Game Causes Computer to Restart...

[cvq07]

When I load a craft and fly it, the game will play for about a minute and a half and then the computer will restart. The computer will only restart in game, most games, but not whilst web browsing etc. Any ideas on what to do? Battlefront II still works so if The Divion and KSP and Prison Architect and GTA still don't work, I've still got the best Star Wars game anyway. Thanks...

[cvq07]

Originally posted by RoofCatA:

overheating most likely.

Be sure you don't use overclocking. Check for dust on radiators/exhaust vents, coolers still rotating, voltage in bios (dying PSU can change voltage), temperatures in bios under load, and eventually thermal grease between chips and their radiators.

Vents cleaned already, no overclocking except for GPU fan speed, but any other solutions.

[Operation40]

it's almost always temps and/or voltages..
saying it's clean is one thing, but I'd check the actual numbers.. http://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.30.zip
checking in BIOS is ok, but the PC isn't really "under load" there.. Mind that the 12v line doesn't move more than 10% while in a full screen video game.

[cvq07]

Originally posted by Operation40:

it's almost always temps and/or voltages..
saying it's clean is one thing, but I'd check the actual numbers.. http://www.cpuid.com/downloads/hwmonitor/hwmonitor_1.30.zip
checking in BIOS is ok, but the PC isn't really "under load" there.. Mind that the 12v line doesn't move more than 10% while in a full screen video game.

Just checked on The Division, the CPU temps didn't even hit 50C before restarting

[cvq07]

Also the game was in windowed borderless mode.

[cvq07]

Originally posted by RoofCatA:

on very rare occasions there may be bad contact somewhere on RAM. Which may be cleaned. Just don't try to rub with wool. They are very sensitive to static. And that wouldn't have as predictable pattern I guess.

And then there is a chance of prank. It's also rather theoretical. Setting one would require skill and time.

Maybe your OC GPU cooler crashes it somehow? Like trying to go even faster when the temperature raises and then fails so badly at internal communication that some numbers overflow or whatever? Just a wild, wild guess.

Did you check the voltages - they should be 12V, 5V, 3.3V?
Even 10% deviation could make your system unstable. Those may change when computer warms up. Like cold start is all cool, and then after a while ingame it's already 13V. Don't expect them to be exact. They rarely are. But should be close.

Or just take it to a specialist. Distance repairs are complicated.

Voltages were lower than that, 12V, 4.4V and 3.2V. I tried Prison architect loaded up a 500 prisoner prisona nd it was fine, I do run mods in KSP though and on 64 bit, could that have anything to do with it?

[AoD_lexandro]

Power supply.

99% of the time, when a system just shuts itself off while gaming, it means the PSU is knackered. Almost certainly the thermal diode controller has died, and the PSU is overheating and shutting down. Thats why system temps all seem fine, as the PSU is never monitored by the system for temps.

Simple test, next time it shuts down touch your PSU backplate and test how hot it is. If its anything above skin temp, its knackered.

[cvq07]

Originally posted by RoofCatA:

4.4V sounds like a disaster. You sure it's not 3.3 typo?
If it's really 4.4V I would switch it off and leave that way before something expensive is burned down beyond repairs. PSU may need replacement in that case. It's not the most expensive part. Some voltage stabilization is done at motherboard level too so it may be involved eventually. Maybe just the voltage detector has broken down and fools the system into self sustain mode, but that's really beyond my security clearance level :) Best guess would be new PSU.

Or you may have some really crazy overclocking going on your system which you don't know about. But in that case (crazy OC) you would be either OC freak yourself or have enough money to pay for real service. Which leaves me out anyway :)

Firstly yes it was a typo.

Originally posted by AoD_lexandro:

Power supply.

99% of the time, when a system just shuts itself off while gaming, it means the PSU is knackered. Almost certainly the thermal diode controller has died, and the PSU is overheating and shutting down. Thats why system temps all seem fine, as the PSU is never monitored by the system for temps.

Simple test, next time it shuts down touch your PSU backplate and test how hot it is. If its anything above skin temp, its knackered.

Secondly, my PSU is a CX750M that is only two years old.

[cvq07]

Originally posted by AoD_lexandro:

Power supply.

99% of the time, when a system just shuts itself off while gaming, it means the PSU is knackered. Almost certainly the thermal diode controller has died, and the PSU is overheating and shutting down. Thats why system temps all seem fine, as the PSU is never monitored by the system for temps.

Simple test, next time it shuts down touch your PSU backplate and test how hot it is. If its anything above skin temp, its knackered.

Also I did the check, my PSU is still coldis when I touch it as it shuts down.

[cvq07]

Originally posted by RoofCatA:

... manufactured by CWT. I hope you didn't think Corsair makes PSU?
http://www.tomshardware.co.uk/forum/id-2547993/psu-tier-list.html
Manufacturer CWT is quite ok, but it's tier IV for some reason. Cheap end.

It doesn't mean though it wouldn't last 3..5 years or even more. Just educational.

