Issue with PC freezes during games

[Guinibee]

I have this post in a different forum and someone suggested my video card(s) could be unstable... I figured I could get more ideas here.

My specs:
4790K OC 4.5ghz
GTX 980 Classified SLI OC’d +75 core/+100 memory
Asus Rog Hero VII
16GB Ram
Asus Rog Swift Monitor

I randomly keep freezing when playing certain games (as of late Payday 2, Heroes of the Storm and CSGO). Either the game freaks out and the side monitor and or main monitor start flashing black and white than either I hard freeze with a sound loop and need to reboot or the game minimizes and I have to shut it down via task manager. I really don’t know what’s up with my computer and I’m concerned considering I have a rather expensive build. All these parts are very new…got them all around xmas time. I’d be highly surprised if anything was a lemon or statically shocked as most of the time everything works fine and when gaming I have no issues.
The first things people mention are the OCs generally. I have run Prime95 26.6 for 6 hrs straight and been 60C max, so the OC is stable and not overheating. I also never froze during that test. I ran Firestrike Extreme with my GPUs never going over mid 60s and during games only get in the 50C range. Overheating and the OC’s should be fine.

What I have tried to fix it:

- Messed with refresh rates on the monitor to make sure I’m always 144hz in game
- Re installed video drivers
- Added more power to GPUs on msi afterburner

What I have on my list to try in order to fix it:

- Clock video cards to stock
- Clock CPU to stock
- Diable SLI
- Disable Steam Overlay (read this affects some people)
- Disable nvidia geforce experience (read this also affects some)
- Wipe SSD and reinstall windows
- Buy new mobo/cpu/ram??? Really don’t think this is necessary…

Does anyone have any other ideas to help me out? I’m really upset with what is happening and would like to fix it.

Thanks

 

[americanbrian]

try a custom fan profile to increase the GPU fan. It could be that the cooler is very effective for the GPU, but not the GDDR. IF the fan stays very low the heat may build up on the GDDR while the GPU stays cool.

Anyway, it is easy to check by just setting the fans on super high (90%+) and seeing if that solves this issue.

 

[JamesSneed]

Your first few troubleshooting steps are spot on. This is typical of an overclock issue so I would right off the bat put the CPU and GPU back to stock speeds like you mentioned. If that doesn't do the trick then I would remove one of the graphics cards as your way of disabling SLI but also completely removing the hardware from causing any issue. If that doesn't do the trick then swap the card out that's left in the PC with the one you had removed. This is of course to determine if you have a bad graphics card.

I expect the OC on the GPU's to be the cause of the crashes.

You didn't mention your power supply but overclocking multiple GPU's along with your CPU has some pretty high power demands. For 20 bucks you can pick up a "Kill A Watt Electricity Usage Monitor" and plug your PC into that to see whats its drawing from the wall which will give you an idea on how much strain your putting on your PSU. If the PSU is having large voltage ripples due to barely handling the power output then disabling the overclocks may also fix your issue by alleviating some of the power demands.

 

[Guinibee]

To answer the fan profile, yes I have it cranked to 50% and a custom curve to keep increasing the hotter it gets.

The PSU I use is a brand new EVGA 1000W Platinum which I got shortly before all the other parts...I know I def have enough power and EVGA makes really high quality PSUs, especially rated at Platinum...I can't imagine this is the issue.

Now here's my issue with the OC's....if I go ahead and set the cards back to stock and I stop crashing that's great, but I have GTX 980 Classies which are built to be OC'd. If I can't OC them without crashing then wouldn't these ones be faulty in a way? Should I contact EVGA about it and see if I can RMA them? I know I'm jumping ahead, but I'm thinking I could test them one at a time with an OC to see if I freeze.

 

[Guinibee]

last night when I took off the OC on the cards I was able to game 3 hours straight with no freezing issues. This is good and bad IMO...

Now ik the problem, but considering the classies are built to be ocd and I can even roll with such a minor oc, wouldn't that indicate that one of these cards is somewhat defective? I'm thinking I should run them individually ocd and see if one of them fails and then rma that one ... Would this be a good next move?

 

[JamesSneed]

I agree with your thought process. Certainly shouldn't be your PSU in this case. I bet one or worst case both cards cant tolerate the overclock. Your idea of testing the OC one card at a time is a good plan. Personally I would have done that at the start meaning to OC each card then set my final OC to both cards to the lowest common denominator. These cards do already come with an OC from the factory so it may just be that "stock" OC is all they can do. I don't own these but maybe you can can search the web for others experiences.