论文标题
多个GPU
Abelian-Higgs cosmic string evolution with multiple GPUs
论文作者
论文摘要
宇宙学相变的拓扑缺陷是通过千里机制的。宇宙字符串和超串可能会导致特别有趣的天体物理和宇宙学后果，但是目前，该研究受到准确的数值模拟的可用性的限制，而硬件资源和计算时间又被瓶颈瓶装。为了消除这种瓶颈，在最近的工作中，我们介绍并验证了针对本地Abelian-Higgs弦乐网络的GPU加速演变代码。尽管这导致速度显着提高，但它仍然受图形加速器上可用的物理内存的限制。在这里，我们通过实施和验证早期代码的多个GPU扩展，并进一步证明其良好的可扩展性，无论是在强和弱缩放方面，都可以进一步证明其良好的可扩展性。 $ 8192^3 $生产运行，使用$ 4096 $ GPU，在Piz Daint SuperComputer上的售价为$ 33.2 $分钟。
Topological defects form at cosmological phase transitions by the Kibble mechanism. Cosmic strings and superstrings can lead to particularly interesting astrophysical and cosmological consequences, but this study is is currently limited by the availability of accurate numerical simulations, which in turn is bottlenecked by hardware resources and computation time. Aiming to eliminate this bottleneck, in recent work we introduced and validated a GPU-accelerated evolution code for local Abelian-Higgs strings networks. While this leads to significant gains in speed, it is still limited by the physical memory available on a graphical accelerator. Here we report on a further step towards our main goal, by implementing and validating a multiple GPU extension of the earlier code, and further demonstrate its good scalability, both in terms of strong and weak scaling. A $8192^3$ production run, using $4096$ GPUs, runs in $33.2$ minutes of wall clock time on the Piz Daint supercomputer.