Most large PSU sellers are bs their customers with their own brand name. Thermaltake, Silverstone, Corsair, EVGA, Antec... NONE of them actually manufactures any PSU. It's all bulk orders in China factories specializing in those.

Next time just don't pay premium price for bs name.



Also it may be completely different problem! We just told the most likely one.

Thanks, although I didn't get it on premium price, Black Friday.

[maj.solo]

Originally posted by RoofCatA:

overheating most likely.

Be sure you don't use overclocking. Check for dust on radiators/exhaust vents, coolers still rotating, voltage in bios (dying PSU can change voltage), temperatures in bios under load, and eventually thermal grease between chips and their radiators.

Yes it is some kind of PANIC reset. The OS + hardware together come to the conclusion it is F-ed up and it is so dangerous to continue that you can not even let the OS continue for fear messing up the system. I think it is called "panic reset".

[cvq07]

Originally posted by maj.solo:

Originally posted by RoofCatA:

overheating most likely.

Be sure you don't use overclocking. Check for dust on radiators/exhaust vents, coolers still rotating, voltage in bios (dying PSU can change voltage), temperatures in bios under load, and eventually thermal grease between chips and their radiators.

Yes it is some kind of PANIC reset. The OS + hardware together come to the conclusion it is F-ed up and it is so dangerous to continue that you can not even let the OS continue for fear messing up the system. I think it is called "panic reset".

How do I deal with this?

[maj.solo]

Originally posted by cvq07:

Originally posted by maj.solo:

Yes it is some kind of PANIC reset. The OS + hardware together come to the conclusion it is F-ed up and it is so dangerous to continue that you can not even let the OS continue for fear messing up the system. I think it is called "panic reset".

How do I deal with this?

I don't think you can deal with it. But I can add information.

The PANIC reset can be FALSE. All it takes is for a few bits of program code on the hard drive to have been corrupted. So that when the OS loads there is some damage to the code there. You run the computer and the damaged binary code is not used.

You surf the web, you do all kinds of things but the OS never goes through the damaged code. Then you start a program that do does use that device driver or that part of the device driver and suddenly pop!!! reset.

in THAT case uninstall / reinstall device drivers. But I / like you will not know where the fault is. I think you have to go online and see how one examine "dump files" on windows. And all kind of un-user-friendly ♥♥♥♥ like that. I think MAYBE you need to set windows to dump the entire ram on disk before, or ask the OS to ask the hardware to dump in some location. And the story goes on and .... you and me ..... do we have the enrgy for this .... I dont.

The individual programs when they crash can also create a dump file. But this is a hard panic reset.

Anyways.

1) since the crash comes when you start any advanced program it might be a creeping fault that is caused by temperature just as roofcat said. You start the program you look at the clock and BOOM.

2) My other possible explanation .... a FALSE reset coused by corrupted binary code caused by damage on the hard drive. Since it only happens when you start KSP then it might be damaged device drivers ... and those you can reinstall fresh and try. Usually when you reinstall the new drivers end up on OTHER sectors then they were before. Maybe a healthier part of the hard drive.

3) If that does not work .. sigh .... is it a total new OS reinstall AGAIN??? I did it 2 years ago FFS!!! not AGAIN ... grrr ... so you reinstall the whole thing ....takes less then an hour. Go through all Windows updates getting drivers, and video CODECS, your favourite programs updated while you are at it takes for me 1 week.

4) ACTUAL faulty hardware. All those GB of RAM we have these days. Even in the old days when the OS detects a fault on the hard drive ( common then ) and registers the bad sector in a fault list and stop adressing that part of the disk ever again. So writing to disk is always done by write reread check that it was written correctly.

But cosmic radiation can change your computer, it can even alter RAM content if the puter is turned on for long enough. Anyways, if damage occurs when the OS is not doing all these reread check and verification, the OS have then left the area flagging it is OK, and then it becomes corrupted, and then the OS gets so so so sooo very surprised!!

Anyways bad disk sectors goes into a fault list .... I don't remember if early computers could detect and flag memory as bad on the fly. I know early computers walked through the entire RAM at start and fixed it there. So any fault in the RAM should be covered.

Next is faults in MMU and other onboard chips bridges etc. It is supposed to be caught during boot self tests. I mean what do you think your puter does before it even loads windows? It is trying to find faults.

But in your case it lets the OS start and everything is fine and dandy. It is therefore sorta think it is a "computer load problem" OR "false reset by corrupted code". It sound like these two.

And then you fix that .... and then it was 1) aaaal along and the reinstall didnt fix ANYTHING.

5) Buggy device drivers ..... he he heh .-..... REALLY .... have it not been cought yet .....I always ASSUME all testing been done and everyone else have suffered and reported bugs .... way before I suffer. I use NVidia cards. And they get updated when MIcrosoft think it is time to update the drivers. So I always assume the entire world have already suffered and discovered the bugs.

