A real good one for you all - PC restart absolutely randomly.
A real good one for you all - PC restart absolutely randomly.
I’ve got a solid setup for someone I know. It’s been behaving oddly lately. Sometimes it restarts itself right after launching a 3D app—like the GPU kicks in and then shuts down again. Usually it works for 10 to 15 minutes, but sometimes it fails completely during heavy stress tests. Two days ago, after running Furmark 2 and OCCT under maximum load for 12 hours, it stayed stable. Then I played Cyberpunk 2077 at night with the character in Dog Town for hours without any problems. After a 24-hour stress test, when I tried to launch any game, it restarted instantly. It’s basically the same rig as mine—5800X3D, X570 Steel Legend, 1000W Focus PSU, with 3600MHz g.skill RAM and an RX 7900XT. All standard checks passed: CPU/GPU heat, motherboard temps, SSDs, memory tests. I updated drivers and firmware multiple times, swapped parts one by one, but mixing them still causes issues. I even tried swapping components between systems, but it only works when everything’s mixed together. I’ve been trying to figure this out for days without success. I’ve changed PSUs, GPUs, CPUs, boards, RAM, coolers—everything I can think of. Hopefully someone spots something I missed; my mind is really tired from all this work.
Was die RAM-Geschwindigkeit bei 3600 auch so? Vielleicht muss ich nachfragen. Die BIOS wurde aktualisiert? Und ist es eine Beta-Version? Dann passiert das nicht sonst. Klingt eher GPU-bezogen und nichts hat funktioniert. Leider handelt es sich nicht um ein PRIME-Modell, sondern um einen FOCUS.
Slower RAM doesn’t solve the problem either; reducing power targets and manually adjusting GPU voltage and clock speed didn’t help. BIOS settings for the motherboard, VBIOS on the GPUs, and firmware for the storage devices are all up to date. This was my first step when he brought the system to me. Even though the unit isn’t a Prime series but a Focus GX, my instinct suggested a PSU might be the issue. It’s the only component I can’t reliably stress test without custom equipment and an oscilloscope. The main concern is that once I put it in the system it functions perfectly, despite occasional instability when the GPU is downclocked—sometimes it stays under 280W, while other times it spikes above 340W. As noted, the system often restarts automatically after pressing Run during Furmark tests or loading screens for any game.
Have you explored methods that significantly boost frame rates? It's not only visual load that strains a GPU, but also the effort to maintain high frame counts.
Tested several games using AFMF. It functions regardless of settings or game. When it fails, it completely stops working. The immediate restart after launching any 3D app is quite notable. This isn't a CPU issue since stressing the CPU doesn't cause problems. Memory seems unaffected. My suspicion leans toward the GPU or PSU, with the PSU being more likely based on my instincts. Anyone know how to specifically test the PSU? As for the GPU, even after overclocking it exceeds 340W but stays under 77°C with a hotspot of 85°C—well within the Nitro+ Vapor-X RX 7900XT's specs, which lists 368W total board power on Sapphire's site.