C8DH plus 5900X causes the CPU to freeze at 0.5 GHz.
C8DH plus 5900X causes the CPU to freeze at 0.5 GHz.
Hey there. Here’s the rundown:
My setup includes an Asus Crosshair VIII Dark Hero, a Ryzen 9 5900X, a Noctua NH-U14S, two 16GB drives, and some other components. The OS seems to be either the newest Windows 11 build on the Corsair or Pop!_OS 21.10 on the Samsung side. The issue started suddenly yesterday with no recent changes.
A few years ago I bought a GPU a month back and had an accident soon after installation—it popped out of its slot because it was too heavy, likely due to a wobbly bracket. After fixing it, the PC turned off and wouldn’t power on. Once reassembled, everything worked normally.
For the CPU, I usually see idle speeds around 3.6-3.8 GHz, which now climbs to about 4.25-4.4 for multi-core and 4.9-4.95 for single-thread tasks. Cooling looks good and temps are normal. I haven’t adjusted throttling settings. PBO is set to Auto or Disabled (it’s Auto by default after a reset), and CO is off. The diagnostic LED shows temps during operation, which would have flagged an error if anything was wrong.
Yesterday I left the PC running for some basic updates and noticed it stuck at 0.51 GHz in Task Manager—definitely slow. Other apps confirmed the same. I tried several fixes: changing power plans, using Ryzen Master, enabling Profile 1/2 settings, disabling PBO, resetting chipset drivers, reinstalling UEFI, and even trying to disable fast boot. None helped much.
Diagnostic tools like OpenHardwareMonitor reported odd power readings—RM said 161% of 142W PPT, CPU Power at 210W, and the package listed over 200W despite each core being under 1W. I was using the latest firmware for the board (4004) and updated it recently, but nothing changed.
I checked RAM, tried different flashing, and even considered reseating components, but nothing resolved it. The system runs smoothly in both Windows 11 and Pop!_OS, and a Live USB with Pop!_OS 21.10 worked too.
It looks like the problem might be hardware-related—possibly a faulty motherboard. I’m still waiting to test another board, but I want to rule out everything else before sending it back. Any advice would be greatly appreciated. Thanks.
Based on your Ryzen master setup, it seems the temperatures are likely within normal ranges or below 80°C. I’d consider re-seating the components. Also check the RAM and power cables. If the issue persists and power consumption remains high during testing, you might need to pursue a motherboard or CPU return warranty. Thanks for the clear and helpful post!
I don't usually run Ryzen Master. I mostly check temperatures with OpenHardwareMonitor or FanControl. Used to use AISuite, but stopped using it for FanControl now. CPU case temps rise up to around 90°C when under load—doesn't matter whether it's single or multi-threaded. (CCD temps are usually about 10°C lower than that.) Idle or light loads bring the case and CCDs to roughly 15-20°C above ambient, while the CCDs stay about 10°C cooler. Not ideal, but expected given the fan curves and limited airflow (Define R5). Apparently AMD said high 80s to low 90s are fine for Zen 3. Still, if this was about throttling, I'd expect much higher temps at such low frequencies (assuming readings are accurate). Thanks. If you need help, it should be reliable.
It has been there in the socket for a whole year, and I haven't taken it out yet.
Recent changes appear to help ease the problem briefly, even after restarting. It’s unclear what exactly causes it to return; adjusting Power Options settings like sleep delays seems to work temporarily. However, power measurements are now inaccurate again.
Sure, I understand you're not relying on it often, but I learned it could serve as an alternative solution.
A minor note about the situation - there was a DRAM diagnostic light on the screen while the PC was running (it froze completely, not responding to any power button press). I removed both memory modules, reinserted one, and ran Windows Memory Diagnostic which came back clean. I also tested Memtest86; it passed as well (though not the full suite). Since it didn’t fail completely, I plan to keep both sticks in place overnight to monitor performance. I also noticed the GSOD display appeared during testing, which could be a Windows feature rather than a hardware issue.