Question GPU Overheating and then shutting down my new computer

[kasual01]

I just build a new computer (all new parts), but I am having issues which I believe are due to the GPU overheating.

My specs are:

• Asus Rog Strix X570-E Gaming
• RX 6600xt
• AMD Ryzen 5 5600x
• 650w Corsair
• 16 Gb RAM

So, I used the new computer without problem for a day, and then when I tried to play a "demanding" game, the computer shutted down. First the screen freezed for a few seconds, then the screen went black but everything was still running (I couldnt hear sound, or type anything). I couldn't turn on the computer for a few minutes after but if I waited a few minutes the computer would turn on normally.

I tried using MSI Afterburner to monitor the GPU (I WAS NOT OVERCLOCKING) and I found that it increased the temperature constantly, until it shutdown thus why I believe the GPU is the problem (or the fans in the GPU).

MSI Afterburner:

• GPU Temp (max): 76°C
• GPU Temp 2 (max): 93°C
• FB Usage (max): 99% (during gameplay, it satyed constantly at this level)
• Power (max): 136 (don't know where this is from but it stayed constant at the max level)
• FAN Speed (max): 51% (it stayed constant at this level even though the temperature was increasing)
• CPU Usage oscilated a lot, max was 47%, but most of the time was very below that
• RAM Usage (max): 5595 MB (I have 16 Gb, so not much)

(I don't know how to insert an image, so I put the link above)

I have the GPU under warranty for now, so I wanted to try and fix it (or at least find what's wrong) without opening it.

I can increase the fan speed manually with MSI (and I can see that it's indeed going much faster when I increase it manually), but I don't know why it doesn't do it automatically since it knows it is too hot.

Does anybody knows why my fans aren't adjusting automatically to the increase in temperature in my GPU and how I could fix it?

 

[kurdtnz]

Just to ask the basics, are all your drivers up to date, graphics, chipset etc? Also, what BIOS version is your motherboard on? And finally, what case are you using and what is the airflow like. how many intake/exhaust fans etc?

 

[kasual01]

All drivers are up to date (GPU, chipset, bluetooth, audio, etc). I updated by the lastest version in manufacturer's site.

BIOS is up to date too. Mine is version 4021, released in september 2021.

Case is Montech X3. The airflow fells good. 3 intakes (all front) and 3 exhaust (2 top and 1 back) fans.

If you need anymore information, just ask.

 

[SkyNetRising]

Case is Montech X3. The airflow fells good. 3 intakes (all front) and 3 exhaust (2 top and 1 back) fans.

Which Montech X3 ? Mesh or glass?

 

[kasual01]

Which Montech X3 ? Mesh or glass?

Glass

 

[SkyNetRising]

Glass

That would mean - it basically has no air intake.
Try removing front panel and see, how that impacts temperatures.

 

[kasual01]

The panel can't be removed (only if I break it). And it has inflow, due to openings to the side. Putting my hand behind the front fans, can few the airflow, at least.

And the main problem isn't that. but that the GPU fans stops at 51% speed and doesn't go further unless I manually increase it.

Edit: found how to remove it (front glass panel) and it improves airflow, but it still doesn't help with the GPU. It is still overheating, but slower now.

 

[Unolocogringo]

You case is TERRIBLE for air flow.
Most of what you feel is recirculated case air. NOT fresh intake.
Those little slots do not let in enough air for one fan much less 2.
Try running it with the side panel removed, if temps drop more than a few degrees, your case is the problem.

 

[kasual01]

I did feel am improvement by removing the front panel. Temperature dropped a lot. But leaving the side panel open didn't do much to help. I will see if I can find a MESH version to switch to improve airflow permanetly.

But the problem that I need help isn't really that. I just need to know why the GPU stops at 51% speed, even though the temperature is rising above 80°C (176F).

 

[Phaaze88]

Gpu Temp 2 is probably the hotspot, or the hottest sensor on the die. It might not be the reason for the shutdowns; hotspot limit is usually a fair bit higher, like 110-120C.
The fans aren't reacting to that, but the average - the Gpu Temp 1 most of us are familiar with.
93C should be within spec if it's a hotspot sensor.

 

[kasual01]

Gpu Temp 2 is probably the hotspot, or the hottest sensor on the die. It might not be the reason for the shutdowns; hotspot limit is usually a fair bit higher, like 110-120C.
The fans aren't reacting to that, but the average - the Gpu Temp 1 most of us are familiar with.
93C should be within spec if it's a hotspot sensor.

But GPU Temp was also at 76°C and increasing. I fell like at this point the fans should be at a higher rotation than 51%.

 

[Phaaze88]

Is this repeatable with Radeon Software running instead of Afterburner?

 

[DRagor]

First thing is, your GPU charts don't really show anything unusual. The temps weren't that high to cause trouble. Plus even if they were, GPU would throttle first rather then shut down. Nothing in the charts suggests throttling happened. Fans doing only 50% speed is how they are set up - you can and maybe should set them up to more aggressive curve and see if that changes anything. What troubles me more is this

I couldn't turn on the computer for a few minutes after but if I waited a few minutes the computer would turn on normally.

This actually suggest the problem might lie in PSU. So what exactly model of Corsair you have?

 

[kasual01]

First thing is, your GPU charts don't really show anything unusual. The temps weren't that high to cause trouble. Plus even if they were, GPU would throttle first rather then shut down. Nothing in the charts suggests throttling happened.

This chart It didn't turn off the computer. I closed the application before it would shutdown the PC. But, yeah, when I played until the PC shutdown, I didn't fell that it throttled at all before it shutdown. It was all working perfectly and then it freezed and then it shutdown.

This actually suggest the problem might lie in PSU. So what exactly model of Corsair you have?

Corsair CV650 650W, 80 Plus Bronze, CP-9020236-BR

 

[DRagor]

Corsair CV650 650W, 80 Plus Bronze, CP-9020236-BR

So you've got pretty much worst model Corsair currently sells in terms of reliability. Not calling it crap but it might actually not be enough for your setup as both your CPU and GPU do use quite a bit of power. Of course it's impossible to say for sure whether the problem is in GPU or PSU but if it were me I would start with swapping PSU. Well actually first thing would be to change GPU's fans curve and see if that solves a problem, then change PSU if it doesn't.