Solved

PC randomly restarts while gaming or stress testing GPU (NO overheat/NO Blue Screen)

My PC randomly restarts without showing Blue Screen, when gaming or stress testing GPU, basically doing stuff which involve using GPU.
Sometimes it doesn't happen, sometimes it does, there are games playing which never restart my PC, while some games restart PC in same spot everytime.
Sometimes FurMark runs without errors for an hour, while other time PC restarts as soon as FurMark runs Burn-In Test. GPU doesn't seem to overheat, it IDLEs at 49C, and MAXes at 76-77C.
I cleaned all radiators and fans, replaced thermal paste, no result.
Also, running Prime95 for 2 hours doesn't affect restarting, so i think it is not a CPU matter.

What is interesting is the behavior with some games:
Batman Arkham Origins - 1280x1024 high details with PhysX - mostly stable, restarts only in detective mode in Jezbell Plaza Fall case
NFS Shift 2 - 1280x1024, high details - restarts mostly on night tracks
Prototype 1 - 1280x1024, high details - restarts every time (no matter if i start new game or start exactly before this level) on level, where you are trapped in military base with couple of mutants and soldiers. Once i managed to pass this level and PC restarted 20-30 minutes later.
Sleeping Dogs - most of the time it run stable on low details with really bad lags (14-38 FPS), once i tried to benchmark on max details, PC restarted before benchmark started and after that trying to run any game results in restart)
Hitman Blood Money - 1280x1024, max details - NO restarts (!not tested after benchmarking Sleeping Dogs)
Dishonored - 1280x720, medium details - NO restarts, even after few hours of gaming (!not tested after benchmarking Sleeping Dogs)
other games that don't restart PC - Batman Arkham Asylum and Arkham City, Black and White 2, Dead Space 2, Fallout 3, Far Cry 3, Mortal Kombat IX, MotoGP 13, NFS Most Wanted (new one), Spec Ops The Line and even multiplayer Warframe - NO restarts (!not tested after benchmarking Sleeping Dogs)

Most of the time i play with game on my monitor and flash videos running on my TV hooked up to my PC.

My Rig:
CPU - Core 2 Quad Q6600 B3 (2.4 GHz/no overclock)
MOBO - Gigabyte GA-P43-E3SG (REV 1.0) BIOS - F14
GPU - Palit GTS 250 1GB E-Green Edition (its pretty much an underperforming utter sh*t)
RAM - 2x1GB Kingston KHX8500D2/1GN (Hyper X) + 1x2GB Kingston 9905316-152.A01LF
HDD - Samsung HD502IJ 500GB
DVD - LG GH20NS15 DVD+RW
LAN - D-Link DFE-530TX PCI Fast Ethernet Adapter (rev.C)
Multidisplay - Samsung LCD 42 (DVI to HDMI) + CTX730M (HDMI to DVI)
PSU - Tacens Radix II 520W (its modular PSU, looks well made)
UPS - APC Smart-UPS 700
CASE - old Sharkoon with front intake fan and rear exhaust fan
SYSTEM - Windows 8.1

Here comes the tricky part, I put GTS 250 on another rig (mobo - geforce 6100sm-m; CPU - Athlon X2 3800+; RAM - 1GB cheap china; PSU - 400W china) and it seem stable, both on FurMark and gaming.
So pretty much sure about faulty PSU i took old GeForce 6800 GS and put it in my PC to confirm the PSU fault, BUT i tested it with FurMark and some games (Hitman Blood Money, Sonic All-Stars Racing and World Racing 2) and no restart... so now I have no idea where should i look to find the problem.
13 answers Last reply Best Answer
More about randomly restarts gaming stress testing gpu overheat blue screen
  1. Instead a faulty PSU you might have insufficient PSU power at given power points, just something to consider... sorry I know it doesn't help you narrow it down any.
  2. APassingMe said:
    Instead a faulty PSU you might have insufficient PSU power at given power points, just something to consider... sorry I know it doesn't help you narrow it down any.


    Well, you are probably right, as 6800GS consume about 55W peak, while GTS 250 even as E-Green Edition (more economical version of the standard GTS 250) consumes much more i think (i don't know how much).

    However, a bad PSU doesn't explain how I could play Dishonored on medium (what used my PC resources to the limits i think) while playing video on second screen.

    Restarts just seem to happen on specific conditions, not depending on power consume.
  3. Do you have another machine to test the GPU in? The problem is likely either the GPU, PSU or the software/drivers. Try running prime95 at the same time as furmark to stress your PSU, you can do this with both GPUs.

    Looks like you've already tested the GPU in another machine. I still recommend running the furmark + prime95 to stress test your psu though. Looks like the suspects are probably down to PSU/drivers. You can do a fresh OS + drivers install if you want to test the latter.
  4. APassingMe said:
    Do you have another machine to test the GPU in? The problem is likely either the GPU, PSU or the software/drivers. Try running prime95 at the same time as furmark to stress your PSU, you can do this with both GPUs.

    Looks like you've already tested the GPU in another machine. I still recommend running the furmark + prime95 to stress test your psu though. Looks like the suspects are probably down to PSU/drivers. You can do a fresh OS + drivers install if you want to test the latter.

