Tech Support Forum banner
Status
Not open for further replies.
1 - 12 of 12 Posts

·
Registered
Joined
·
8 Posts
Discussion Starter · #1 ·
Hi there,

I recently built a new PC and since then have been plagued with BSOD restarts when playing games (namely Skyrim, Saints Row 3, smaller less intensive titles don't seem to be affected e.g. minecraft, atom zombie smasher). It seemingly happens at random anywhere from at the start of the game, to a few hours in.

I have checked GPU temperature and that comes back at between 50and 60 degrees at the time of the crash which is apparently fine according to the card specs (max temp: 100 degrees).

I am running:
· Windows 7
· x64
· Retail version
· Hardware and OS < 3 months old

· CPU: Intel i5-2500K
· GPU: Nvidia GeForce 550Ti
· MotherBoard: ASUS P8Z68-M PRO
· Power Supply: CiT 750W

I have also included the autoruns output, even though the instructions said to download it but never said to run it.

Thanks in advance for any help,
Gaz
 

Attachments

·
Registered
Joined
·
2,210 Posts
I only have to look at a number of these to notice a very specific pattern. All of them seem to be occurring as TDR failures. This is when your video card drivers fail to respond to DirectX in a timely manner and will force DirectX to reset the driver in order to recover it. Often times this can work, though in worse cases it will attempt to do this multiple times in succession, tripping a BSOD after a few tries since it notices it hasn't resolved the issue.

In your case, however, this is not happening. Rather, when attempting to even reset the driver, that in itself has failed. It means the driver has absolutely no intention of stopping and Windows is not having a good time trying to fix it. As a result, there is an immediate BSOD.

You can review this article to get a good idea of possible causes of TDR failure and potential diagnostics and remedies you can do to ascertain and resolve the issue. If you want my immediate position on the matter, I've noticed all of these are occurring when dealing with the VirtuWDDM.sys driver, which to my understanding is the driver used to mix both GPU power of the Sandy Bridge with that of a discrete, separate video card (LucidLogix's Virtu). It's dated from June, so you may wanna update it. Otherwise, you may be forced to uninstall and not utilize it until they work out the kinks. Again, that's just immediate speculation, given the consistent pattern of this BSOD.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #3 ·
Hmm interesting, completely forgot about that useless tool. I installed that right at the start and it's never worked for me. Just got a cannot run with current GPU config error. Wasn't that bothered but at the same time didn't remove it.

I'll get rid of it and see if that fixes the issue, cheers for the theory and quick reply. I'll keep you updated.

Gaz
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #4 ·
Well I removed Virtu and am still experiencing problems, it seems the theory was wrong unfortunately. I'm now working my way through the article you mentioned.
 

·
Registered
Joined
·
2,210 Posts
Umm, I meant the ones you are experiencing now after uninstalling the Virtu drivers, but yes, I see that these do include those.

Just so you know, your latest crashdump is unrelated. Driver Verifier found that the MagicISO driver you have wasn't handling its resources properly (as in bad programming of driver code). Given that this driver (mcdbus.sys) is dated way back in February 2009, I can kinda understand why. You may wanna update it before it imposes any issues in the future (if it hasn't already).

As for the other crashdumps, I've actually found something very peculiar that passed my notice the first time I looked, that did show up in all your TDR crashdumps:

Code:
BugCheck 116, {fffffa8004d0c4e0, fffff8800fc7fae4, ffffffff[B]c000009a[/B], 4}
That c000009a is an NTSTATUS error code with the following description:

Code:
Error code: (NTSTATUS) 0xc000009a (3221225626) - Insufficient system resources exist to complete the API.
Apparently a particular resource has been exhausted, which is causing the inevitable crash. What resource could it be, though? Well, while the minidumps don't hold that information, fortunately, something else did: the system event log:

