Abstract:

With the rapid development of digital economy and Artificial Intelligence (AI) industry, the demand for computing power shows explosive growth. The contradiction between the high energy consumption of data centers and the pressure of new energy consumption in new power systems is increasingly prominent. As an important path to solve the spatial-temporal supply and demand mismatch between computing power and power and achieve the green and low-carbon development goals, computing-power synergy has attracted extensive attention from industry and academia. Firstly, this paper expounds the core concept and development background of computing-power synergy, and systematically sorts out the policy orientation, technical progress and application situation at home and abroad. Secondly, it discusses in depth the key technologies of AI-empowered computing-power synergy from four dimensions: multi-element time series intelligent prediction, 4D spatial-temporal joint scheduling, adaptive migration of heterogeneous computing power, and quantitative evaluation of multi-dimensional value. Then, it introduces China Mobile’s technical layout and practical exploration in this field, and analyzes the application effects with one typical case: Shanghai-Anhui-Xinjiang cross-domain computing-power synergy. Finally, it summarizes the current challenges and prospects the future development direction.

Key words: computing-power synergy, AI, computing power network, new power system, demand response