Computer shuts down under load/gaming - tried most fixes

[Michael_H]

Hello all - my recent upgrades have turned my build into a nightmare, the type of story I always assumed "couldn't happen to me".

The problem:
While gaming (or sometimes under load, usually with a movie playing), the computer will shut down and restart. No blue screen, no warnings, just gone. It will reboot and tell me that Windows did not shut down correctly.

I can browse the internet and type all day long, but for example when playing Just Cause 2 or Dragon Age the computer will shut down after roughly 20-25 minutes.

My specs:
ASUS P8P67 Motherboard
i5-2500k (not overclocked)
Sapphire ATI 5770
G.SKILL Ripjaws X Series 16GB (4 x 4GB) 240-Pin DDR3 SDRAM DDR3 1600
CORSAIR Builder Series CX600 V2 600W
COOLER MASTER Hyper 212 Plus
ASUS DVD Burner
Rosewill R5604-TBK Case
2x 7200 RPM hard drives

What I've done so far:

• First I thought it was a cooling issue, assuming the system was shutting down from overheating during gameplay. I installed the 212+ in lieu of the stock cooler and watched temps obsessively. My CPU runs cool even during Intel Burn Test (never crests 60C), and the shutdowns still happen.
• I ran memtest86+ for 19 hours with no errors.
• I then thought my power supply wasn't strong enough (cheapie 400W job) and upgraded to a beefy 600W, 40A on a single 12V rail supply. The shutdowns still happened.
• I had the motherboard replaced under warranty. The shutdowns still happened.
• I had the video card replaced under warranty. The shutdowns still happen.

I'm at a loss and getting frustrated. I've sunk a lot of time and money into this computer so I could use it for gaming, and that doesn't seem to be in the cards. What other steps can I take? If any more detail is needed, please let me know.

Some background, if it helps:
I built the first iteration of this machine about 2.5 years ago, it was an AMD Athlon 64 X2 3.0 GHz, DDR2 memory. I updated this summer to a new motherboard, Intel chip, and gobs of new RAM, planning on doing the other components eventually. "Eventually" turned into "part of the troubleshooting", and here we are.

 

[Brainonska511]

Could your video card be overheating?

 

[Michael_H]

I ran Just Cause 2 for a while, and some quick alt-tabs to GPU Shark showed ~60°C. I also ran Furmark for about 5 minutes at my usual gaming resolution of 1920x1080, and it seemed to be pretty solidly at 72°C. (I'm also now a little seasick.) Are those normal ranges for a graphics card?

 

[Brainonska511]

I ran Just Cause 2 for a while, and some quick alt-tabs to GPU Shark showed ~60°C. I also ran Furmark for about 5 minutes at my usual gaming resolution of 1920x1080, and it seemed to be pretty solidly at 72°C. (I'm also now a little seasick.) Are those normal ranges for a graphics card?

Sounds normal. Do you have the latest Catalyst drivers installed? Any critical errors appearing in the Windows event viewer (besides "unexpected shutdown")?

 

[Michael_H]

Sounds normal. Do you have the latest Catalyst drivers installed? Any critical errors appearing in the Windows event viewer (besides "unexpected shutdown")?

The Catalyst drivers are up date, even the latest hotfix that I believe they pushed in the past few days.

The Critical events are all listed as Event ID 41 "Kernel-power", which says "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly."

The Error message logs have corresponding events that say "The previous system shutdown at 4:58:32 PM on ‎6/‎13/‎2011 was unexpected."

Looking further through the Error messages, these seem to line up with the crashes:

Event ID 1060: "\SystemRoot\SysWow64\drivers\pfc.sys has been blocked from loading due to incompatibility with this system. Please contact your software vendor for a compatible version of the driver."

I'm not sure what that means, but the fact that these errors line up with the crashes might mean something?

(Everything else aside, I had no idea Event Viewer existed. So I learned something today.)

edit: Windows (7, if that matters) is also completely up to date.

 

[Michael_H]

Related to the above error message, I found the following:

http://extended64.com/forums/p/1126/2184.aspx

There someone mentions that pfc.sys is also listed as part of a freezing (though not crashing) problem, and it was installed by Adobe Audition. I have a pretty old version of Audition (1.5) that I use for work and is one of the first things I install on a machine.

Does this sound like something that could be the culprit?

edit: Also, on some occasions the computer has hung and frozen, rather than shut itself off. Usually, however, it restarts.

 

[Brainonska511]

edit: Also, on some occasions the computer has hung and frozen, rather than shut itself off. Usually, however, it restarts.

