🛠 Thermal management and monitoring under heavy load

Objectifs

Learn to monitor the temperatures of hybrid cores (Lion Cove and Skymont) and configure thermal thresholds. This tutorial uses Linux utilities to visualize the impact of the included aluminum heatsink on boost frequencies (up to 4.8 GHz).

Matériel nécessaire

• LattePanda Mu Ultra with installed aluminum heatsink
• Mini Carrier Board
• Optional external fan for extreme tests
• SSH terminal or connected screen

Étapes

1. Install thermal monitoring tools specific to modern Intel processors.

sudo apt install lm-sensors intel-gpu-tools stress-ng

1. Detect available thermal sensors on the module.

sensors-detect
sudo service procps restart

1. Read current temperatures of different cores. Note that P-cores and E-cores have different thresholds.

sensors

1. Launch an intensive CPU load to test active cooling. Use stress-ng to saturate cores.

stress-ng --cpu 8 --timeout 60s

1. During load, monitor frequency and temperature in real-time to check if thermal throttling occurs.

watch -n 1 "grep MHz /proc/cpuinfo | head -n 2; sensors"

1. Configure the CPU frequency governor to optimize performance/thermal balance.

sudo apt install cpufrequtils
cpufreq-set -g powersave

Vérification

Temperatures must not exceed 100°C under sustained load thanks to the included heatsink. If frequencies drop significantly, check the thermal contact of the aluminum heatsink.

Pour aller plus loin

Automate fan management via the carrier board GPIO or explore power profiles in BIOS to limit TDP.