China is looking at NVIDIA off-the-shelf solutions for use in modern weapons

The Chinese academic journal Propulsion Technology published a study by the Beijing Energy Engineering Research Institute and Dalian University of Technology. It states that increasing computing performance even with a low-power AI module from NVIDIA is crucial for “real-time optimization” of various aircraft systems, such as fuel supply and engine fault diagnostics.

The paper discusses how NVIDIA’s low-cost Jetson TX2i system-on-chip for industrial applications can significantly enhance the capabilities of hypersonic weapons to reach speeds in excess of Mach 7.

This SoC, which is not restricted by U.S. export controls and is much more affordable than expensive options, could enhance China’s military capabilities. The NVIDIA Jetson TX2i processor was originally designed for industrial applications, but the research team adapted it to work in hypersonic aircraft.

Aircraft benefit greatly from the chip’s ability to process complex aerodynamic computational models. This computing power reduces the time required for critical calculations from several seconds to 25 milliseconds.

Using the TX2i module significantly reduces the cost of hypersonic technology research and development. With a peak performance of 1.26 FP16 TFLOPS while consuming up to 20 watts, the TX2i offers significant computing power at a fraction of the cost of more powerful NVIDIA AI chips.

Jetson TX2i Module specifications:

– NVIDIA Pascal™ architecture with 256 NVIDIA CUDA cores, 1.3 TFLOPS (FP16);
– Dual-core Denver 2 64-bit CPU and quad-core ARM A57 Complex;
– 8GB 128-bit LPDDR4 (ECC support), 1600MHz – 51.2 GB/s;
– 32GB eMMC 5.1;
– Voltage input: 9V – 19.6V DC;
– Module power: 10W – 20W.

The study also emphasized the use of heterogeneous computing on CPUs and GPUs, introducing a new architecture to address the challenges posed by the sequential nature of hypersonic flow simulations. This solution is crucial because it allows for more efficient processing of the complex computations required by such performance-demanding applications.

Despite the potential of using U.S.-made chips such as the TX2i, the team recognized that Chinese chips could meet the needs of the country’s military applications just as well. This indicates a possible shift toward greater self-sufficiency in critical military technologies.

The fact that China is looking at NVIDIA’s off-the-shelf solutions for use in modern weapons is a clear indicator of the strategic importance of integrating advanced computing solutions that can challenge traditional defense systems.