Sounds like Adobe Audition installed a driver of some sort that's causing system instability. I'd suggest uninstalling Adobe Audition and see if that solves the problem. If not, maybe try this: http://crypt-sdk.blogspot.com/2010/10/windows-7-pfcsys-problem.html (Warning: Editing the registry can result in OS damage; consider making a restore point before executing any changes)

 

[BoomerD]

http://www.mydigitallife.info/analy...using-windows-suddenly-reboot-via-whocrashed/

 

[Michael_H]

Those are all great resources but did not seem to sort out the issue. I've had several crashes in the past few days. Adobe Audition was uninstalled and I tried all the other fixes mentioned, and the computer shuts down.

Today it has been doing something else odd - while playing Dragon Age the video locked up and I heard a loud buzzing noise coming from the speakers (not the computer itself).

Any other diagnostics I can try?

 

[C1]

You build systems differently than I do. After a totally new build, I test the system thoroughly between incremental installs of major pieces of software. For example, after a new Win 7 install, does the system run/perform reliably? If not, it's easier to diagnose/fix than after adding all kinds of complicated software.

In your case now, I would just keep uninstalling things until stability is achieved then work from there.

If there is a question about heating, buy a brick/block of dry ice, put it in a paper bag and put that inside the computer case (will last 24 hours) and see if the problem still exhibits. Lay the case on its side to ensure that the cold air off the ice continuously flows over the components. Smaller pieces can be used to cool just selected components such as video or memory for pin point diagnostic purposes.

 

[sygyzy]

Could your video card be overheating?

I would put my money on this. I had the exact issue as you described and it only happened when I taxed the graphics card. In other words, things were fine in Windows 7, Chrome, etc, but if I fired up Left 4 Dead or something, it would crash almost immediately. The problem is an overheating graphics card. I was using an aftermarket cooler and I guess somehow it must have come loose or something. I inspected it and it looks perfectly fine (no gaps between chips and heatsinks). I bought a new card and everything works perfectly.

 

[Michael_H]

I would put my money on this. I had the exact issue as you described and it only happened when I taxed the graphics card. In other words, things were fine in Windows 7, Chrome, etc, but if I fired up Left 4 Dead or something, it would crash almost immediately. The problem is an overheating graphics card. I was using an aftermarket cooler and I guess somehow it must have come loose or something. I inspected it and it looks perfectly fine (no gaps between chips and heatsinks). I bought a new card and everything works perfectly.

I'm open to this being the case. It's a bit puzzling, though: the video card and case are unchanged, that is, I've used them for a year with no problems at all. The problems started when I upgraded my CPU/Motherboard/RAM while retaining the same GPU/case/PSU. The PSU has since been upgraded. I didn't notice any higher than normal temps during benchmarking (see above post - about 75°C is about all I saw)...

 

[Michael_H]

You build systems differently than I do. After a totally new build, I test the system thoroughly between incremental installs of major pieces of software. For example, after a new Win 7 install, does the system run/perform reliably? If not, it's easier to diagnose/fix than after adding all kinds of complicated software.

You'll have to forgive me if my methodology seems poor; I'm "new" at this. I may have gotten lucky with my first computer build a few years ago - I put everything together and it worked perfectly. I had no reason to do anything but gleefully use it however I wanted.

When I upgraded my motherboard/CPU/RAM, I used it the same way and it was then that I started running into issues. For future builds I will probably go a little bit slower, but this has been my "learning experience", so to speak.

 

[classy]

I am betting bad OS drivers. I would do a clean install then run some games. Also what kind of sound card drivers are you running? This sounds like an OS/software issue.

 

[Michael_H]

A fresh Windows install might be in order, after which I can carefully monitor what gets installed I suppose. Steam first, of course.

For sound card drivers I am using the video card, and it's listed a "AMD High Definition Audio Device". Device manager says the drivers are up to date. There's also "Realtek High Definition Audio" listed, and those drivers are also up to date. I don't use the digital out on my motherboard though - perhaps I can get rid of the "Realtek Audio Manager" in my system tray.

Oh, I was also poking around in CCC to examine fan speed settings on the GPU per the suggestion of overheating, and saw that Catalyst Overdrive was enabled. I disabled it to see if that helps any.

 

[C1]

Im inclined to agree with "classy".
And yes, I would not entertain multiple audio management applications for the obvious reason of a possible conflict. Not knowing about the AMD HD, I personally would go with the more known quantity (ie, RT HD).

Ive had cases where it depended in which order application software was installed (ie, A before B versus B before A). Otherwise driver conflicts would occur. Ive also had cases where uninstalling a problematic application would not eliminate the problem caused by its driver. Hence, my preference for testing with incremental installs (particularly applications involved with audio/video such as games).

