Help.. Intense applications causing reboot on close...

F

friend'scatdied

[H]ard|Gawd
Joined
Feb 20, 2005
Messages
1,283
Ok, this is my current setup:

MSI K8N Neo4 Platinum/SLI
512MB x 2 Dual Channel Corsair Value Memory
160GB SATA 7200RPM Seagate "Barracuda" HDD
Chaintech 6600GT (500/1000)
AMD Athlon64 3000+ Socket 939 Winchester
Windows XP Pro w/ SP1a/SP2 installed and fully patched.

Okay, this is my first personal build ever, and Windows XP is giving me some issues.

The first time I formatted and installed Windows, everything ran fine and drivers installed smoothly. I started by installing the Forceware 70.78 Quadro drivers as someone suggested to me and ran the games I could run in OpenGL. They ran OK.

However, then I tried the Painkiller full game that came with my 6600GT. It ran incredibly smooth and framerates soared. However, after I try to exit the game, without any notification from Windows whatsoever, the screen goes blue for a split second as the entire system crashes and auto-reboots.

Being quite worried, I tried uninstalling various drivers with no luck. Since I had disabled System Restore, I was hasty to reformat (although probably not the best option).

This time, I chose the quick NTFS setup and started it up. This time, I installed the drivers one by one. Games still ran fine at either D3D or OpenGL at this point, and didn't crash. I installed the drivers that came on the disk for my 6600GT. Again, fine. Then, as I installed Razer Diamondback drivers that came with my Diamondback, things started crapping out on me. I uninstall the Razer drivers and to my horror the crashes are still there post-game.

I system restore back to the previous pre-Razer state. Phew, the problem seems to be fixed now. However, soon I chance it and try the drivers again and the problem arises again. Then, I uninstalled and system restored back. This time, I wasn't so lucky as the crashes still continue. As I got the same post-game crashes, I did Norton and PC-Cillin checks and successfully removed one worm and two backdoors (the names of which I forget). The crashes still continue. After each crash I check the BIOS for CPU temp just in case. They turn up at about 40'C max, and my setup got at a peak of 43'C after 1 hour of the Prime95 torture test (I use a Zalman CNPS 7000B AlCu with Arctic Silver 5 so these temperature are underwhelming for the circumstances).

I highly doubt these are hardware problems and I really think these are moreso OS problems considering in the actual games performance is sky-high and just as it should be according to benches. It is only when I try to exit back to my desktop that crashes occur. Except for games, everything runs stably and smoothly (Firefox, AIM, Winamp, etc.). Spyware is probably not an issue as updated Spybot/Ad-Aware checks yield nothing. However, I threw down $800 for this setup to play games smoothly and I don't want to risk long term damage or stability issues with these crashes.

Anyone know some answers? I came here because I heard this was a great place for diagnosing hardware problems.. Thanks in advance.
 
Crashes like that are almost always traceable to hardware, even if the exhibit the problem in software. It could be a driver issue though. Do you have the latest chipset drivers installed?

You might want to run some memtest too. I shy away from any value type of RAM, though the Corsair is supposed to be OK if you aren't overclocking.
 
Yeah, I have the latest NF4 chipset drivers according to MSI's LiveUpdate. I haven't overclocked anything at all, and even when I lower all of the BIOS settings for the RAM (underclocking basically), the problems still persist (RAM is at 200MHz with 2.6V and CAS Latency of 2.5 right now). Everything is at the clock setting it was out of the box.

I'll try memtest as suggested though (software and BIOS tests ran just fine for the memory though I know memtest is more thorough and better anyway), but I don't understand how memory-taxing programs can run fine at one setup and die at another. Thanks for the input.

It's annoying because I came from a 2.2 GHz P4 with 512MB PC2100 RAM and even it survived post-game (although it performed horribly in-game).
 
I know that if the system encouters a fatel error it will reset automatically.



  1. Open system in Control Panel.
  2. On the Advanced tab, under Startup and Recovery, click Settings.
you will see auto reset.
 
A problem has been detected and windows has been shut down to prevent damage to your computer.
IRQL_NOT_LESS_OR_EQUAL
Technical Information:
*** STOP: 0x0000000A (0x9D086240,0x00000002,0x00000000,0x80S1EC1C)

Apparently, since the problem eventually occurs even after reformat, a reformat now likely won't fix the problem. Thanks for the fatal error thing though, now I know what the problem is.

Can anyone give me some steps on how to resolve this problem without needing a reformat, since it'll just come up again? I know how to access the BIOS and do registry stuff if needed, anyway. Thanks in advance again. I'll be glad to provide any information needed along the way.
 
well first things first, irq are set up in the bios and saved in the cmos. They are not saved on the harddrive. next we cant see what is conflicting because your device manager is conflicking right? startup in safe mode see if you can get into it then. or reset the cmos.
 
