On the Qinghai clear power and inexperienced computing scheduling heart in Xining metropolis, real-time knowledge on the display screen displays energy technology from totally different sources, together with wind, photo voltaic and hydroelectric energy, in addition to the power consumption metrics of computing facilities.
Leveraging its distinctive geographical benefits, Xining, capital of Qinghai province in northwest China, is selling the synergy between inexperienced electrical energy and clever computing.
Lately, the fast improvement of generative AI has led to a meteoric rise in functions using giant AI fashions, leading to an explosive demand for computing energy.
As of June 2024, China had greater than 8.3 million operational computing racks, with a computing capability reaching 246 EFLOPS. This substantial enhance in computing capability has been accompanied by a big rise in power consumption. In line with the China Academy of Data and Communications Know-how, knowledge facilities, together with computing facilities, accounted for about 1.6 % of the full electrical energy consumption in China in 2023.
Qinghai province, located on the Qinghai-Xizang Plateau, has pure circumstances and useful resource benefits that make it a pacesetter in clear power and an excellent location for creating inexperienced computing.
Fan Kewei, common supervisor of the Qinghai clear power and inexperienced computing scheduling heart, defined that computing energy utilization is intently tied to particular utility eventualities. In comparison with conventional knowledge storage and common computing, the quickly rising clever computing energy is extra appropriate for integration with inexperienced electrical energy.
“Clever computing shouldn’t be all the time lively. Duties like large-scale rendering and knowledge coaching are usually initiated solely when wanted, and will be completed when there’s an abundance of renewable power and decrease electrical energy costs,” Fan stated.
Presently, Qinghai’s clever computing capability has reached 1.23 EFLOPS, and the province has established the nation’s first 100% clear power traceable massive knowledge heart. Moreover, its giant knowledge facilities keep a median energy utilization effectiveness (PUE) beneath 1.2, putting them on the forefront nationally.
Digitalization and intelligence are vital within the development of a brand new energy system, with ample computing energy poised to offer important help.
The options of randomness, intermittency and volatility related to renewable power make balancing electrical energy provide and demand more and more advanced as extra renewable sources hook up with the grid. AI functions, represented by giant fashions, can carry out computational evaluation on meteorological circumstances, historic renewable power technology knowledge, and electrical energy demand, facilitating correct predictions of renewable power technology and consumption, thereby enhancing the protection and stability of the ability system.
Utilizing massive knowledge, cloud computing and blockchain applied sciences, the province additionally developed the electricity-carbon calculation mannequin to attain a complete view of carbon emissions in key industries, parks and enterprises, serving to to appreciate the objective of carbon discount and inexperienced improvement.
Editor：ZHONG Jianli
