[SOLVED] Restarts while gaming

[r3ddr4g0n]

Hi, I've been trying to troubleshoot what is going on with my system.

S:
Chief Complaint: The system restarts when I'm playing video games, and rarely when I'm not playing any games. When it restarts, the screens say "no input detected" and the case fans go to max speed. It doesn't reboot or post, it just stays like that until I hit the reset button.

HPI: The restarts began mid-October.
The timing of the restart is random but usually happens pretty soon after rendering in a game.
Once it restarts, it usually will restart a few more times until I get frustrated and stop trying to get it to run.
The times that it restarted outside of a game, it was just after gaming.
It has restarted before even getting to Windows boot screen.

History: I started this build about 10 years ago so pretty familiar with all the components, and I've replaced/upgraded as components broke/became obsolete. In 2017 the old ASRock mobo failed, I replaced at that time. In January 2020, the CPU cooler mounting bracket broke spontaneously and the cooler fell off, I replaced the bracket. I upgraded from Windows 7 to Windows 10 in July. Has been running stable since.

O:
Mobo: ASUS M5A97 R2.0

CPU: AMD FX-8350 Black Edition
Cooler: ZALMAN CNPS9500A

GPU: GIGABYTE GeForce GTX 1060

Drives:1x Intel 320 Series 2.5" 120GB SATA II as Windows partition
2x WD Caviar Black WD1002FAEX 1TB in RAID 1

RAM: G.SKILL Ripjaws Series 8GB (4 x 4GB)

PSU: CORSAIR TX Series CMPSU-750TX 750W

Screens: LG 43LF5100 43"

Tests completed:
Complete breadboard outside of case - booted and restarted during a game.
Mobo - BIOS is up to date, no settings changed recently.
CPU - removed CPU cooler, cleaned, re-seated, and reapplied thermal paste. It ran games for a day and then restarted.
GPU - Tried the other PCI-E 2.0 slot, and replaced with an EVGA GTX 1060, and this ran games for a day before going back to restarting every game. Instead of going black screen on restart, it would post and boot, might be how that GPU handles restart?
RAM memtest - all 4 sticks and each individual stick passed.
PSU paperclip - all rails reading correct voltages confirmed by voltmeter.
Software - this occurred across multiple games (NMS and SWTOR), and I wiped drivers with DDU and re-installed. No changes.

A:
I suspect that this is either a failed mobo or a CPU/GPU thermal issue, so I ran Open Hardware Monitor and tried to make it restart. When it did, the GPU was never higher than 65, north/south bridge temps were fine, but the CPU temp readings were all over the place, as low as 6C (is that even possible under load?) up to 25C. I'm wondering if the motherboard temp sensor failed and the CPU isn't getting enough cooling because it's reading low.
All the fans are working properly, even GPU fans which I set to max and still had a reset.
Windows Event Viewer doesn't have anything other than an unexpected restart, but I noticed that there are hundreds of paging operation errors for the SSD. This has been happening since July, and I have my system page file on the RAID array, so I don't think this is an issue.

P:
Happy to perform any other diagnostics. Wondering if I should just buy a new MOBO/CPU and hope the problem fixes itself.

 

[Flayed]

Hi, I've been trying to troubleshoot what is going on with my system.

S:
Chief Complaint: The system restarts when I'm playing video games, and rarely when I'm not playing any games. When it restarts, the screens say "no input detected" and the case fans go to max speed. It doesn't reboot or post, it just stays like that until I hit the reset button.

HPI: The restarts began mid-October.
The timing of the restart is random but usually happens pretty soon after rendering in a game.
Once it restarts, it usually will restart a few more times until I get frustrated and stop trying to get it to run.
The times that it restarted outside of a game, it was just after gaming.
It has restarted before even getting to Windows boot screen.

History: I started this build about 10 years ago so pretty familiar with all the components, and I've replaced/upgraded as components broke/became obsolete. In 2017 the old ASRock mobo failed, I replaced at that time. In January 2020, the CPU cooler mounting bracket broke spontaneously and the cooler fell off, I replaced the bracket. I upgraded from Windows 7 to Windows 10 in July. Has been running stable since.