    Well, I just ran both tests (Prime95 on Blend and FurMark on Burn-In 1280x1024). While PC didnt restart I noticed that with both tests running GPU load never exceeded 38%, BTW I tested with 6800 GS.
    So still no definitive answer if PSU is really faulty. Do You have any other advices? Thank You for your time.

    BTW. If You can please un-select the best answer, as I accidentally selected it. I believe that there can be found a definitive solution, so I would not want the topic to be marked as solved. Thank You.

    Edit: I managed to hit 100% on CPU (Prime95) and 95-96% on GPU (FurMark), but it didn't cause PC to restart :??: Any suggestions?
  5. Can you run the tests on your test pc (not the pc you're having trouble with) and with the GPU in question to ensure that the GPU is fine. Once the GPU is ruled out though, it's still down to the PSU or Drivers.

    I would probably lean to the PSU as the source of the issue, but the Drivers will be a cheaper thing to check with a clean install and almost as likely the source of the issue. Regardless, unless you have money to burn, I would try a clean install as that could fix your issues if they are software related.

    In regards to the PSU being the issue, it's possible that some of the games causes the GPU to surge in power usage for a quick second which overpowers your PSU's ability to provide clean power, this would explain the not crashing during the burn-in (burn-in is continuous) and the repeatable crashes during games. It's also possible that the PSU is able to keep up with the lower power demand for your other GPU with no crashing issues while still being unable to provide the required power for your new GPU
  6. You could also try uninstalling all graphic drivers and re-installing the latest ones for your new graphic card, this has been known to fix similar issues before. But if you are going to re-install, then this would be done by default so.... it's up to you.
  7. APassingMe said:
    You could also try uninstalling all graphic drivers and re-installing the latest ones for your new graphic card, this has been known to fix similar issues before. But if you are going to re-install, then this would be done by default so.... it's up to you.


    You convinced me, that the PSU is most likely the problem. I thought about it a little more.
    PSU is Tacens Radix II Smart 520. It is dual rail - 12V1 (17A) and 12V2 (18A) and MODULAR:fou:


    Right now on modular ports I have connected:
    1. SATA Cable -> connected DVD+RW drive and a HDD internal case (currently without HDD and OFF)
    2. SATA Cable -> connected HDD
    3. MOLEX Cable -> connected only one front intake fan
    4. Empty

    While GPU is connected with dedicated 6pin (+2 pins unused) cable right from inside of the PSU.

    Maybe the problem is that one of the two rails is too overloaded? Is it possible?

    I tried to find which of 4 modular ports are connected on each of rails, without opening the PSU, but there is no info on this on the web.
    So I will probablly end up opening this "little devils box".
  8. First my recommendation is to never open a PSU up, Secondly... if you don open one up, don't touch anything. It's not worth the danger of get electrocuted and no it doesn't have to be plugged in to kill you.... there is residual charge so unless you know how to discharge it... don't do it.
  9. APassingMe said:
    First my recommendation is to never open a PSU up, Secondly... if you don open one up, don't touch anything. It's not worth the danger of get electrocuted and no it doesn't have to be plugged in to kill you.... there is residual charge so unless you know how to discharge it... don't do it.


    Can I check rails without opening it ? Is it posibble that rails are a problem or shouldnt I even mind this?
  10. Best answer
    Beyond that issue, I don't know how many amps your GPU pulls but it's probably a case of the PSU being slightly overrated more than anything else since I have never heard of that PSU company it's probably a cheap PSU that just meets the minimum requirements if that.

    I would recommend getting a quality PSU in the 600 to 650 watt range from a known manufacturer, personally I recommend a FSP sourced PSU but that's because I don't keep up with the other manufacturers to really compare and see which are the best and FSP is good enough for me. You can start here if you want to check for manufacturers

    http://www.tomshardware.com/reviews/power-supply-psu-brands,3762-5.html
  11. The only way that the rails would be the issue would be if you were trying to pull more than what it could supply through a given rail (shouldn't be an issue unless the PSU has problems considering your load). You might just want to upgrade to a more power efficient GPU and go that route as the 250 really is pretty old. You could go with a 750 ti and probably make your money back over the life of the GPU (assuming 24x7 on operation) in power savings.

    If this is a home-built machine (one that you intend to keep upgrading as time goes past), I would recommend looking into upgrading your PSU around the same time you upgrade your GPU if that ever happens.
  12. APassingMe said:
    Beyond that issue, I don't know how many amps your GPU pulls but it's probably a case of the PSU being slightly overrated more than anything else since I have never heard of that PSU company it's probably a cheap PSU that just meets the minimum requirements if that.

    I would recommend getting a quality PSU in the 600 to 650 watt range from a known manufacturer, personally I recommend a FSP sourced PSU but that's because I don't keep up with the other manufacturers to really compare and see which are the best and FSP is good enough for me. You can start here if you want to check for manufacturers

    http://www.tomshardware.com/reviews/power-supply-psu-brands,3762-5.html


    I will get a new PSU. Thank You for helping me :)

    One more little question, can I use my PC with this old PSU, until new PSU arrives or there is a possibility of damage to the components?
  13. That's a shot in the dark either way, it will probably be okay but there is a chance you could harm components if the PSU decides to get fancy on you. Your call.
Ask a new question

Read More

Gaming Blue Screen Overheat Shutdown Q6600 Nvidia Palit GTS 250 System Restart Games Geforce FurMark GPUs Stress Test