Question PC reboots randomly in game

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Sep 29, 2022
18
1
15
Hello!


Here is my problem :


.Symptoms : the PC reboots only in game

-It’s always a hard reboot (without BSOD), without an error message before nor after the reboot. I have this problem since i built the PC, around march 2021

-The frequency of the reboots seems different on every game : in worst cases, it can be 6/7 reboots within 20 hours of gameplay, while in other games, i had only one reboot within 100 hours. Moreover, it’s possible to suffer from 3 reboots during the first 10 hours of gameplay, than not a single one for the 10 hours after. The problem doesn’t seems to be more common with the most taxing games (i had more reboots with Gloomhaven rather than with Cyberpunk 2077…).

-Reboots can happened between 5 min and 2 hours of gaming

-The problem seems vaguely cyclic : it’s possible to not have a single reboot for one or two weeks, then suddenly to have one every two days.


System :

OS : Windows 10 problème

CPU : AMD 5600X

MB : ASUS Prime B550M-K

GPU : Inno3D Geforce RTX 3070

RAM : GSKILL Aegis DDR4 16go (3200mhz)

PSU : Antec Neo Eco 700W Gold, then Bequiet Purepower 80+ Gold 700W


State of the PC :

-Bios and drivers are up to date. I did some tests last year with the drivers of that time, with the same results.

-All Bios settings are unmodified, no overclocking

-The PC has been formatted last year, a few months after the issues began.

-No overheating problems

-The PC and monitors are plugged on the same system than my previous PC, which had no problems. I tried to change the power strip, without any result. I also use the same devices as before (M/K, speakers…).

-In power plan settings, minimum processor state is set to 5%, max at 100%.

-I don’t remember any lighting strike close to my apartment


Windows event logs:

For every reboot, I have a Critical Error Kernel Power, with a Cache Hierarchy Error (Machine Check exception).


Tests:

-OCCT: I did every single test for 1 hours: no reboot /error.

-Memtest86: no reboot/error

-Windows Memory Diagnostic: nothing

-CHKDSK: nothing

-WhoCrashed: no result (no valid crash dump has been found)

-Windows Defender and Malwarebytes: nothing

-Prime95: no reboot, but two errors, after 28 and 30 min (fatal error: rounding was 0,5, expect less than 0,4). No more errors after this (1 hour test). I don’t know if it’s enough to really suspect that the CPU is the culprit.


What I tried:

-Reseat everything and change the power supply cable

-I changed my PSU: no differences

-I used two diffrent PCI-E cables for my GPU

-I desactivated XMP profile


What I couldn’t try:

-Test another GPU/MB/RAM, simply because I don’t have anything else available

-Replicate the problem: I know that some games have more chances than others to cause a reboot, but it’s still possible to have a few hours without any issue


I have to admit that I don’t see any consistency with these results. At first, I thought that it was an overheat problem (it’s obviously not), then that it was a faulty PSU, but now I’m just completely lost.


If Someone has an idea to solve this problem, thanks in advance!


PS: English is not my first language, sorry for the mistakes
 
Sep 29, 2022
18
1
15
unplug the hdd from system and did you check the error code prime 95 give to you on test .

I will try to unplug the HDD as soon as possible.

Considering Prime 95: there are actually 2 messages: "Fatal error: rounding was 0.5, expected less than 0.4", and "hardware failure detected runningf 1120K FFT size". It's the same error for both threads that failed.

I will try another test with Prime 95 to see if i can reproduce the issue.

I have may have something new: i had 3 crashes in 2 days with Darktide, so i tried to monitore the game with OCCT and HW Info. And luckily (...) i had another reboot 20 min after.
This is OCCT, 10 min before the crash:
https://www.dropbox.com/s/jkqtbpx903s8dv6/Darktide 1.PNG?dl=0

I also have HW Infos Logs:
https://www.dropbox.com/s/yws91pg5qhcpame/Logs HWinfo Darktide.csv?dl=0

I hope it helps...
 

scout_03

Titan
Ambassador
as soon as cpu hit 70 c system drop what cooler you have on it check if it not jiggling when fans ram up and last time you dust off system 1 would try a repaste and test temp again other then that you will have to strip system to minimal part since prime says a defective hardware maybe the files you use to test in prime are to big for cpu l3 cache .
 
Sep 29, 2022
18
1
15
as soon as cpu hit 70 c system drop what cooler you have on it check if it not jiggling when fans ram up and last time you dust off system 1 would try a repaste and test temp again other then that you will have to strip system to minimal part since prime says a defective hardware maybe the files you use to test in prime are to big for cpu l3 cache .

Nothing is jiggeling, and i already repasted the cpu. Temperatures seems very reasonable to me.
Considering Prime 95 results, i'm not sure that i understand your point. If i understand well, it said that the CPU has 2 faulty threads. I tried to test it again, with the same results: threads n 2 and 6 are victims of harware failure during the process.
 
Sep 29, 2022
18
1
15
to bad we would not be able to solve the issue unless someone else have a clue .

Well, it seems that this type of problem is really hard to diagnose. I saw a lot of people on other forums having the same problem, and it seems that it's either a PSU issue, or just the processor.
I wish i could test every component with another PC, but it's just not possible. I see one last solution: RMA.