 

[Michael_H]

I would put my money on this. I had the exact issue as you described and it only happened when I taxed the graphics card. In other words, things were fine in Windows 7, Chrome, etc, but if I fired up Left 4 Dead or something, it would crash almost immediately. The problem is an overheating graphics card. I was using an aftermarket cooler and I guess somehow it must have come loose or something. I inspected it and it looks perfectly fine (no gaps between chips and heatsinks). I bought a new card and everything works perfectly.

I'm not going to say that I have absolutely nailed this down, but you guys seem to have been on to something. I investigated an overheating graphics card as one of my very first steps, and it seemed to be okay. Given the suggestions here, though, I decided to check it out further.

In order to track down this potential graphics card heating issue, I knew I would need an elegant, subtle solution. After a few hours and some brainstorming sessions, I'm pretty proud of what I came up with:

Who knew? Just Cause 2 was a known offender, and would consistently cause a crash after about 15-20 minutes. Using my potentially-licensed-to-NASA solution I was able to run for well over an hour before stopping of my own volition, something that hasn't happened in months.

I will try this for another couple of days, after which I will start looking into better fans, or perhaps a case with more fans/better cooling. (This one only has two, one in the front and one in the back.)

I'm not too familiar with aftermarket cooling options for graphics cards, if such a thing exists.

 

[Brainonska511]

Who knew? Just Cause 2 was a known offender, and would consistently cause a crash after about 15-20 minutes. Using my potentially-licensed-to-NASA solution I was able to run for well over an hour before stopping of my own volition, something that hasn't happened in months. I will try this for another couple of days, after which I will start looking into better fans, or perhaps a case with more fans/better cooling. (This one only has two, one in the front and one in the back.) I'm not too familiar with aftermarket cooling options for graphics cards, if such a thing exists.

Was it only under intensive GPU activity? (for example, you could run a processor/RAM stress test and it wouldn't crash). There is a heatsink on the motherboard northbridge which could not be getting enough airflow. But it certainly seems like the graphics card is the culprit. Just that the big fan solution doesn't necessarily narrow it down.

If your card is overheating though, maybe you could try for an RMA with the manufacturer.

 

[Michael_H]

I could consistently reproduce the crash with a heavy graphics load, while heavy CPU load by itself never caused an error. I do a lot of audio encoding/processing which is all CPU, and ran lots of CPU stress tests without issue. The only RAM test I ran was memtest86+, and that ran fine.

This is already an RMA'd card, so I'm thinking perhaps the case/fan setup I have now just isn't getting enough fresh air to the card?

 

[Brainonska511]

I could consistently reproduce the crash with a heavy graphics load, while heavy CPU load by itself never caused an error. I do a lot of audio encoding/processing which is all CPU, and ran lots of CPU stress tests without issue. The only RAM test I ran was memtest86+, and that ran fine.

This is already an RMA'd card, so I'm thinking perhaps the case/fan setup I have now just isn't getting enough fresh air to the card?

Could be the case design limits the amount of cooler air reaching the card. Also could just be bad luck too - just two faulty cards...

 

[Michael_H]

Could be, that's just my luck. Thanks for all your help!

 

[exdeath]

I second video card or motherboard.

 

[Michael_H]

I had a thought:
Here's my layout. According to my P8P67 manual, I'm supposed to install a single graphics card into PCIe slot 1 (top in this picture), "for better performance". If I were running two cards in Crossfire, then I would use PCIe slot 2 (bottom) as well.

I put a rectangle roughly over where my graphics card lies, and note that the ass end of it goes right over the P67 chipset with a teensy amount of clearance.

Could that be overheating and causing shutdowns? If I move my single card to slot 2 to increase airflow over the chipset, does that mean I'll lose performance?

Here's what it says in the manual, I have to admit I don't understand it fully:

1) PCIe 2.0 x16_1 slot [blue] (single at x16 mode)
2) PCIe 2.0 x16_2 slot [black] (at max. x4 mode, compatible with PCIe x1, x2 and x4 devices)

Thoughts?

 

[Puppies04]

Seeing as it is a 5 minute job and you already know how your system performs why not try it and see .

 

[Brainonska511]

1) PCIe 2.0 x16_1 slot [blue] (single at x16 mode) 2) PCIe 2.0 x16_2 slot [black] (at max. x4 mode, compatible with PCIe x1, x2 and x4 devices)

That makes it sound like the black slot isn't a true x16 slot. If it is only wired to be an x4 slot, you're card will lose 75% of its current bandwidth.