply on Amazon.

For now, I've configured War Thunder (that's the game I'm currently playing) so that I can use the integrated graphics card (an Intel HD 3000).  It works but my frame rate dove down to 30-40fps.  :(  I suppose I can either try to do War Thunder on my iMac or use the desktop system I was using before I bought the Alienware.  I also have a Dell Latitude E5530 that has an Intel HD 4000 that might get better frame rates.  I do have options.

I'll update this blog entry as I decide what to do next.