Certain games power-off computer after new graphics card - need help!

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510
Hello,
My build is as follows:

PROCESSOR: i5 4670k (not overclocked)
VIDEO CARD: Gigabyte GTX 970 (G1 Gaming edition)
MOTHERBOARD: MSI Z87-G45 Gaming
MEMORY: Corsair Vengeance 8gb DDR3 (CMZ8GX3M2A1600C9)
PSU: Corsair VS 650
OS: Windows 8.1 64 bit
I also have a Samsung 840 PRO SSD and an old HDD that I forget the name of (no packaging laying around)

So here's my problem. Whenever I play some very specific games, I will get up to a certain point or play for a certain time, and then my computer will power-off. No warning, no BSOD, nothing that indicates the cause of the failure. There are only 2 games I have tried that do this: Metro: Last Light and Call of Duty: Modern Warfare 2.

While I had already looked up the issue, a few people had experienced the same issue as mine, but most involved overheating. I have checked temps before and directly after the issue and they seem to be perfectly fine. It is bizarre, because I can play much more demanding games, such as Arma 3 and Skyrim (modded with all those graphic injectors) perfectly fine without issues for hours, but I can only play MW2 and Last Light for minutes before a power-off.

I don't think the power supply is the problem, but there is definitely a chance.

I did, however, run across a certain forum that brought up a subject of interest. It was detailing the RAM's compatibility with the motherboard, and involved the QVL. Out of interest, I checked if my model of RAM was qualified with my motherboard, but it wasn't. However, the RAM's involvement in this issue is questionable, because my issues have only started AFTER the installation of my 970, replacing my 560ti.

So here I am, quite unsure of the cause of my issues. Could it be any kind of GPU - RAM incompatibility? Could it be my operating system? Perhaps my PSU is not up to the task? Or could it simply be that my Graphics card drivers are not 'optimised' for use with older games?

Any help is truly appreciated.
Thanks,
Brennon.
 
Solution
have you used benchmark to see if stressing the card resets the computer?
msi kombustor?
if it doesn't reset, then i would stress test the ram and see if thats the problem.
also have you checked your hdd for bad sector? or do you have a ssd
also are you oc the card? i was running into some problems with tdu2 with some oc on the 970 and i realized after hours of crashes, i reset to standard clock on the 970 and played for hours after that no problems once the oc was removed.
i was also getting the driver stopped responding nvidia managed to recover, so thats what made me remove the OC.





chasehornby55

Reputable
Jun 23, 2014
142
0
4,710
you should monitor your temps in the game because it could be a heat spike then the temps settle when you clog back on. use msi afterburner and install the remote server as well so you can monitor the temps on any smart device on the same wifi network while in game. You could also try this if your comfortable, you could either increas the cards power a little bit and see if it becomes stable or reduce the core and memory clock speeds and see if that helps. do the temp checking first though.
 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510


Unfortunately I have no device that is capable of hosting the remote server (unless of course it actually CAN be run on a PC), so I am going to have to need some kind of alternative to that :??: . Thanks.
 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510


I have friends that own compatible devices so I may have to organise that with them. May take a bit of time, but thanks for that suggestion :)
 

scorpionadiga

Reputable
Jan 9, 2015
26
0
4,540


Since there is no indication of failure. It is not an overheating problem or high voltage contact.. So I would assume its 99% chance that the problem is with your PSU.. either its a fail unit...... or time worn out it's efficiency.... Your Gpu requires high power (Watts) relatively.. So if I am right the PSU is prioritizing the GPU with not enough power to energize other components... So just try another PSU to check.. (This issue is most common when changing GPU's)

If that doesn't work out... I would suppose its either your Rams failure... or your HDD or SSD where your windows run !
 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510




I was also thinking that. My current PSU is a corsair VS 650. I am not entirely sure if the PSU actually meets the requirements for the gtx 970. I had originally bought the 650 PSU in preparation for an older graphics card, the gtx 770, which probably would have been fine, but it was a bit of an oversight when I got my 970. Is a 650 watt PSU adequate?
 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510


Cheers for the reply RobCrezz.

Would you kindly be able to suggest some PSUs that would be sufficient for gaming? I might consider getting a new one in the near future. It is bizarre however, that the problems only occurred AFTER the installation of my new graphics card. Can there be a problem due to that? I was able to play the games perfectly fine even with the same PSU but older GPU...

Thanks.
 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510


My original graphics card was a Gigabyte GTX 560ti.

