Critical processor malfunction detected.
Critical processor malfunction detected.
New 5900x is running well, temperatures are stable. I adjusted the curve in PBO to a negative offset of 30 (max). For a few days of gaming and idle time it worked. Then I restarted my PC after booting up, and the browser crashed with a fatal hardware error—CPU core issue. I reduced the offset to 25. This might be related to my undervolting. Fresh OS with essentials installed, no overclocking, just set the PBO curve to 30 for a short while. No long benchmarks were done.
I also have a 5900X and I've been trying PBO undervolting. I started with -30 and everything worked until I ran a Windows Defender scan on one of my NVMe drives. I prefer this as a hardware test because it simulates real usage, pushing the CPU to 90-100% across all cores without heavy AVX workloads, which lets boost clocks rise higher than Cinebench. Using HWInfo64 showed good performance with lower temps and higher clocks. Then suddenly the CCD1 temperature spiked above 86°C and the system shut down instantly. That’s why undervolting can be risky. I’ve now set it to 28 and all tests have passed so far.
I've noticed many users experiencing problems with their 5950x and 5900x displays. Most were using the curve optimizer, but -30 didn't work well even though it functioned okay in some games and benchmarks. On Reddit, people with 5900x reported crashes after fresh installs with default settings; they had to add a +4 offset in the curve optimizer for stability. It seems -25 might be a better setting.
That's unusual. My computer didn't crash either; it just restarted and I checked the event viewer. -25 appears to be functioning now, which is good, but I'm still hoping it resolves. I'd rather stay above -25 since I don't notice much improvement at lower temperatures.