PC restarts after around 20 minutes of playing (probable cause: GPU)

[Tumpei]

 

I recently upgraded most of my setup expect for the GPU because .. well everyone knows why. I had this problem with my old setup where a game like PUBG ran well for around 20 minutes and seemingly out of nowhere the pc would restart itself. I thought it must be either a cooling or power issue since my cable management and airflow was. Now I've upgraded everything else except the GPU and RAM from my old setup and the problem persists.

 

My GPU temps never go past 65-68C and fans never seem to go higher than ~40%. The restarts have now started to happen in less demanding games like League of Legends and Divinity 2 in addition to PUBG. I've tried running FurMark and 3DMark stress tests but can't seem to make the PC restart that way. I have done clean installations of Windows and graphics drivers and tried relatively low settings in all games.

 

So I guess my question is, is that how can I be sure its my GPU causing the issue and is there anything I can do to fix it?

 

 

Specs: 

CPU: AMD Ryzen 7 5800X

CPU COOLER: NH-D15

MOTHERBOARD: ROG Strix X570-F Gaming

MEMORY: 4x4GB VENGEANCE LPX, DDR4 3000MHz

GPU: NVIDIA GeForfce GTX 1060 6GB

STORAGE: Samsung SSD 1TB 970 EVO Plus

STORAGE: Samsung SSD 250GB 840 EVO

STORAGE: Seagate Barracuda 1 TB

PSU: Corsair HX850, 80 PLUS Platinum

CASE: LANCOOL II Mesh RBG

 

Two 144HZ monitors

[lafrente]

Try using ram sticks one by one first

[Alexsheff93]

i honestly think its the ram take all your ram sticks out and test each slot one of them will be faulty and if its not a ram slot it'll be a dodgy ram stick. considering your GPU temps there completely fine. Ram is usually the one that culprit that never goes noticed till all the other kinks have been worked out.

[Tumpei]

Update: Spent 6 hours today testing my RAM. Ran MemTest86 on every stick separately and every slot and it gave me no errors. I still think it is GPU related but don't know how to continue. 

 

I don't know if it's any indication but when running stress tests GPU Shark this is what I noticed, are these normal numbers?