PC freezes during gameplay but restarts smoothly without encountering a crash screen.
PC freezes during gameplay but restarts smoothly without encountering a crash screen.
So two weeks prior, my entire PC experienced its first major crash while playing games. Now, over time, similar incidents have occurred multiple times. One instance involved the game freezing completely, though the system remained operational until everything halted after a few seconds before restarting. Another case featured a complete crash followed by freezing and subsequent failure. I've picked up some advice from others' experiences, but this doesn't seem to point to a failing power supply unit. Still, it's hard to say for certain.
When assembling the PC, I immediately lowered my CPU voltage (using Curve Optimizer, all cores, negative offset -24) and managed to run several AAA titles without problems for about a month and a half, covering games from 2008 to 2024. However, I encountered some issues even on a desktop setup, which I resolved by removing certain ASUS software and slightly adjusting the offset. After the first crash two weeks ago, I increased the CPU voltage to -20, but the crashes continue, leaving me uncertain if this is the root cause—especially since the part claims it should handle that level.
I also undervolted my GPU, but accidentally left it off for a short period, which led to crashes. There were no BSODs or significant minidumps, but I enabled a full dump after the fifth crash. The Event Viewer offered little help; only Kernel-power 41 (63) was recorded when the system lost power. While CPU and PSU failures typically trigger specific alerts, everything else seemed normal.
Temperatures didn't seem to be the problem—CPU stayed between 50-60°C, occasionally reaching 70 or 80°C during heavy loads. I reached 95°C during stress tests, which was unusual since fans were working hard. The GPU consistently hovered around 40-55°C, matching fan behavior. I've configured fan curves via Fan Control and installed games on a Samsung SATA SSD with 99% health reported by SMART. The drive itself is new (2 years old) and shows no signs of distress.
I never enabled EXPO on my 32GB RAM kit; at 4800 MHz it performed well, though I've faced similar issues before. I considered increasing the clock speed to 5200 MHz but stopped before the problem reappeared. RAM temperatures were around 31°C during games.
Actions taken:
- Ran sfc /systemscan and DISM commands, which resolved some errors.
- Reduced curve optimizer undervolt offset from -24 to -20 after the latest crash.
- Conducted OCCT, Aida64, and Karhu mem tests without errors.
- Updated GPU drivers.
- Verified specs: AM5 ASUS TUF GAMING B650E-E WIFI, RTX 5070 Ti, Ryzen 7 7700x, Corsair RM750x, etc.
Except the GPU, everything else is about two months old. The GPU I received in early December was undervolted during assembly and had no issues. CPU was used for a short time before being replaced with a used unit, which has worked without problems so far. I'm not sure if it's the cause, but replacing it would be costly.
Otherwise, I plan to disable Curve Optimizer and report back once more.
I'll provide further details if needed. I really hope this gets resolved.
Edit: It seems the aggressive undervolting of the CPU was the main issue—running it at -15 after the crash confirmed it. I'm 99% confident it was the cause.
It's not, yet all items on the official QVL list for my board are also no longer available. At the time I bought the current kit, it was quite costly considering the specifications (most were CL36 and 38), and they seem even more expensive now given current RAM prices. Also, I recall seeing a few kits with similar specs from another vendor while searching for memory kits, so it's likely my current TeamGroup kit will still function properly. It was probably made a few years ago, and the QVL list has been updated over the past few years.