PC turns off and auto-restarts while playing a game

eamonx

TS Rookie
Hello,

I was playing a game on Steam called Final Fantasy XII Zodiac Age. The game is not supposed to be so intensive since it's a port of a Playstation 2 title.
But anyway, as I play it, usually after 15-40 minutes, my computer suddenly switches off instantly, then boots up. I thought either the CPU or GPU was overheating, so I ran a temp monitor program while running the game. CPU remained in the ~35 degrees region while GPU hit a max of 70 degrees. The monitor also showed that the CPU wasn't heavily loaded, but the GPU was - which I believe means the game isn't CPU-intensive.
Correct me if I'm wrong but anything under 70 for CPU and 85 for GPU is supposed to be healthy, right?
Nothing else triggers this problem. I've played much beast-ier games like Resident Evil 2 remake back in Jan/Feb and no problem.

My spec is:
- Intel Core i7-6700K
- Noctua NH-D14 CPU fan cooler
- 16GB(4GBx4) G.SKILL Ripjaws V DDR4 3000MHz
- MSI Nvidia GTX 1080ti Gaming X
- 1TB SSD
- 4TB HDD
- Asus Z170-A Motherboard
- Corsair HX1000i PSU

I haven't done any overclocking at all. BIOS is at default settings. I've read that despite being a PS2 game, FF12 is badly optimized and requires a high spec to run on PC.
I have a suspicion that power is the issue here. This may be my ignorance speaking here, but could it be that the GPU is drawing more power out of the PSU than normal, and a 'fail-safe' is getting triggered to prevent it going over the limit?

Thanks in advance!
 

Cycloid Torus

Stone age computing - click on the rock below..
System looks balanced and more than sufficient for your game. Temps look fine.

Test RAM overnight with Memtest86 - continuous, rack 10 or so runs.

"Nothing else triggers this problem." - Remove and reinstall game.
 

eamonx

TS Rookie
Thanks for helping.

I removed and reinstalled the game. After starting it up, my PC once again switched off after literally 2-3 mins in. So that eliminates the possibility of corrupt files I guess?

Tonight I'll run Memtest86.
 

Cycloid Torus

Stone age computing - click on the rock below..
"So that eliminates the possibility of corrupt files I guess?"

Umm, but it increases the likelihood that the program you are running is 'broken'. If you find another game/program which crashes then it might be hardware. The RAM test is to try to reduce the possibility of hardware issue further.

A quick look at messages indicate 'crash' is frequently found with 'Final Fantasy XII Zodiac Age' - so it might be just the program. Check 'Reliability History' to see if it is a recurring issue.
 

eamonx

TS Rookie
Sorry for the late response.

Ran Memtest86 overnight and came up with 0 errors.
I also tested a different game (Resident Evil 2 remake) and played for an hour. No problems occurred.

How do I check 'Reliability History' exactly? I'm a bit confused.
 

Cycloid Torus

Stone age computing - click on the rock below..
Enter "Reliability History" in search box and have a look. There is a 'more detail' capability which tells what is messing up and where.

Keep playing other games for a bit. Sounds like a software glitch in the crashing program.
 

eamonx

TS Rookie
Okay, so you mean the windows program "Reliability Monitor" in settings?

Well, in Critical Events, it shows as "Windows" with summary "Windows was not properly shut down." When I click on "view tech. details" the only extra info is when it refers to the last time Windows was not properly shut down.

Steam or the game's exe is not shown at all in Reliability Monitor.

I'll try playing some other games and report on any outcomes. But so far, you're right, it looks like the problem is directly caused by the software itself.
 

Cycloid Torus

Stone age computing - click on the rock below..
When a game crashes, it shows up - so what we have may just be a plain old conflict in the program - which is going about its business that includes "oh, bumped into a problem, wrote over the contents of memory, oops , time to reboot"

Seems like most likely explanation - cause could be a bad driver, but which? - wouldn't a bad driver cause other problem - why this specific program? - need a log to tell - try Administrative Tools, Event Viewer and look at Applications - might show up there.
 

eamonx

TS Rookie
In Event Viewer, I check the logs in Applications and Services... There's only a few sub-sections (GoProDesktopApp, Hardware Events, Internet Explorer, isaAgentLog, Key Management Service, Microsoft, Microsoft Service Alerts, Windows PowerShell).

None of these have logged any errors that occur around the same time as the Windows shut down.

If the issue might be a bad driver, then could it be due to the Nvidia drivers? I hear Nvidia is notorious when it comes to their video drivers because many times it releases a new version, it causes more problems than it fixes.

What do you reckon I should do next?
 
Last edited:

Cycloid Torus

Stone age computing - click on the rock below..
Play other games until you find one which crashes, then we can compare. Possibly lurk in the game forum for similar reports. Since you downloaded a fresh copy of the game, it seems the issue is inside that program as written.
In Event Viewer also check Event Viewer (local) > Windows Logs > Application
 
Last edited:

eamonx

TS Rookie
Hi, I thought I'd reply back to this thread after ages due to finding a significant discovery.

I believe I have found the reason for the PC turning off. I have had similar problems when playing a recent game called Shenmue III. During various moments, the PC would just instantly switch off, so I'd lower the performance settings to get through them. However, I reached an unskippable cutscene where I could not get through even with the lowest settings.

So I went about searching online for a real solution to this problem and apparently someone had found that this issue was due to the way you use 6+2 pin PCIe power cable(s) from the PSU to the GPU.

The guy had used a single PCIe power cable with dual 6+2 pin connectors to fit into the 2 6+2 pin slots on the GPU. The same configuration I had.

He changed this. Instead, he used TWO 6+2 pin PCIe power cables to fit into the slots. So rather than using one single cable with dual connectors, he used two cables. Apparently, that solved the issue of PC switching off during games.

I just tried this now, and I managed to get through the unskippable cutscene in Shenmue III - even on the highest performance settings. I believe I have finally solved the problem!

I'll try some other games, and will report back if I encounter the same problem again. But if I don't ever write here again, consider this problem permanently solved! :)
 
  • Like
Reactions: Cycloid Torus

Cycloid Torus

Stone age computing - click on the rock below..
Even though the PSU has plenty of juice, the 8-pin and 6-pin power requirement of the GPU could overwhelm it if it (and other things) crowded a single rail. The other possibility is a loose connection at the PSU connection of the modular wires which can creep loose over a couple of years of on/off cycles. The other possibility is the PSU setup which says in specifications, "Adjustable Single/Multi 12V Rail Yes", is not properly setup for the GPU.
 
  • Like
Reactions: eamonx