This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Tianhe Supercomputer | |
|---|---|
| Name | Tianhe Supercomputer |
| Manufacturer | National University of Defense Technology |
| Release | 2010s |
| Cpu | Intel Xeon (in upgraded systems) |
| Gpu | NVIDIA Tesla (in later configurations) |
| Memory | multiple terabytes |
| Storage | petabyte-class storage arrays |
| Os | customized Linux distributions |
| Flops | petaflops-class peak |
| Location | National Supercomputer Center, Guangzhou; National Supercomputer Center, Tianjin |
Tianhe Supercomputer
The Tianhe Supercomputer family comprises a series of high-performance computing systems developed by the National University of Defense Technology and deployed at Chinese national supercomputing centers in Guangzhou and Tianjin. These systems have been central to national programs in weather forecasting, computational fluid dynamics, seismic simulation, nuclear fusion research, and large-scale big data modeling, featuring in international rankings and attracting attention from institutions such as Top500 and Green500.
Tianhe systems were designed to deliver petascale performance by integrating custom interconnects, multicore processors, and accelerator technologies to support projects led by entities including the China Meteorological Administration, the National Natural Science Foundation of China, and the China Academy of Engineering Physics. The deployments serve research centers at the National Supercomputer Center in Guangzhou and the National Supercomputer Center in Tianjin, and interface with programs run by the Ministry of Science and Technology (PRC), the People's Liberation Army's research units, and civilian universities like the Tsinghua University and the Peking University.
The Tianhe architecture evolved across generations, initially combining many-core processors with proprietary networking switches developed by the National University of Defense Technology. Later iterations incorporated Intel Xeon processors and NVIDIA Tesla accelerators, reflecting collaborations and procurement involving Intel Corporation and NVIDIA Corporation. The systems use large shared-memory nodes, high-bandwidth interconnect fabrics inspired by designs from the Cray Inc. and informed by standards such as InfiniBand. Storage is provided by parallel file systems compatible with technologies from IBM and Seagate Technology, while memory hierarchies span DRAM and node-local caches similar to configurations used by the Oak Ridge National Laboratory and the Lawrence Livermore National Laboratory.
Tianhe machines achieved top positions on the TOP500 list, briefly ranking as the world’s fastest supercomputer following benchmark runs that reported peak performance in the petaflops range. Performance evaluations referenced LINPACK benchmarks used by TOP500 and energy metrics assessed by the Green500 project. Competing systems during the same eras included Jaguar (supercomputer), Sequoia (supercomputer), and later Summit (supercomputer) and Fugaku, placing Tianhe within a global cohort spanning centers like Argonne National Laboratory and the Lawrence Berkeley National Laboratory.
Tianhe systems run customized Linux-based operating environments, incorporating components derived from distributions used by large installations at Los Alamos National Laboratory and Sandia National Laboratories. System software stacks include MPI implementations such as OpenMPI and vendor-tuned message passing libraries, numerical libraries like LAPACK and FFTW, and middleware for job scheduling akin to SLURM Workload Manager and PBS (software). Application codes ported to Tianhe include climate models from European Centre for Medium-Range Weather Forecasts, fusion codes similar to those used at ITER, and molecular dynamics packages comparable to GROMACS.
Tianhe platforms support research in atmospheric modeling for institutions like the World Meteorological Organization partners, seismic hazard simulations used by the China Earthquake Administration, aerodynamic design for aerospace firms such as AVIC, and materials science studies related to work at the Institute of Metal Research (Chinese Academy of Sciences). They have been used for data analytics projects by telecom companies akin to China Mobile and for genomics pipelines similar to those run at the Beijing Genomics Institute.
Development traces to initiatives launched by the National University of Defense Technology with funding and strategic direction from national programs associated with the 863 Program and later the National High Technology Research and Development Program. Early prototype systems debuted in the late 2000s, with operational deployments and upgrades across the 2010s that involved collaboration with suppliers including Inspur and research exchanges with institutions such as Rensselaer Polytechnic Institute and Technische Universität München.
Energy management for Tianhe installations required advanced cooling systems, integrating concepts used at facilities like the European Centre for Medium-Range Weather Forecasts data centers and employing chilled-water loops, raised-floor airflow designs, and custom heat exchangers similar to those adopted by Google and Microsoft hyperscale centers. Energy efficiency metrics were tracked for submissions to Green500, prompting optimizations in power delivery and workload placement to reduce PUE values comparable to practices at Oak Ridge National Laboratory.
Tianhe deployments intersected with geopolitical and export-control debates involving agencies such as the United States Department of Commerce and firms like Intel Corporation and NVIDIA Corporation due to concerns about advanced computing for defense-related research. Academic collaborations with institutions including Massachusetts Institute of Technology and Stanford University proceeded alongside scrutiny about technology transfer, while press coverage from outlets such as The New York Times and South China Morning Post highlighted discussions on dual-use capabilities and export regulations enacted by bodies like the Bureau of Industry and Security.