Chinese Algorithm Supercharges Nvidia GPUs for Science
中国算法大幅提升英伟达GPU科学计算能力
Chinese researchers have developed a high-performance algorithm that boosts Nvidia GPU speed by 800 times for solving complex material design problems. Created at Shenzhen MSU-BIT University, the algorithm enhances peridynamics (PD), a cutting-edge method for simulating cracks and fractures. This breakthrough allows industries like aerospace and defense to run advanced simulations on affordable, widely available GPUs, bypassing restrictions from US sanctions.
中国研究人员开发出一种高性能算法,使英伟达GPU在解决复杂材料设计问题时提速800倍。该算法由深圳北理莫斯科大学团队研发,优化了用于模拟裂纹和断裂的前沿方法——围观动力学(PD)。这一突破使航空航天和国防等行业能够在低成本、广泛可用的GPU上运行高级仿真,规避美国制裁带来的限制。