O:
Mobo: ASUS M5A97 R2.0

CPU: AMD FX-8350 Black Edition
Cooler: ZALMAN CNPS9500A

GPU: GIGABYTE GeForce GTX 1060

Drives:1x Intel 320 Series 2.5" 120GB SATA II as Windows partition
2x WD Caviar Black WD1002FAEX 1TB in RAID 1

RAM: G.SKILL Ripjaws Series 8GB (4 x 4GB)

PSU: CORSAIR TX Series CMPSU-750TX 750W

Screens: LG 43LF5100 43"

Tests completed:
Complete breadboard outside of case - booted and restarted during a game.
Mobo - BIOS is up to date, no settings changed recently.
CPU - removed CPU cooler, cleaned, re-seated, and reapplied thermal paste. It ran games for a day and then restarted.
GPU - Tried the other PCI-E 2.0 slot, and replaced with an EVGA GTX 1060, and this ran games for a day before going back to restarting every game. Instead of going black screen on restart, it would post and boot, might be how that GPU handles restart?
RAM memtest - all 4 sticks and each individual stick passed.
PSU paperclip - all rails reading correct voltages confirmed by voltmeter.
Software - this occurred across multiple games (NMS and SWTOR), and I wiped drivers with DDU and re-installed. No changes.

A:
I suspect that this is either a failed mobo or a CPU/GPU thermal issue, so I ran Open Hardware Monitor and tried to make it restart. When it did, the GPU was never higher than 65, north/south bridge temps were fine, but the CPU temp readings were all over the place, as low as 6C (is that even possible under load?) up to 25C. I'm wondering if the motherboard temp sensor failed and the CPU isn't getting enough cooling because it's reading low.
All the fans are working properly, even GPU fans which I set to max and still had a reset.
Windows Event Viewer doesn't have anything other than an unexpected restart, but I noticed that there are hundreds of paging operation errors for the SSD. This has been happening since July, and I have my system page file on the RAID array, so I don't think this is an issue.

P:
Happy to perform any other diagnostics. Wondering if I should just buy a new MOBO/CPU and hope the problem fixes itself.

I would suspect either the GPU or PSU for the restarts. Memory can cause restarts but not just during gaming usually. Plus you have run memory diagnosics. The CPU temp reading are suspect though although not sure if that would cause the issue. Perhaps if you can set the CPU fan to run at a fixed RPM rather than auto based on temperature that would help.

Does it restart during a run of prime95? If it doesn't that would point towards something other than the CPU

 

[r3ddr4g0n]

I would suspect either the GPU or PSU for the restarts. Memory can cause restarts but not just during gaming usually. Plus you have run memory diagnosics. The CPU temp reading are suspect though although not sure if that would cause the issue. Perhaps if you can set the CPU fan to run at a fixed RPM rather than auto based on temperature that would help.

Does it restart during a run of prime95? If it doesn't that would point towards something other than the CPU

Gotcha. I tested the PSU and found all the voltages to be normal, and I swapped out a different GPU card and had similar symptoms. I ran Prime95, it didn't reset while running Prime95, but it did throw a LOT of errors. I'll re-run all the mem tests, I've had bad sticks pass before. Any other ideas based on the above?

 

[Flayed]

Errors in Prime95 indicate a problem with the CPU. It is very unusual to get errors in Prime95 if the CPU is not overclocked

 

[r3ddr4g0n]

Errors in Prime95 indicate a problem with the CPU. It is very unusual to get errors in Prime95 if the CPU is not overclocked

Gotcha. I had a feeling. Would the CPU overheating cause that? When I took off the thermal paste, it was extremely sticky, almost glued on. Not like the previous times I've re-seated the CPU. Thank you for the help tho, I really appreciate you helping me add another symptom to diagnose this.

 

[Flayed]

No I don't think it is due to heat. CPU's are normally very good at math so when their sums don't add up something is very wrong. I'm thinking it would be something like a failed memory location inside a register of the CPU maybe.

Or perhaps the CPU is unstable at stock voltage/frequency due to age. You could increase the voltage to the CPU to see if that would stabalise it.

Otherwise you are looking at having to replace the CPU

 

[r3ddr4g0n]

I did more testing, I ran Prime95 again for 15 minutes with no errors. I have no idea why it threw a ton in at the beginning. I ran Prime95 + FurMark for 10 mins, max CPU usage all cores and max GPU usage hoping that I could force a reset from a dying PSU, also no luck. I tried installing nVidia drivers from before this all started happening, no luck.

Today when I booted up, it said I needed to restart to fix drive errors in the SSD. Since the OS is on the SSD, I'm wondering if the SSD is failing causing resets when DirectX or some other OS program is used. Weird that there isn't any other instability when not running a game.

One thing that helped is that I can run a game for 5-30 minutes now if I turn v-sync on.

 

[r3ddr4g0n]

Got a working solution so far. I reinstalled another GTX 1060 and didn't use nVidia's drivers, instead I let Windows automatically install a driver and it installed an older version from 7/24/2019. Haven't had a crash since but will update if it happens again.

