Breakthrough progress has been achieved at the Phase II Project of the National Supercomputing Shenzhen Center. LineShine, a fully independently developed supercomputer deployed at the site, delivers sustained computing performance beyond 2 ExaFLOPS, securing the top spot in global supercomputer rankings.
Against the massive energy consumption challenges brought by exascale-level computing, China IPPR leverages profound research accumulation and cutting-edge design concepts to develop the world’s largest centralized liquid cooling system. The innovative infrastructure solution significantly optimizes computing facility architecture and delivers superior energy efficiency performance.
Centered on intelligent integrated computing architecture, the LineShine supercomputer fully supports exascale computing power, exascale storage and petascale-level data communication. It integrates computing power, algorithm iteration and application delivery, covering benchmark scenarios across multiple industries and forming a replicable and scalable implementation model for high-performance computing infrastructure. The platform delivers robust technical support for industrial innovation and technological upgrading in the Greater Bay Area.
Building on its successful design experience for the project’s Phase I construction, China IPPR continued to provide full-process design and technical services for the Phase II supercomputing facility. The project adopts 14 sets of 20kV power distribution circuits and builds a highly reliable power supply core for the 2 ExaFLOPS supercomputer via high-voltage direct current power supply and full-link redundancy design. Equipped with self-developed full-liquid cooling heat dissipation technology, the site features the world’s largest centralized liquid cooling system.
Adopting a centralized CDU architecture, the design effectively addresses layout challenges arising from ultra-high single-cabinet power density and delivers outstanding liquid cooling efficiency. The secondary pipeline system stretches 3,214.7 meters in total length, while the overall system net weight reaches 243.9 tons.
The project establishes a high-load, high-temperature liquid cooling system with maximum primary and secondary main pipe diameters of 1,200 mm. The circulating water pump boasts an installed power of 315 kW, while the high-temperature liquid cooling system delivers a total cooling capacity of 57,900 kW, with the single plate heat exchanger achieving a maximum heat exchange capacity of 8,700 kW.
The finely optimized liquid cooling circulation architecture enables efficient heat dissipation for high-density computing equipment, maintaining the facility’s Power Usage Effectiveness (PUE) at an ultra-low level.
Focused on technological innovation and design optimization, China IPPR maintains leading industry expertise in data center design and consulting. The company prioritizes core technical breakthroughs including liquid cooling system development and pre-project technical comparison, continuously reducing PUE indicators and improving overall energy utilization efficiency. Such technical capabilities support the stable operation of large-scale supercomputing and intelligent computing systems, delivering reliable, green and high-efficiency customized solutions for modern data center construction.
Moving forward, China IPPR will further strengthen technological innovation capabilities, deepen its layout in high-performance computing and green data center sectors, and advance the low-carbon, high-efficiency upgrading of global data center infrastructure through continuous technical iteration and engineering practice.
