Workflow
珀尔马特超级计算机
icon
Search documents
超级计算机实现量子微芯片高精度模拟
Ke Ji Ri Bao· 2025-11-19 22:53
Core Insights - The collaboration between Lawrence Berkeley National Laboratory and the University of California, Berkeley has achieved a breakthrough in simulating quantum microchips using the "Perlmutter" supercomputer, which utilized over 7,000 NVIDIA GPUs, laying a solid foundation for optimizing next-generation quantum technologies [1][2] - The simulation allows for precise modeling of quantum chips, enabling performance predictions before manufacturing, thus identifying potential issues to ensure the final product meets design expectations [1] Group 1 - The simulated quantum chip features a multilayer structure measuring only 10 square millimeters in size, with a thickness of 0.3 millimeters and etching widths as fine as 1 micron [1] - The research utilized the full resources of the "Perlmutter" supercomputer, calling upon 7,168 GPUs simultaneously to capture the chip's structural and functional characteristics within 24 hours [1] - Unlike traditional simulations that treat chips as "black boxes," this research leverages the advantages of parallel computing to reveal the internal physical mechanisms of the chip [1] Group 2 - The simulation employed full-wave physical precision, taking into account material properties (including metals like niobium), chip layout, and resonator construction, while also modeling the dynamic interactions between qubits and circuit components [2] - The team plans to continue a series of simulation experiments to deepen the quantitative understanding of chip design and explore its adaptability in larger systems [2] - Future focus will be on the resonance characteristics between qubits and circuit components, with benchmark validation through multi-frequency simulations [2]