Thanks for the suggestions for the PSU.
 

pharohbender

Distinguished
Apr 20, 2009
12
0
18,520
have you used benchmark to see if stressing the card resets the computer?
msi kombustor?
if it doesn't reset, then i would stress test the ram and see if thats the problem.
also have you checked your hdd for bad sector? or do you have a ssd
also are you oc the card? i was running into some problems with tdu2 with some oc on the 970 and i realized after hours of crashes, i reset to standard clock on the 970 and played for hours after that no problems once the oc was removed.
i was also getting the driver stopped responding nvidia managed to recover, so thats what made me remove the OC.





 
Solution

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510


Cheers for the suggestions. That MSI Kombustor looks neat, I will have to try that out. Also will have to check for bad sectors, I run an SSD with windows 8.1 installed on it with a HDD that I use for extra space. Be aware that I have both Last Light AND MW2 installed on the HDD, so perhaps I might try moving them to the SSD.

As for the overclocks, I have tried downclocking by around 100mhz for both core and memory clocks to no avail. However it is interesting to know that I would get bluescreens with my 560ti when playing the BF3 campaign. I downclocked to the original gtx 560ti clock (the card was factory overclocked) and the bluescreens stopped.

Thanks for help so far guys!
 

pharohbender

Distinguished
Apr 20, 2009
12
0
18,520
are you running something like msi afterburner, ramp up the fan speed keep the temps lower or create your own fan setting in the program so when gaming the fans don't stay to low as for nvidia's profile the cooler the better.


 

quantumsheep0

Reputable
Dec 22, 2014
8
0
4,510
Sorry for the long reply guys, been busy with school.

Anyways, I figured I might play around a little bit more with the clock speeds just to see if something would happen. Turned down the clock speed to the stock Nvidia 970's clock speed, and tried it out. Turns out, at default boost clocks, my 970 hates MW2 and Last Light, just as my 560ti hated BF3.

Thanks all for your contributions, as it probably would've taken me a lot longer to figure out the problem without the help. Now I can play the games I want without a worry! I will keep this thread bookmarked for future reference.

Thanks again!
 

Bratushka

Distinguished
Oct 16, 2010
17
0
18,510
I had similar issues as you describe. For checking GPU temperature I used a free gadget with my Windows 7. I had two video cards in SLI and monitored each. My particular brand/model of GPU would go into protect mode at 206 F. I found this out from the GPU mfg. site. The failures which ran the gamut from going into power Save mode and shutting down to crashing the game. My cards were eVGA and I used their GPU control software called precision X. While you can tinker with all kinds of stuff I was only interested in manual control over the GPU fan speeds. I'd run the fans as fast as the software allowed which was about 85%. Default speed was about 45%. Long term I had a video card that was slowly failing but speeding up the fans kept things going for a year or more. There are a number of these utilities out there and I believe each GPU manufacturer has their own version and all are free. Some of the aftermarket ones may not be. You'll have to look.

At one time Crucial.com was the best source for checking RAM in your PC. They have a utility that runs that will tell you what RAM you need, how they need to be configured -all the same size or mixable-, the maximum capacity you MN can handle, etc. They sell RAM which is why their utility is so good. You don't have to buy from them but the info you can glean will be accurate. You can double check if you have any issues.

On the power supply, yes too weak an output can cause issues. Some PCs, a couple of my old Dells were like this, have the mounting holes in opposite corners which makes using a power supply other than theirs difficult without some body work. I don't know if they are still like that or id everybody has standardized. I know that was a big pain when I started upgrading with more powerful GPUs because at the time Dell's biggest PSU was 250 watts. I long since abandoned off the shelf PCs because there was too much of that "nobody's stuff will fit but ours". Again, check the GPU's manufacturer data sheet for power requirements. If you can't make sense of it call their support line and ask what kind of power capacity do they recommend. Bear in mind if you are running a lot of USB devices that don't have their own power supplies/adapters the PC's power supply is driving them too which puts an additional load on it.

Another thing I remember was having an ATI GPU that requires two power supply connections but you weren't supposed to use them from the same line. You were supposed to use one plug from one line and another from a different line.

I think your clocking adjustment is disguising something else that's going on because in briefly poking around I didn't see others were having an issue with the games and that card. Also, since it's possible to wreck stuff messing with GPU settings without fully understanding what the settings do I wouldn't think game manufacturers would mainstream market a game that wouldn't work with a common video card without doing things like that. Games are usually designed for the broadest range of hardware and for the least tech-savvy of players while both GPU and games can include the neat stuff that get the more technically inclined excited.

But, whatever works!