Novel Solution for the Phone & Server Overheating Problem
Novel Solution for the Phone & Server Overheating Problem
NYCU Research Team Breaks Through Multi-Core Chip Heat Management
Tags: National Yang Ming Chiao Tung University, Taiwan, Electronics & Robotics, Computing Technology
A research team from NYCU’s CERES Lab, led by Associate Professor Kun-Chih Chen, has developed an innovative machine learning-based thermal management system for multi-core chips. This technology uses adaptive reinforcement learning to predict and control temperatures in real-time, preventing overheating in devices like smartphones and computers. It improves system performance by reducing temperature prediction errors and efficiently managing heat through proactive measures. Applications include multi-core chips in smartphones, servers, and other high-performance devices. The team's work was awarded the 2024 IEEE TVLSI Best Paper Award.
IP Type or Form Factor: Software & Algorithm
TRL: 4 - minimum viable product built in lab
Industry or Tech Area: Semiconductors; Computing Infrastructure & Networks