Yes Microsoft will update your GFX card drivers after awhile. :) late ... yeah yeah late ... but it will happen. Let the 20 year old kids download the latest and rage on forums about FPS.

----

So how can one help you? I dont know. I can't help myself. It is some kinda hard reset that is for sure. Cause unknown. Examine some dump file needed. By someone.

It all takes time. And the sad part is if it is true hardware problem. Then it costs more then time then you have to open the wallet too.

[maj.solo]

Voltages are being lowered, they try to make the chips dissipate less heat. It is being done in steps slowly through the decades. What you could try is check the voltages at idle just after you login and run nothing.

make a note.

Then run something and see if they drop. Could be the power unit can be something else.

If you have a lead .... and there is a narrow section. It causes resistance. There is always a voltage drop behind a resistor. And the resistor warms up. A damage lead is one thing. But if voltage drop is reported at many positions in the computer then it must be the PSU.

[cvq07]

Originally posted by maj.solo:

Originally posted by cvq07:

How do I deal with this?

I don't think you can deal with it. But I can add information.

The PANIC reset can be FALSE. All it takes is for a few bits of program code on the hard drive to have been corrupted. So that when the OS loads there is some damage to the code there. You run the computer and the damaged binary code is not used.

You surf the web, you do all kinds of things but the OS never goes through the damaged code. Then you start a program that do does use that device driver or that part of the device driver and suddenly pop!!! reset.

in THAT case uninstall / reinstall device drivers. But I / like you will not know where the fault is. I think you have to go online and see how one examine "dump files" on windows. And all kind of un-user-friendly ♥♥♥♥ like that. I think MAYBE you need to set windows to dump the entire ram on disk before, or ask the OS to ask the hardware to dump in some location. And the story goes on and .... you and me ..... do we have the enrgy for this .... I dont.

The individual programs when they crash can also create a dump file. But this is a hard panic reset.

Anyways.

1) since the crash comes when you start any advanced program it might be a creeping fault that is caused by temperature just as roofcat said. You start the program you look at the clock and BOOM.

2) My other possible explanation .... a FALSE reset coused by corrupted binary code caused by damage on the hard drive. Since it only happens when you start KSP then it might be damaged device drivers ... and those you can reinstall fresh and try. Usually when you reinstall the new drivers end up on OTHER sectors then they were before. Maybe a healthier part of the hard drive.

3) If that does not work .. sigh .... is it a total new OS reinstall AGAIN??? I did it 2 years ago FFS!!! not AGAIN ... grrr ... so you reinstall the whole thing ....takes less then an hour. Go through all Windows updates getting drivers, and video CODECS, your favourite programs updated while you are at it takes for me 1 week.

4) ACTUAL faulty hardware. All those GB of RAM we have these days. Even in the old days when the OS detects a fault on the hard drive ( common then ) and registers the bad sector in a fault list and stop adressing that part of the disk ever again. So writing to disk is always done by write reread check that it was written correctly.

But cosmic radiation can change your computer, it can even alter RAM content if the puter is turned on for long enough. Anyways, if damage occurs when the OS is not doing all these reread check and verification, the OS have then left the area flagging it is OK, and then it becomes corrupted, and then the OS gets so so so sooo very surprised!!

Anyways bad disk sectors goes into a fault list .... I don't remember if early computers could detect and flag memory as bad on the fly. I know early computers walked through the entire RAM at start and fixed it there. So any fault in the RAM should be covered.

Next is faults in MMU and other onboard chips bridges etc. It is supposed to be caught during boot self tests. I mean what do you think your puter does before it even loads windows? It is trying to find faults.

But in your case it lets the OS start and everything is fine and dandy. It is therefore sorta think it is a "computer load problem" OR "false reset by corrupted code". It sound like these two.

And then you fix that .... and then it was 1) aaaal along and the reinstall didnt fix ANYTHING.

5) Buggy device drivers ..... he he heh .-..... REALLY .... have it not been cought yet .....I always ASSUME all testing been done and everyone else have suffered and reported bugs .... way before I suffer. I use NVidia cards. And they get updated when MIcrosoft think it is time to update the drivers. So I always assume the entire world have already suffered and discovered the bugs.

Yes Microsoft will update your GFX card drivers after awhile. :) late ... yeah yeah late ... but it will happen. Let the 20 year old kids download the latest and rage on forums about FPS.

----

So how can one help you? I dont know. I can't help myself. It is some kinda hard reset that is for sure. Cause unknown. Examine some dump file needed. By someone.

It all takes time. And the sad part is if it is true hardware problem. Then it costs more then time then you have to open the wallet too.

I only downloded GPU drivers, would downloading the ones for the motherboard be of any use, or the CPU etc. Again only the Division GTA and KSP don't work, oddly specific, would it be possible it could be down to the fact I think my version of windows was cracked. To clarify, I had built my computer and put windows on myself, but there was a hard drive issue and then went down to the shop had an SSD put in and had them install windows?