System experiences instability when under heavy usage.
System experiences instability when under heavy usage.
My recent PC experiences frequent crashes during gaming sessions. After installation, everything functions normally, but games like Fallout 4, No Man's Sky, Overwatch, and Diablo 3 frequently terminate to desktop. No error alerts appear; the applications simply end abruptly. I also tried Cyberpunk, which caused total freezes and system crashes. Performance checks in 3Dmark yielded mixed results, with only a handful succeeding before an error message appeared. I updated GPU drivers via GeForce Experience and refreshed BIOS on my motherboard. Running stress tests through MSI Afterburner showed stable temperatures, though performance remains inconsistent. All hardware specs are current, and no overclocking is enabled. When I disable the CPU overclock, the issue disappears.
It comes from two separate cables. The GPU uses a 12-pin adapter that converts to three 8-pin connectors. My PSU only has slots for two GPUs, so one 8-pin needs its own and the other two share the remaining ones. This was the largest PSU I found that would fit in a mini ITX case.
I own another machine that's 3950x with a 2080 and 1000w PSU. Everything is liquid cooled, which meant I thought it might be a backup option. I didn’t want to drain the loop to get parts unless it was really needed.
Identifying the issue often begins with checking the CPU, as it's linked to a complete system failure.
I transferred the 2080 from another system and installed the stock air cooler on it, then connected it to the PC running the 5900x. I executed a speed stress test and it succeeded immediately. The 3090 ti hadn’t finished once during its more than ten attempts. After removing the 2080, I reinstalled the 3090 ti, then took out the 1000w PSU from the first system. The 1000w model features six VGA ports, allowing all three 8-pin cables for the 3090 to be connected properly.
I added a few more utilities besides afterburner. I’m unsure if you can view GPU VRAM stats, but I couldn’t locate them, so I used TechPowerUp GPU-Z. What’s the safe range for RAM temperatures and could they be causing the instability you’re experiencing? Vram reaches up to 112°C, with a hotspot around 98°C on the GPU during heavy tests, while the overall GPU stays at 89°C under full load. I’m also concerned about the setup that might be placing DRAM directly against the side of the case separating it from the motherboard—could this lead to random errors?
Reach extreme temperatures with 90°C. Begin by increasing fan speed to full capacity at 60°C in the drivers.