6th of November, 2024

Portada » NVIDIA Supercomputers and China Restrictions

NVIDIA Supercomputers and China Restrictions

31 octubre, 2024
English
Os supercomputadores NVIDIA e as restrições da China

NVIDIA supercomputers are among the world’s leading supercomputers, but NVIDIA faces restrictions on exports to China.

In October 2022, the U.S. Department of Commerce issued specific controls on China. The measures cover both advanced logic chip design software and semiconductor equipment. In addition, they include services that U.S. persons provide to produce advanced logic and memory chips.

Similarly, the Commerce Department restricted the export of certain chips to China, especially those with applications in artificial intelligence (AI) and supercomputing. The restrictions affect chips as well as software, hardware, equipment and technology used in the development and manufacture of these components to China, including Hong Kong and Macau, as well as Russia.

Among the impacted products are A100 and H100 integrated circuits, as well as DGX systems and any boards containing these circuits. In response, U.S. technology companies such as NVIDIA, AMD and Intel announced that they would manufacture chips for China with specifications just below the restricted levels.

NVIDIA Supercomputers

Researchers and developers turn to NVIDIA solutions to accelerate a wide range of important applications. These range from molecular dynamics simulation to weather forecasting. With more than 3,500 supported applications, NVIDIA computing powers promising areas of discovery, such as materials science, climate prediction, wind tunnel simulation, and genomics.

NVIDIA powers more than 75 percent of the supercomputers on the global TOP500 list. This includes 24 of the top 30 systems on the Green500 list. NVIDIA is currently the leading AI chip designer by revenue and usage in artificial intelligence research.

Supercomputing

In addition, certain future NVIDIA integrated circuits must meet licensing requirements if they meet certain performance and chip-to-chip I/O thresholds. This requirement also applies to systems or boards that integrate these circuits. In addition, licenses are required to export a variety of products, such as networking products, intended for certain end users and end uses in China.

NVIDIA’s computing solutions include supercomputing platforms and servers. These integrate GPUs, DPUs, interconnects, and high-performance computing (HPC) and artificial intelligence (AI) software packages. Finally, NVIDIA offers NVIDIA AI Enterprise software, the DGX Cloud service, and a growing collection of acceleration libraries, APIs, SDKs, and domain-specific application frameworks.

 

 

[themoneytizer id="51423-6"]