This paper presents our endeavors in developing the large-scale, ultra-high-resolution E3SM Land Model (uELM), specifically designed for exascale computers furnished with accelerators such as Nvidia GPUs. The uELM is ...This paper presents our endeavors in developing the large-scale, ultra-high-resolution E3SM Land Model (uELM), specifically designed for exascale computers furnished with accelerators such as Nvidia GPUs. The uELM is a sophisticated code that substantially relies on High-Performance Computing (HPC) environments, necessitating particular machine and software configurations. To facilitate community-based uELM developments employing GPUs, we have created a portable, standalone software environment preconfigured with uELM input datasets, simulation cases, and source code. This environment, utilizing Docker, encompasses all essential code, libraries, and system software for uELM development on GPUs. It also features a functional unit test framework and an offline model testbed for comprehensive numerical experiments. From a technical perspective, the paper discusses GPU-ready container generations, uELM code management, and input data distribution across computational platforms. Lastly, the paper demonstrates the use of environment for functional unit testing, end-to-end simulation on CPUs and GPUs, and collaborative code development.展开更多
【目的】针对极限学习机(ELM)神经网络在室内可见光定位(VLP)中收敛不稳定,易陷入局部最优状态,导致定位精度降低的问题,文章引入了麻雀搜索算法(SSA)确定ELM神经网络的初始权值和阈值,提出了SSA-ELM神经网络算法。【方法】首先,采集定...【目的】针对极限学习机(ELM)神经网络在室内可见光定位(VLP)中收敛不稳定,易陷入局部最优状态,导致定位精度降低的问题,文章引入了麻雀搜索算法(SSA)确定ELM神经网络的初始权值和阈值,提出了SSA-ELM神经网络算法。【方法】首先,采集定位区域内接收信号强度(RSS)与位置信息作为指纹数据;然后,训练SSA-ELM神经网络并得到预测模型,将测试集数据输入预测模型得到待测位置的定位结果;最后,设计了仿真实验和测试平台。【结果】仿真表明,在立体空间模型中0、0.3、0.6和0.9 m 4个接收高度,平均误差分别为1.73、1.86、2.18和3.47 cm,与反向传播(BP)、SSA-BP和ELM定位算法相比,SSA-ELM神经网络算法定位精度分别提高了83.55%、45.71%和26.26%,定位时间分别降低了36.48%、17.69%和6.61%。实验测试表明,文章所提SSA-ELM神经网络算法的平均定位误差为3.75 cm,比未优化的ELM神经网络定位精度提高了16.38%。【结论】SSA对ELM神经网络具有明显的优化作用,能够显著降低定位误差,减少定位时间。展开更多
文摘This paper presents our endeavors in developing the large-scale, ultra-high-resolution E3SM Land Model (uELM), specifically designed for exascale computers furnished with accelerators such as Nvidia GPUs. The uELM is a sophisticated code that substantially relies on High-Performance Computing (HPC) environments, necessitating particular machine and software configurations. To facilitate community-based uELM developments employing GPUs, we have created a portable, standalone software environment preconfigured with uELM input datasets, simulation cases, and source code. This environment, utilizing Docker, encompasses all essential code, libraries, and system software for uELM development on GPUs. It also features a functional unit test framework and an offline model testbed for comprehensive numerical experiments. From a technical perspective, the paper discusses GPU-ready container generations, uELM code management, and input data distribution across computational platforms. Lastly, the paper demonstrates the use of environment for functional unit testing, end-to-end simulation on CPUs and GPUs, and collaborative code development.
文摘【目的】针对极限学习机(ELM)神经网络在室内可见光定位(VLP)中收敛不稳定,易陷入局部最优状态,导致定位精度降低的问题,文章引入了麻雀搜索算法(SSA)确定ELM神经网络的初始权值和阈值,提出了SSA-ELM神经网络算法。【方法】首先,采集定位区域内接收信号强度(RSS)与位置信息作为指纹数据;然后,训练SSA-ELM神经网络并得到预测模型,将测试集数据输入预测模型得到待测位置的定位结果;最后,设计了仿真实验和测试平台。【结果】仿真表明,在立体空间模型中0、0.3、0.6和0.9 m 4个接收高度,平均误差分别为1.73、1.86、2.18和3.47 cm,与反向传播(BP)、SSA-BP和ELM定位算法相比,SSA-ELM神经网络算法定位精度分别提高了83.55%、45.71%和26.26%,定位时间分别降低了36.48%、17.69%和6.61%。实验测试表明,文章所提SSA-ELM神经网络算法的平均定位误差为3.75 cm,比未优化的ELM神经网络定位精度提高了16.38%。【结论】SSA对ELM神经网络具有明显的优化作用,能够显著降低定位误差,减少定位时间。
