Hello,
I recently bought a high-end computer (7 months ago), but the problem is that on some demanding games (e.g. Cyberpunk, BF2042, RD2, Star Citizen, etc...) I couldn't play for more than an hour without a crash back to Windows (or in the worst case a complete PC crash). I didn't pay much attention to it until a few days ago when I decided to do more research.
I took the opportunity to update the Bios (MSI Click Bios 5), and update most of my drivers, but nothing changes, there are still these crashes.
At this point, I think it comes from the PSU which must be defective but I'm not sure.
So, I tried a lot of Benchmark/Stress tests including several with OCCT with its "Power" test that puts CPU and GPU load to the max, and as soon as the first test starts, the software crashes.
I did 2-3 other tests and quickly noticed a problem, as soon as I start the Stress Test the CPU goes up to 100°C (Of course, I stop the test after a few seconds). And the second I stop the test, the temperature drops back to ~45°C (really in 1 second) (Well, I didn't study physics in college but is it normal for a component to go from 40->100->45°C in matter of 5 seconds?)
(The moment in question after the test)
(On the left, we see the vMax at 100°C, and on the right the chart showing the evolution of the CPU temperature on a 3-4s test)
I also decided to relaunch Cyberpunk to see if I still have a crash after an hour (just to check, you never know
). Except that the game crashes (back to Windows) after 2 minutes now...
I also launched League of Legends (which worked without problems from the start), and I got a BSOD.
In short, here's my setup (what seems important):
I recently bought a high-end computer (7 months ago), but the problem is that on some demanding games (e.g. Cyberpunk, BF2042, RD2, Star Citizen, etc...) I couldn't play for more than an hour without a crash back to Windows (or in the worst case a complete PC crash). I didn't pay much attention to it until a few days ago when I decided to do more research.
I took the opportunity to update the Bios (MSI Click Bios 5), and update most of my drivers, but nothing changes, there are still these crashes.
At this point, I think it comes from the PSU which must be defective but I'm not sure.
So, I tried a lot of Benchmark/Stress tests including several with OCCT with its "Power" test that puts CPU and GPU load to the max, and as soon as the first test starts, the software crashes.
I did 2-3 other tests and quickly noticed a problem, as soon as I start the Stress Test the CPU goes up to 100°C (Of course, I stop the test after a few seconds). And the second I stop the test, the temperature drops back to ~45°C (really in 1 second) (Well, I didn't study physics in college but is it normal for a component to go from 40->100->45°C in matter of 5 seconds?)
(The moment in question after the test)
(On the left, we see the vMax at 100°C, and on the right the chart showing the evolution of the CPU temperature on a 3-4s test)
I also decided to relaunch Cyberpunk to see if I still have a crash after an hour (just to check, you never know
). Except that the game crashes (back to Windows) after 2 minutes now...I also launched League of Legends (which worked without problems from the start), and I got a BSOD.
In short, here's my setup (what seems important):
- Motherboard: MSI PRO B660-A DDR4 - B660/LGA1700/DDR4/ATX
- CPU: Intel Core i5-13600K - 5.1Ghz/24Mo/LGA1700/BOX
- GPU: Asus TUF RTX 4080 O16G GAMING
- RAM: Kingston KF432C16BBAK2/16 8GB x 4 (at 3200MHz in PC25600) (XMP enabled)
- Watercooling: MSI MAG CoreLiquid C240 (Only for the CPU)
- PSU: Seasonic ATX 850W 80+ Gold - G12 GM-850
- With 6 fans, almost no dust and in a quite ventilated place
For me,
- either it comes from the PSU which is defective
- or it's the cooling that is defective or poorly arranged (I ordered Noctua NT-H2 thermal paste because the PC was not assembled by me but by a company and I think they must have used the thermal paste provided by one of the components)
- or it's a CPU temperature sensor that displays incorrect results and pushes the CPU to drop its frequency to the minimum until it turns off (but I doubt it)
- or it's the CPU?
I hope you can guide me on what's wrong.
Thank you in advance for your advice.
(I can attach screenshots of HWiNFO64, OCCT, BlueScreenView, Windows Event Viewer if you want more details)
Facebook
Twitter
Instagram