That STOP-message can be caused by a number of things, both software and hardware.
Does it say anything meaningful after IRQL_(etc) ?

According to google, it can be caused by everything from SP2 being unhappy, bad ram, or your PSU being too weak. (And a number of other seemingly random things.)


How did that memtest go?
 
It said some mumbo jumbo that if this is a persistent problem, I should reinstall my hardware one by one and change BIOS settings. I checked the Device Manager and the IRQs are fine.

From some references I take the hardware part with a grain of salt. It can really be anything, including bad drivers, though I won't rule hardware out. I think drivers triggered this anyway, since it didn't occur at all until I installed certain drivers.

Memtest didn't yield any errors.

Prime95 torture test runs alright for hours on the machine without causing the BSOD. It's supposed to be quite taxing on memory and processor as well.

If it was really some defective hardware, why would it crash only when returning to the desktop anyway? It's even crashed on the smallest thing.. I've had it crash once while trying the DX9 test in DXdiag.
 
If you're getting frequent CTD's when using 3D apps, it's probably related to your video card.
 
BEST JD TCR said:
well first things first, irq are set up in the bios and saved in the cmos. They are not saved on the harddrive.
Click to expand...

That's not actually true anymore. With the advent of Plug and Play and ACPI compatible BIOS's (and OS), Windows actually assigns the resources for the hardware. Back in the day, yes, the motherboard would assign resources. Or, today, if you don't have Plug and Play enabled in the BIOS. Note that this doesn't mean that you still might get a resource conflict from the board level (say, sound card occupying a slot below the video card) but that is more how the motherboard handles interrupts.

More reading here.

To the OP, this seems simliar to the problem that some people are having with Nvidia hardware and Athlon 64's. I think the answer was to "NoExecute" option in boot.ini to alwaysoff. Had to do with a conflict with DEP (data execution protection).

Try it out. Right-click My Computer > properties (or window-break) > advanced > startup and recovery settings > edit > change the line "multi(0)disk(0)rdisk(0)partition(1)\WINDOWS="Microsoft Windows XP Professional" /fastdetect /NoExecute=OptIn" to read /NoExecute=AlwaysOff

This will require a reboot.

That has apparently helped other people whose machines are crashing / rebooting on exiting a game.


 
Thanks, I'll keep that in mind. Actually, considering it's from full screen and back, that could certainly be it. Though, again, I won't rule out anything else. I'll check my BIOS and stuff.

I'm still befuddled on why it BSODs only after I try to return to the desktop? Why doesn't it just BSOD me when I try to enter the game or during the game? Why would it start after a particular set of drivers? This is what I don't understand about this particular one...

Thanks for the extra info. I'll try it out.
 
Here read this describing exactly your same problem (almost down to the same hardware) and its resolution as posted above. Also, here is what MS has to say about the noexecute switches.

Takes 30 seconds to edit and a few minutes to find out.

 
K, Yeah, I know UT2004 is getting the BSOD and PainKiller just "feels" Unreal-engine based (not sure if it is).

I did the DEP fix and also uninstalled my current video card drivers for the 71.81's, which supposedly have fixes for the problem.

Thanks a lot for the great and quick input, I hope I'm lucky in fixing this. *crosses fingers*
 
Looks like it uses the PAIN engine (coupled with, what else, Havok physics engine). But, the two could use some of the same hardware hooks that cause the problem? It's a stretch, but at this point...

 
arkamw said:
That's not actually true anymore. With the advent of Plug and Play and ACPI compatible BIOS's (and OS), Windows actually assigns the resources for the hardware. Back in the day, yes, the motherboard would assign resources. Or, today, if you don't have Plug and Play enabled in the BIOS. Note that this doesn't mean that you still might get a resource conflict from the board level (say, sound card occupying a slot below the video card) but that is more how the motherboard handles interrupts.

I was reviewing what I read in "Upgrading & Repairing PC 16th ed w/ scott muller".
Newer systems as well as plug & play operating systems, such as win 95b or later, windows 98, and windows 200/xp, all support a function known as PCI IRQ steering. Older system BIOSs and windows 95 or 95 a do not have support for pci irq streeing.

Generally, the Bios assigns unique irqs to pci devices. Even when irq steering is enabled, the bios still initially assigns irqs to pci devices. Although windows has the capability to change these settings, it typically does not do so automatically, except where necessary to eliminate conficks.

So all in all, we were both right; and we were both wroug.
Click to expand...
 
It worked! You guys own. :)

Thanks. It's quite devastating laying down a large sum of money on your first build and having it not work out right. Everything seems to be fine now, though, and I still get screaming performance.

I repeat, you guys own.
 
You must log in or register to reply here.
Back
Top