Code:
Event[3597]:
  Log Name: System
  Source: Microsoft-Windows-Resource-Exhaustion-Detector
  Date: 2011-12-01T17:45:07.101
  Event ID: 2004
  Task: Resource Exhaustion Diagnosis Events
  Level: Warning
  Opcode: Contains the results of the diagnosis.
  Keyword: Events related to exhaustion of system commit limit (virtual memory).
  User: S-1-5-18
  User Name: NT AUTHORITY\SYSTEM
  Computer: Gaz-PC
  Description: 
Windows successfully diagnosed a low virtual memory condition. The following programs consumed the most virtual memory: TakeOnHDemo.exe (405900) consumed 982044672 bytes, atkexComSvc.exe (1672) consumed 265945088 bytes, and Steam.exe (4204) consumed 242057216 bytes.
This is just one of the entries present. Evidently you have either your memory being exhausted and/or your paging file. This explains why this is only happening to resource-intensive games like Skyrim and not to smaller ones. I recommend you beef up your paging file and possibly even add more RAM to your system if necessary, or it's time to turn down your graphics settings (anything RAM-intensive like textures and whatnot).

I also want to add that after a bit of googling there's a possibility that atkexComSvc.exe has a memory leak. It's the ASUS Com Service. You should go in your services list and see if you can't find and squelch it. I don't see what purpose it would have to keep it active.

I do wanna warn that it may break Windows, so it'd actually be better to go through your programs list and uninstall anything that has anything to do with your motherboard, besides the chipset drivers. I notice other gimmicky mobo utilities installed on your PC that we here at TSF have found to have a notorious habit of crashing PCs and causing woe. It's best to clean them out.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #8 ·
Interesting, thanks a lot for looking into this for me. I'll try everything you suggested once these memtests are complete (all looking good so far).

It's still weird as I have 4Gb of RAM which I thought should be enough considering Skyrim for instance isn't even large address aware (or wasn't last time I played it) so can only use 2GB of it. And the page file size is a good couple of GB too as I've already considered that possibility.

I do have two more completely different sticks of RAM on order though so shall see if that's the problem or not.

Thanks again,
Fingers crossed this will finally solve it.
Gaz
 

·
Registered
Joined
·
2,210 Posts
The Skyrim process environment itself may be 32-bit (doubtful), but even so if it were an exhaustion of its own environment it would present an application crash, as it's a user-mode application. It is not just the Skryim process, but that along with other processes that are expanding in tandem with each other and eventually exhausting all your memory. The result is that the GPU no longer has anything to work with, and is being pressured to acquire more when it already has nothing. This will lead to the TDR failure.

So it's not the issue with Skyrim lacking resources, but rather your GPU. Furthermore, the error isn't so much, "You don't have any more resources." Rather, it's, "I lack the resources I need to complete this operation." Skyrim or something else may be requesting a large amount of resources in one go, which cannot be satisfied given the current resource constraints already present. You may even have 1 full GB free then all of a sudden Skyrim is asking the GPU to start loading a whole bunch of world textures and when the GPU tries it goes, "I'm sorry, but there's not enough resources to satisfy your entire request." That one request alone could exhaust all that memory in one fell swoop.

I do want to add that 4GB is actually pretty much the norm nowadays. Even on my workstation I already take up 2 GBs from using Windows 7 and Firefox with 30 tabs. I can imagine Skryim being far more hungry than that, especially when you start bumping texture quality up.
 

·
Registered
Joined
·
8 Posts
Discussion Starter · #10 ·
After a lot of searching I found something else that may be the issue. Seemingly completely unrelated to the memory issue but want to check it with you anyway.

Apparently Nvidia recommend a "minimum +12V continuous current rating of 24A", whereas my PSU only gives 22A. How much of a problem is this likely to cause?? It seems like such a small amount, but might account for BSODs when the system is working hard.
 

·
Registered
Joined
·
2,210 Posts
Might be. 2 amps is a very big difference. If you want, you can also check voltages using HWInfo and see if there's heavy voltage fluctuating (close or more than 5% skew up or down for 5v, 3.3v or 12v rails). You can do logs and all by clicking the Sensors only option when you start HWInfo.
 
1 - 12 of 12 Posts
Status
Not open for further replies.
Top