PC Restarting Under Load

[NameIsGreg]

Also try TM-5 with anta absolute config or equivalent

Thanks for the link @3x0, I ran it for an hour and half and had no errors again. Should i run it for longer or do you think this is sufficient?

 

[3x0]

TM5 combined with HCI MemTest is sufficient to rule out memory instabilities.

I'd try Corecycler with an overnight test to see if any reboots happen under lightly threaded workloads https://github.com/sp00n/corecycler/releases
Default settings are OK, just close all background apps.

 

[NameIsGreg]

TM5 combined with HCI MemTest is sufficient to rule out memory instabilities.

I'd try Corecycler with an overnight test to see if any reboots happen under lightly threaded workloads https://github.com/sp00n/corecycler/releases
Default settings are OK, just close all background apps.

@3x0 I forgot to put it on last night but I let it do a full iteration this morning and no instability and 0 errors, I can test it overnight tonight if you think I should

 

[3x0]

I can test it overnight tonight if you think I should

It's best to do it overnight since there are a lot of cores to test, running one 10 minute test for a lot of cores is not enough

 

[NameIsGreg]

Basically test the CPU by running a Prime 95 or similar -- if that passes great.

Then run a GPU stress by running furmark -- if that passes, also great.

Then run them both together, at the same time, even start and stop the furmark test a few times with the CPU test running - this will pull max wattage from your PSU, and it will power cycle if the Power Supply cant handle transients.

The reason for this is your error mode (no log, no dump) is as if someone just ripped the power cord from the computer - that's really the only explanation for your hardware testing fine individually but the rig rebooting with no minidump or whea error or any other faults logged. It really might be as simple as the PSU overvoltage protection tripping.

@phanbuey So I ran prime95 for 1hr and it passed, after that ran furmark for 1hr and it passed. Then ran them simultaneously and they both passed - i restarted furmark around 4 times in the hour and was toggling the "render furry object" throughout the test to simulate transient loads and my system got pretty hot but it never crashed.

@3x0 okay will do that tonight.

 

[phanbuey]

I hate to say it but this is sounding like a bad mobo

New psu works - passed stress
Gpu is fine
cpu is fine
Ram is fine

The only remaining components are ssd and motherboard… unlikely that it’s the ssd.

 

[Koozwad]

@Koozwad
iirc the rtx 20xx was worse, and i dont have trouble with the 750w, even if i put max load on gpu (binned, 2-2.2ghz constant boost).
the pcb has diagnostic leds, that would show if there is a problem power wise (that i did have with 2 different 850 units from other brands; one power led getting triggered during bench/gaming).

I don't ever recall hearing about the 20xx series having the issue, but then again I kinda skipped that one completely. Went from a 980 to a 3090 and I was out of the loop for some time. The 20xx series was still using NVIDIA's 'max' TDP of 250W, while the 30xx series went quite far over that for the first time(ever?).

I would advise anyone with this issue and a 30xx card to first and foremost try some verified stable PSUs that can handle the spikes. From memory even some 1000W Titanium efficiency high-end Seasonics couldn't handle the spikes, to name one example.

 

[Waldorf]

@Koozwad
regarding transient spikes it was worse (considering tdp), but mainly on the xx80/ti, especially with those coming with mfg oc or binned chips for LC cards (like mine).
the 850w gold/plat units i tried from Tt and (i think it was) a Kingwin (or similar) that got decent reviews, both had issues with my card at "stock" (raised power/thermal limits),
while no issues with my friends 2080Ti and 2080S (mid range cards with 25mhz factory "oc").
then again, i had seen my power diagnostics led go "off" without OS/game crashing, so it might have not even be noticeable if there was an issue with it.

part of the reason in dont recommend any earlier releases of RMx series (only the latest).

 

[NameIsGreg]

I hate to say it but this is sounding like a bad mobo

New psu works - passed stress
Gpu is fine
cpu is fine
Ram is fine

The only remaining components are ssd and motherboard… unlikely that it’s the ssd.

Well that kinda sucks, I don't really understand how I wouldn't of seen a crash when stress testing though?

I had another crash whilst playing factorio, and it just came out of no where, my pc fans didn't turn up so my parts won't overheating. I just dont understand it.

Anyway I can test the motherboard to try and actually pinpoint the issue before I get a new one?

I've also been looking around for people who have this same restarting issue and I have seen that people changing their fans (mainly the controller) that can fix it and some people were even saying a wiring issue in the case could cause it? Do you think its likely that something could be shorting in my pc somewhere, thats what I think it could be. What do you guys think?

 

[phanbuey]

Well that kinda sucks, I don't really understand how I wouldn't of seen a crash when stress testing though?

I had another crash whilst playing factorio, and it just came out of no where, my pc fans didn't turn up so my parts won't overheating. I just dont understand it.

Anyway I can test the motherboard to try and actually pinpoint the issue before I get a new one?

I've also been looking around for people who have this same restarting issue and I have seen that people changing their fans (mainly the controller) that can fix it and some people were even saying a wiring issue in the case could cause it? Do you think its likely that something could be shorting in my pc somewhere, thats what I think it could be. What do you guys think?

Yeah that makes sense - given your testing it's either coming from the mobo, or something touching the mobo causing a short - can be fans or RGB, can be a bad component on the mobo itself etc, can just be a raised part of the case touching a cap etc.

Very hard to tell since it's intermittent and not reproducible, but you've eliminated essentially all other components.

Maybe check SSD for SMART errors, but otherwise there's only one major component left. Maybe also try a different outlet, or an APC.

 

[Shrek]

No I haven't, I'll give that a go. I've been using Memtest64 and Prime95.

Also did a memtest86 test a while back but it reported 0 errors.

Tried Windows Memory Diagnostic?

 

[jsfitz54]

@NameIsGreg : Something to consider

Sometimes on the back of the motherboard, where a component is soldered in through the board, the leg that protrudes through the board was not trimmed short enough at the factory and it needs to be trimmed down with flush cut trimmers.(could be cause for shorting, as the case is expanding and contracting with heat)

example: https://www.walmart.com/ip/Side-Cut...elry-Processing/15546552395?classType=REGULAR

If this is impossible try wedging a stiff cardboard shim, such as that found in a credit card offer. (or an old plastic credit card or some plastic of the like that can be cut like a thin shim)
The tension of screwing the board to the standoff of the motherboard could puncture the cardboard shim if too tight so be careful there.

You may want to check that there are no extra standoffs causing a short.