Global pandemics can wreak havoc and lead to significant social, economic, and personal losses. Preventing the spread of infectious diseases requires implementing interventions at different levels of government, and evaluating the potential impact and efficacy of those preemptive measures. Agent-based modeling can be used for detailed studies of epidemic diffusion and possible interventions. We present Loimos, a highly parallel simulation of epidemic diffusion written on top of Charm++, an asynchronous task-based parallel runtime. Loimos uses a hybrid of time-stepping and discrete-event simulation to model disease spread. We demonstrate that our implementation of Loimos is able to scale to large core counts on an HPC system. In particular, Loimos is able to simulate a US-scale synthetic interaction network in an average of 1.497 seconds per simulation day when executed on 16 nodes on Rivanna at the University of Virginia, processing around 428 billion interactions (person-person edges) in under five minutes for an average of 1.4 billion traversed edges per second (TEPS).

翻译：全球大流行病可能造成严重破坏，导致重大社会、经济和个人损失。防控传染病传播需要在不同政府层级实施干预措施，并评估这些预防性措施的潜在影响与效力。基于智能体的建模可用于疫情扩散及相关干预措施的精细研究。我们提出Loimos——一种基于Charm++异步任务并行运行时构建的高度并行化疫情扩散仿真系统。Loimos采用时间步长与离散事件混合仿真方法对疾病传播过程建模。实验证明，Loimos的实现在高性能计算系统上可扩展至大规模核心数。具体而言，在弗吉尼亚大学Rivanna集群的16个节点上运行时，Loimos能以平均每仿真日1.497秒的速度模拟美国规模级合成交互网络，在不足五分钟内处理约4280亿次人际交互（个体间边接触），实现每秒约14亿条边的遍历速度（TEPS）。