Shutdown after lengthy gaming sessions

Soldato
Joined
19 Apr 2012
Posts
5,183
My brother has reported that he is getting shutdowns (hard reboot) during lengthy gaming periods. I'm not at home so looking at the best way to test this by me giving him verbal indicators.

From what I remember his setup is:

1600AF
MSI B450 tomahawk Max
RX580 GPU
Case is MSI mag vampiric (I think but will confirm)

I think if it were a PSU issue, it would happen fairly quickly. In starting to think it's heat related once the case is getting up to temperature.

Should I get him to run a CPU and GPU benchmark test to get the temps up quicker? Something like prime95 and furmark at the same time?
 
Associate
Joined
8 Feb 2014
Posts
846
Location
Aberdeen
One thing I've done before is run with the case off. It's not conclusive of course but could be a step in the process.

Using something that records cpu/gpu temps might help too.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
One thing I've done before is run with the case off. It's not conclusive of course but could be a step in the process.

Using something that records cpu/gpu temps might help too.

Yeah that's a good idea. He did say one night it lasted an hour or so but last night managed 5 hours before it started.
I think it's down to the increase in temps around the house with the nice weather.
 
Associate
Joined
18 Oct 2002
Posts
1,553
Does the case only have one rear fan?
I run a ryzen 3700x and the other day I gamed for 6 hours in the man cave where the temp was 35 c and my CPU topped out at 78c.
But I do have 5 case fans - perhaps an extra fan or two for the case would help.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
Does the case only have one rear fan?
I run a ryzen 3700x and the other day I gamed for 6 hours in the man cave where the temp was 35 c and my CPU topped out at 78c.
But I do have 5 case fans - perhaps an extra fan or two for the case would help.

I seem to remember him having one rear and one roof extract fan but I'll check. I've told him to pop open the case anyway when gaming to see if that helps.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
Ermmmmmmmm

He needs to install something like hwmonitor and see if it is actually the temps first?

I told him to install that a few days ago and to monitor it.
I was sort of interested to see if it woul dbe beneficial to run software to load up the CPU and GPU to speed up the fault finding and what software would be best.
 
Associate
Joined
26 May 2012
Posts
1,581
Location
Surrey, UK
One thing to test is to reset any overclocks to base values. See if that makes a difference.

Otherwise could be a gpu issue if games specifically are causing crashes. Try reverting back to some older and more stable drivers, if on recent drivers. Or if you're lucky to have a spare around, swap out the gpu to see if it still crashes. I had similar issues a month ago and it turned out to be dying GPU.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
One thing to test is to reset any overclocks to base values. See if that makes a difference.

Otherwise could be a gpu issue if games specifically are causing crashes. Try reverting back to some older and more stable drivers, if on recent drivers. Or if you're lucky to have a spare around, swap out the gpu to see if it still crashes. I had similar issues a month ago and it turned out to be dying GPU.

I'm not so sure if he has any OC. I believe all run at normal out of the box levels.
I'll be home next week so hopefully all going well with relaxing of lockdown, I can pop over to his and have a look.
 
Soldato
Joined
15 Oct 2019
Posts
11,656
Location
Uk
Sounds like something is over heating which is causing the shutdown. the cpu/gpu can be monitored but it could also be the psu.

Has the weather being cooler over the last couple of days made a difference?.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
Sounds like something is over heating which is causing the shutdown. the cpu/gpu can be monitored but it could also be the psu.

Has the weather being cooler over the last couple of days made a difference?.

I've not seen him online to ask him. I did wonder if it could be the PSU too as he was given it FOC from a work mate when he was building. I believe it was an older 750W bronze PSU.
 
Associate
Joined
28 Nov 2019
Posts
126
The CPU should massively thermal throttle before causing a reboot, and even then a reboot will only occur if the CPU has declocked to it's minimum clock speed and the temperatures are still increasing; this kind of thing could happen if voltages are too high (in an overclock), or the fans on the CPU cooler stop spinning. Basically, given this is happening over a long gaming period, I would expect he would see a performance impact in game before he sees it reboot if it is the CPU. The same principle is true for the GPU too. The most likely cause of a reboot in a gaming session is an unstable overclock of either the CPU or GPU.

