Tasks requiring constant CPU usage cause significant instability in this system.
Tasks requiring constant CPU usage cause significant instability in this system.
Hello everyone, I just shared details about my 9800X3D running on an Asus Strix B650E-I mini-ITX board. I've tested both at standard voltages and with negative offsets. The main issue appears when temperatures rise, especially near a Liquid Freezer III 360 AIO. It only happens in hot conditions—above 25°C in the UK it’s likely to occur. Between 20-25°C seems fine. When loading certain games or doing tasks like precaching or shader compilation, sudden usage spikes cause temperatures to soar and the PC freezes completely. I can’t shut it down until I press the power button. Stress testing didn’t reveal problems, but when I tried Rainbow Six Siege, the CPU would spike to 100% during system checks, temps jumped, and the game froze. Monitoring tools like HWinfo showed no data after that point. I’m adjusting PBO settings to tweak voltage or fan curves, but nothing seems to resolve the problem.
It seems you've already tackled many troubleshooting aspects. To refine the approach, consider comparing socket and CCD temperatures—HWInfo may only record CCD temps, but if the VRM or socket area gets hot in a compact case, instability can appear before core temperatures reach limits. Verify cooling contact with the pump/AIO mounting; issues like poor paste spread or incomplete contact can trigger sudden spikes and freeze events. It might help to reapply fresh paste to the block. Ensure your BIOS is up-to-date for the Strix B650E-I, as X3D chips are known for stability challenges early on. Try capping power limits (e.g., ~85W PPT) instead of offsets and observe if spikes diminish. After a hang, check Event Viewer for WHEA errors—especially related to CPU cache or memory—to confirm if the problem lies with the CPU or software. Running tests in an open case can reveal if airflow is the limiting factor beyond the AIO. This appears more like a sensitivity issue with cooling or firmware/voltage handling under load rather than a faulty component. If previous steps fail, consider applying manual limits to prevent runaway boosting. Note: this advice comes from chat gpt, but following these steps should help.
I'm experiencing a few problems with my CPU as well (another update), but it's not related to temperatures. The system still shows normal temperature codes even though I'm not overclocking. I'm also seeing all the CPU timing codes when everything seems okay. I'll check your post if I find anything useful.
I replaced the CPU and tested it with the standard configuration. No overheating issues were observed. After removing the CPU module, it was evident that the thermal paste wasn't evenly applied in the middle. While some spread to the edges, the center remained underperforming. This inconsistency points to a potential problem. I started Rainbow Six Siege and successfully passed the Detecting System Configuration stage using default settings, reaching a peak temperature of 89°C—comfortable and within normal operating ranges. It didn’t function properly at -23°C, so I adjusted the offset to -20°C for further testing. It’s great to know it’s working fine at default! Thank you!
I tried the same approach, noticed there are brief intense spikes that aren't explained. I used a bit more paste than usual but stayed moderate, and it still works similarly. If you're like me, I'll keep monitoring since it doesn't seem to be the fix.