My PC hard reboots when the power plan's processor state is at 100%

[clutchrence]

Hello everyone!

As the title suggests, my PC randomly freezes when the maximum processor state is at 100% (whether on high performance or balanced, as long as the processor state is at 100% or boost clock). It hard reboots even I'm just casually using it (browsing and listening to music) or playing games.

For now, I'm only with 99% maximum and minimum processor state in the power configuration plan to prevent random freezes while using my PC. Since it is at 99% maximum and minimum processor state, the clock speed currently sits at 2.81 GHz even at max load.

PC SPECS  

CPU	Intel i5-8400 (with Intel Stock Cooler)
GPU	Gigabyte AORUS RX 570 (4GB)
MOBO	Gigabyte H310M H 2.0
RAM	Samsung Kingston 8GB (2666MHz) & SK Hynix V-Color 8GB (2666MHz)
SSD & HDD	240GB SSD & HDD 1TB
PSU	Thermaltake TR2 S 600W

[LeonWaksman]

Download and run HwiNFO (you may use the portable version).  Monitor the temperatures of various parts (i.e. CPU, SSD, RAM,...).

Leon

[clutchrence]

I've been monitoring it and I don't see any suspicions, and it seems like everything is stable in temperatures.

[pressed_for_time]

You seem to have a processor and motherboard combo that no longer supports turbo boost. Could be a motherboard or power supply issue as well  as the CPU itself.

[LeonWaksman]

1. You may run CPU test using Intel® Processor Diagnostic Tool. 

2. Did you updated BIOS and all Drivers (including GPU Driver)?

3. You may disable the RX570 GPU and run Integrated GPU only. This will verify that the problem was not caused by the GPU. 

Leon

[PC1997]

Hello, @clutchrence

In addition to the points already mentioned, I’ve prepared a guide that I compiled with the help of LLMs.

Short version: locking Windows to 99% disables Intel Turbo Boost, keeping your i5-8400 at its 2.8 GHz base clock. The fact that your PC only freezes/reboots when Turbo is allowed (100%) strongly points to an instability that appears when the CPU briefly draws extra current (boost). The usual culprits—in order of likelihood—are: marginal PSU or power delivery, overheating/VRM stress, BIOS/microcode/driver issues, or mixed RAM instability.

Here’s a focused plan to pin it down and fix it:


0) Quick sanity tweaks


• In your Windows power plan, set Minimum processor state = 5%, Maximum = 99% (you currently have min=99%, which just wastes power/heat at idle).

• Make sure the CPU fan is on the CPU_FAN header and spins freely; blow out dust.


1) Check what Windows is seeing


• Open Event Viewer → Windows Logs → System:

• Kernel-Power (Event 41) after a sudden reboot = power loss (PSU/VRM/short/overheat).

• WHEA-Logger (Event 18/19/47) with “cache hierarchy / internal parity” = CPU/VRM/IMC instability.

• Display driver / LiveKernelEvent 141 = GPU/driver (less likely for a hard reboot). Note what you find.


2) Temperatures & limits


• Use HWiNFO64 (Sensors) while you briefly allow 100% max again:

• Watch CPU Package temp, VRM MOS (if exposed), CPU Core power, and Thermal Throttling flags.

• If temps spike into the 90s °C right before a reset, reseat the cooler and replace thermal paste. The stock cooler should be fine for an i5-8400, so if temps are fine but it still reboots, think power delivery.


3) Isolate power vs. components


• GPU-off test: Remove power to the RX 570 and plug your monitor into the iGPU (Intel UHD 630). Uninstall AMD drivers (use DDU if you can) and run at 100% CPU allowed.

• Stable now? The GPU/PSU combo is the issue (transient spikes from GPU + CPU boost).

• Per-stick RAM test: Run with only one 8 GB stick at a time (no XMP on H310 anyway). If one stick crashes at 100% and the other doesn’t, you’ve found it.

• PSU cable check: Ensure the 24-pin and CPU 8-pin are fully seated; GPU 6/8-pin firmly seated; no splitters/adapters if possible.


4) Stress tests (short, careful runs)


• Prime95 Small FFTs (CPU only) for 5–10 minutes with 100% enabled. If it reboots quickly, it’s CPU/power, not GPU.

• FurMark/Unigine (GPU only) at stock clocks. If GPU-only is fine but Prime95 reboots, it’s CPU/power/VRM. If both together (OCCT “Power” test) cause instant reset, that screams PSU.

Keep runs short—your goal is to reproduce and observe, not roast the hardware.


5) Firmware & drivers


• Update BIOS for the Gigabyte H310M H 2.0 to the latest stable release.

• Install latest Intel Chipset + MEI drivers (Gigabyte support page).

• Perform a Clean AMD driver install (DDU → latest Adrenalin).

• In BIOS, try Disabling Intel Turbo Boost (for testing). If that mirrors the 99% fix exactly, it reinforces the power/VRM suspicion.


6) If it’s power delivery (very likely here)


Your Thermaltake TR2 S 600W is a budget, older-design unit that can struggle with transient spikes. Turbo boost can create fast current transients even during light tasks. Replacing it with a modern, well-reviewed unit often fixes this kind of “100% only” crash.

Good replacement targets (your rig needs ~250–300W max real load):

• 550–650 W, 80+ Gold (or solid Bronze) from reputable lines, e.g. Corsair RM/RMx or CX-M (2017+), Seasonic Focus GX, be quiet! Pure Power 12/11, EVGA G6/G5. Even a quality 450–550 W would be plenty for i5-8400 + RX 570, but headroom is nice.


7) Workarounds if you can’t swap parts yet


• Keep Max processor state = 99% (and Min = 5%) to stay stable.

• Or use ThrottleStop / Intel XTU to limit Turbo power (reduce PL1/PL2) or apply a light CPU undervolt (if your BIOS/microcode still allows it). Reducing PL2 in particular trims those spike currents that trip weak PSUs/VRMs.

• On the GPU, set a small power limit −10% in Radeon software; it often reduces combined spikes with negligible performance loss.


Physical checks


• Reseat RAM and GPU.

• Inspect the motherboard around the CPU socket for bulged/leaking caps.

• Ensure the case has at least one intake + one exhaust fan for VRM/CPU airflow.


What I’d do first in your case


• Set Min=5% / Max=100% just for testing, start HWiNFO sensors.

• Run Prime95 Small FFTs: if it reboots within minutes with temps <90 °C → power/VRM.

• Try iGPU-only: if that stabilizes 100%, the PSU is the prime suspect.

• Update BIOS; if still bad, replace the PSU with one of the units above.