So here are some things to check in this order:
1) Are the fans on the CPU cooler and GPU cooler spinning when under load?
2) Are the fins on the coolers clear of dust and other blockages?
3) Has he done some overclocking? If yes, then reset the overclock of the component(s) in question back to defaults
4) Run a mem test on the memory to check for errors

Failing that, the PSU has already been mentioned as a possible cause, but as another possiblity (as an outlier when it comes to reboot issues, but it is still worth checking nonetheless)... Is he using an NVMe SSD and if so, does it have a heatsink? My cousin had an issue with an NVMe that overheated and caused a reboot during a disk heavy operation; to be fair though in his case, the NVMe completely dissapeared until it had cooled down so it was kind of obvious what the cause was. Still, it is possible that an NVMe could cause a reboot but also still show up. So that is something else worth considering if he has an NVMe SSD.
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
The CPU should massively thermal throttle before causing a reboot, and even then a reboot will only occur if the CPU has declocked to it's minimum clock speed and the temperatures are still increasing; this kind of thing could happen if voltages are too high (in an overclock), or the fans on the CPU cooler stop spinning. Basically, given this is happening over a long gaming period, I would expect he would see a performance impact in game before he sees it reboot if it is the CPU. The same principle is true for the GPU too. The most likely cause of a reboot in a gaming session is an unstable overclock of either the CPU or GPU.

So here are some things to check in this order:
1) Are the fans on the CPU cooler and GPU cooler spinning when under load?
2) Are the fins on the coolers clear of dust and other blockages?
3) Has he done some overclocking? If yes, then reset the overclock of the component(s) in question back to defaults
4) Run a mem test on the memory to check for errors

Failing that, the PSU has already been mentioned as a possible cause, but as another possiblity (as an outlier when it comes to reboot issues, but it is still worth checking nonetheless)... Is he using an NVMe SSD and if so, does it have a heatsink? My cousin had an issue with an NVMe that overheated and caused a reboot during a disk heavy operation; to be fair though in his case, the NVMe completely dissapeared until it had cooled down so it was kind of obvious what the cause was. Still, it is possible that an NVMe could cause a reboot but also still show up. So that is something else worth considering if he has an NVMe SSD.

Thats some good points acually. Forgot about the downclocking when things thermal throttle.
He isn't using an NVME, Just an SSD for Windows and an older HDD for a games drive.

I guess we will just have to monitor everything.

As questioned above, It does seem to have happened since the higher temps reached us. He never had an issue prior to this I believe.
 
Soldato
Joined
25 Oct 2010
Posts
5,231
It could be a PSU issue, what's the exact make and model?

Not all PSU issues are related to a lack of power being supplied, sometimes the protections can trigger and cause a shut down for a number of reasons, including a fault with the PSU itself. In these scenarios the fault may be periodic.

Monitor temps and what have you but if the problems persist I'd try it with a different PSU and see if it helps, maybe you could loan him yours for a day or two?
 
Soldato
OP
Joined
19 Apr 2012
Posts
5,183
It could be a PSU issue, what's the exact make and model?

Not all PSU issues are related to a lack of power being supplied, sometimes the protections can trigger and cause a shut down for a number of reasons, including a fault with the PSU itself. In these scenarios the fault may be periodic.

Monitor temps and what have you but if the problems persist I'd try it with a different PSU and see if it helps, maybe you could loan him yours for a day or two?

That is a possibility. He has been looking at getting a newer PSU anyway so maybe he might look into getting one sooner.
I may be able to loan him my one so will see what he thinks when I chuck him some suggestions from here.
 
Associate
Joined
28 Nov 2019
Posts
126
As questioned above, It does seem to have happened since the higher temps reached us. He never had an issue prior to this I believe.
So one thought that I have is that as temperature rises, electrical resistance increases. If your brothers CPU or GPU are overclocked and in colder weather there is only just enough voltage for it to be stable at the highest core temperature in that cooler weather, then increasing vcore very fractionally could fix it. The idea is this, as the warmer weather sets in, the ambient temperature increases, which in turn increases the max temperature reached by the CPU and GPU cores, this increases the electrical resistance further than it would have done in the colder months and if the vcore voltage isn't high enough then it could reset in hotter months but not colder months.

Before changing voltages though, try running some stress tests on the CPU (Cinebench R20 is free and might push it enough to see it lockup/restart if there is a thermal issue with the CPU, else something like Prime 95) and GPU(Heaven Unigine should help there). Also run a mem check on the RAM just to check it hasn't developed a fault.
 
Back
Top Bottom