 

[Flayed]

A different 1060? So maybe it was the graphics card after all

 

[r3ddr4g0n]

A different 1060? So maybe it was the graphics card after all

It happened again after playing for a few hours. So weird. I tried the same routine I did, wipe drivers, install new GPU, and let Windows install the drivers. Crashed before it even finished loading the game.

 

[Flayed]

It happened again after playing for a few hours. So weird. I tried the same routine I did, wipe drivers, install new GPU, and let Windows install the drivers. Crashed before it even finished loading the game.

Another new GPU? Do you have a box full of 1060s or something?

 

[r3ddr4g0n]

Another new GPU? Do you have a box full of 1060s or something?

Nah just two. But the issue happens with both of them. Still can't tell if this is a failing PSU, motherboard, or now possible driver issue.

 

[saltyGoat11]

Nah just two. But the issue happens with both of them. Still can't tell if this is a failing PSU, motherboard, or now possible driver issue.

If your computer restarts, I would assume it's hardware and not software (software issues like drivers will mostly show blue screens). I've seen a lot of incidents where the memory would cause random restarts (although what you describe isn't random since it happens only when gaming).
SO, stuff to try: you have 4 mem sticks, try removing a pair and check with only 8GB see if you can reproduce.
PSU - I'd borrow / buy with intent to return just to make sure it's not it.
GPU - Since you tested 2 with the same results although you used a different driver, I'd assume that's not the issue.
Temps: 6 degrees sounds very low and inaccurate. try to reset the bios (jumping the board) (instructions how to for your motherboard)

Good luck my friend. keep us posted

 

[r3ddr4g0n]

If your computer restarts, I would assume it's hardware and not software (software issues like drivers will mostly show blue screens). I've seen a lot of incidents where the memory would cause random restarts (although what you describe isn't random since it happens only when gaming).
SO, stuff to try: you have 4 mem sticks, try removing a pair and check with only 8GB see if you can reproduce.
PSU - I'd borrow / buy with intent to return just to make sure it's not it.
GPU - Since you tested 2 with the same results although you used a different driver, I'd assume that's not the issue.
Temps: 6 degrees sounds very low and inaccurate. try to reset the bios (jumping the board) (instructions how to for your motherboard)

Good luck my friend. keep us posted

Thank you! I'm interested in any ideas anyone has for tests. I reset the BIOS using the instructions, and while I was in the BIOS I noticed that my 12V rail was reading 11.6 in the BIOS monitor. I busted out the voltmeter and it showed 12.4V. Both of those are within tolerance, but I don't know why there's two different readings.

Then I tried 2x 4gb sticks of RAM. Crashed with both sets. But I had joined a discord and noticed that the sound kept going, and I was still connected, for quite a while! I was able to talk to people while the screens were off. So I need to google those symptoms because it isn't a true "restart".

EDIT: This is super weird. With one graphics card, the screens say no signal but the computer still plays audio and I can speak through discord. With the other graphics card, the whole system resets.

 

[Flayed]

You need to narrow this down a bit somehow as it stands it could be anything. I'd be tempted to boot off a linux usb stick just to see if the issue happen outside of windows which would at least eliminate a driver or windows cause.

 

[saltyGoat11]

EDIT: This is super weird. With one graphics card, the screens say no signal but the computer still plays audio and I can speak through discord. With the other graphics card, the whole system resets.

My guess would be: the PSU either not feeding enough or stable current to the motherboard/GPU
I'd try a different PSU just to make sure.

 

[r3ddr4g0n]

You're right. I'll test it with a different PSU and let you know the results. Thanks both for helping me figure this out.

 

[r3ddr4g0n]

You need to narrow this down a bit somehow as it stands it could be anything. I'd be tempted to boot off a linux usb stick just to see if the issue happen outside of windows which would at least eliminate a driver or windows cause.

My guess would be: the PSU either not feeding enough or stable current to the motherboard/GPU
I'd try a different PSU just to make sure.

I figured out what it was. After re-testing the memory and trying each stick individually, I noticed that the crashes got a lot worse when I put them back in. I noticed the motherboard had flexed a lot when I put them in and the motherboard ATX 24 pin connector was loose. I had re-seated the 24-pin connector when I breadboarded the computer, but when I wiggled it while the computer was on, the computer immediately reset, and I noticed that the fans went higher or lower depending on if I wiggled the 24-pin connector. So I re-seated it firmly and made sure the clip was in place, and haven't had a reset since 11/7.

 

[r3ddr4g0n]

Just another update, the crashes returned a few weeks later. Ultimately I grew suspicious of the PSU, despite passing the pin test. I bought a new one and replaced it, and haven't had a crash in over a year.