[SOLVED] System does full shutdown under load

[Fabian van Doorslaer]

SPECS
CPU: AMD Ryzen 7 3700X (with Noctua NH-D15)
GPU: AMD RX Vega 56
RAM: Corsair Vengeance LPX DDR4 3200 C16 2x16GB
MBD: Asus ROG CROSSHAIR VIII HERO
PSU: be quiet! STRAIGHT POWER11 CM 750W
SSD: Samsung 970 Evo Plus NVMe PCIe M.2 500GB
SSD: Samsung 860 Evo 1TB
SSD: Samsung 850 Evo 500GB
HDD: WD 4TB

PROBLEM
My system does a full shutdown under load (gaming or benchmarking), all of a sudden everything goes black and that’s it - no sound, no fans, no restarting.
I can perfectly power it back on afterwards, and everything works fine, - although I’ve noticed ‘0C’ on my ASUS Q-Code. I’m certain it hasn’t always been like that, but I don’t know when that changed. (Don’t need to go into that, it says ‘reserved for future AMI SEC error codes’ in the appendix, so ASUS won’t help us.)
I’ve already did a lot of research, and found that in the windows event viewer, a critical event always gets registered after a shutdown. It’s ID is ‘41’ and it involves kernel power, so I went on the microsoft advanced troubleshooting forums and basically what they say is: or the event has a code in it that explains what the problem is - or not. Mine of course doesn’t provide a code, so back to guessing.

I’ve run some testing and benchmarking already, and my system is perfectly fine for the rest. I don’t do real overclocking, but I did however try some light boost driving in the AMD Ryzen Master utility, and in the ASUS UEFI. I checked if the problem still persisted or not, and it does. Both with the light boosts enabled or disabled. Right now I’m on ‘standard everything’ and it’s still a problem. I’ve updated all drivers, chipsets, and what have you not.

QUESTION
The next step is probably a ‘better’ PSU, but I’m a bit hesitant cause I already have a potent and expensive one, and I’ve calculated my needs in multiple PSU calculators before buying mine, it should be more than fine. Does anyone see something that I don’t in this situation? I really don’t want to buy a new PSU if I’m not 100% certain that that’s the problem.

 

[jay32267]

Have you checked the temps or the voltages while it's under load?
If not....I would do both.

 

[Mr.Spock]

your PSU has a multi-rail design IIRC, it might be beefy but maybe you're pulling more than an individual rail can handle?

 

[Fabian van Doorslaer]

Have you checked the temps or the voltages while it's under load?
If not....I would do both.

Thankyou very much for answering. I will definitely check that, would you know a one-stop utility for both? I can use ryzen master for the temps, not sure about the voltages though?

 

[Fabian van Doorslaer]

your PSU has a multi-rail design IIRC, it might be beefy but maybe you're pulling more than an individual rail can handle?

Thanks for the suggestion. Uh, I didn't know about single/multiple rail designs, so I went searching. The gist of it is that with modern PSU's (like mine) it really doesn't matter that much anymore, would you agree?

 

[jay32267]

Thankyou very much for answering. I will definitely check that, would you know a one-stop utility for both? I can use ryzen master for the temps, not sure about the voltages though?

You're welcome.
HWInfo will work for the voltages.
You want the main voltages....+12, +5 and + 3.3 to remain within 5% of these values under heavy and light loads.
Temps.....I like the CPU under 70C and GPU under 80C.

 

[Fabian van Doorslaer]

You're welcome.
HWInfo will work for the voltages.
You want the main voltages....+12, +5 and + 3.3 to remain within 5% of these values under heavy and light loads.
Temps.....I like the CPU under 70C and GPU under 80C.

Ok I think I got you. So I ran Cinebench R20 twice, the second one following the first one immediately - all while closely watching HWinfo. I don't think I've ever seen anything strange popup, but just to be sure here's a screenshot from right after the tests.

HWinfo Screenshot

I don't know if this matters, my score was 4501.
(Also, the PC obviously didn't shut down, so for now I'm not able to recreate the problem. Should I do a heavier benchmark?)

 

[jay32267]

I'm not seeing anything strange.
Your voltages look good.
Your CPU temps look good.
I can't see GPU temps.....as I think you need to scroll down further in HWInfo.

 

[Fabian van Doorslaer]

I'm not seeing anything strange.
Your voltages look good.
Your CPU temps look good.
I can't see GPU temps.....as I think you need to scroll down further in HWInfo.

Thanks so much for baring with me.
Ok so I was running PCMark 10 right after this and during the video editing phase of the benchmark, it happened again. Lights out, nothing - full shutdown all of a sudden. I had been monitoring HWinfo in the 10 minutes beforehand and nothing got even close to worrying. I'm gonna try to recreate this again and film the GPU part of HWInfo

 

[jay32267]

Thanks so much for baring with me.
Ok so I was running PCMark 10 right after this and during the video editing phase of the benchmark, it happened again. Lights out, nothing - full shutdown all of a sudden. I had been monitoring HWinfo in the 10 minutes beforehand and nothing got even close to worrying. I'm gonna try to recreate this again and film the GPU part of HWInfo

You can do that.
I think you can ALSO set up hardware info to log the data to a file.
Then...I think it "should" save all the data right up to the crash.
Although I have never tried this because I have never had to try this.

 

[Fabian van Doorslaer]

You can do that.
I think you can ALSO set up hardware info to log the data to a file.
Then...I think it "should" save all the data right up to the crash.
Although I have never tried this because I have never had to try this.

Yeah I saw that option as well. But not sure how and if it works, just like you.

BUT I managed to film a crash with the camera pointed at HWinfo stats for CPU and GPU temps. It's really strange, the system crashes at the same place in the PCMark10 bench everytime (me filming this is the 3d time) - it's during the video editing part. I'm actually quite sure that nothing worrying is going on, but if you could have a look I would know for sure.

I filmed for more than 10min (the video editing phase is only halfway in the benchmark) but I've cut the video to the last 30sec before the crash since nothing's happening. At least you can see it with your own eyes now...

YouTube link: PCMark 10 benchmark - shutdown

 

[jay32267]

I don't see anything.
At least we know your voltages and temps seem good.
Do you have another GPU to try.
I'm suspecting it's possibly the GPU.

 

[Fabian van Doorslaer]

I don't see anything.
At least we know your voltages and temps seem good.
Do you have another GPU to try.
I'm suspecting it's possibly the GPU.

Thanks for watching the video.

No I don't own another GPU. I can try to get one of course, though that will take some days unfortunately.. What makes you think that it's the GPU? Is there something I could still try in the meantime?

 

[jay32267]

I think it might be the GPU because I think we've somewhat ruled out the temps and the power supply.....and if you DID have another GPU it would be a real quick way to either determine it IS the GPU or eliminate it as the problem.

I'm just trying to do the easiest things to eliminate "or include" as many things as possible.

 

[Fabian van Doorslaer]

I think it might be the GPU because I think we've somewhat ruled out the temps and the power supply.....and if you DID have another GPU it would be a real quick way to either determine it IS the GPU or eliminate it as the problem.

I'm just trying to do the easiest things to eliminate "or include" as many things as possible.

Ok that makes perfect sense, I'll try to find one haha. Thank you so much already!

 

[jay32267]

Ok that makes perfect sense, I'll try to find one haha. Thank you so much already!

You're welcome.