Adaptive techniques for clustered N-body cosmological simulations
Adaptive techniques for clustered N-body cosmological simulations
复制标题
集群 N 体宇宙学模拟的自适应技术
DOI：
10.1186/s40668-015-0007-9
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
F. Governato
中科院分区：
文献类型：
--
作者：
Harshitha Menon;Lukasz Wesolowski;G. Zheng;Pritish Jetley;L. Kalé;T. Quinn;F. Governato
Changa是一个用Charm++实现的N体宇宙学模拟应用程序。在本文中，我们提出了CHAGA的并行设计，并解决了由于集群数据集的高动态范围而带来的许多挑战。我们提出了基于自适应技术的优化方案。我们评估了CHAGA在高度聚集的数据集上的性能：一个是25MPC体积的20亿粒子实现的z∼0$zSIM0$快照，另一个是矮小星系的5200万粒子多分辨率实现的快照。对于25MPC的体积，我们显示了高达128K核的Blue Waters强大的可扩展性。我们还演示了在20亿粒子模拟的多步运行中将内核扩展到128K。虽然多步运行的可伸缩性不如单步运行，但128K核的吞吐量要高出2倍。我们还展示了在具有120亿和240亿粒子的两个大型均匀数据集上，Blue Waters高达512K核的强大可伸缩性。
ChaNGa is an N-body cosmology simulation application implemented using Charm++. In this paper, we present the parallel design of ChaNGa and address many challenges arising due to the high dynamic ranges of clustered datasets. We propose optimizations based on adaptive techniques. We evaluate the performance of ChaNGa on highly clustered datasets: a z∼0$z \sim0$ snapshot of a 2 billion particle realization of a 25 Mpc volume, and a 52 million particle multi-resolution realization of a dwarf galaxy. For the 25 Mpc volume, we show strong scaling on up to 128K cores of Blue Waters. We also demonstrate scaling up to 128K cores of a multi-stepping run of the 2 billion particle simulation. While the scaling of the multi-stepping run is not as good as single stepping, the throughput at 128K cores is greater by a factor of 2. We also demonstrate strong scaling on up to 512K cores of Blue Waters for two large, uniform datasets with 12 and 24 billion particles.