arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1106.4964hep-latcs.PF
收藏

重叠算子在多GPU上的高效实现

Efficient implementation of the overlap operator on multi-GPUs

Andrei Alexandru, Michael Lujan, Craig Pelissier, Ben Gamari, Frank X. Lee

摘要:

格点量子色动力学计算是显示GPU在高性能计算领域潜力的首批应用之一。我们的兴趣是找到有效使用GPU进行重叠运算的方法。这些代码占用大量内存,需要并行使用多个GPU。在本文中,我们展示了我们用来有效地实现这个操作符的方法。我们在具有类似互连的GPU集群和CPU集群上运行我们的代码。我们发现,为了匹配性能,CPU集群需要比GPU多20-30倍的CPU核心。

英文摘要:

Lattice QCD calculations were one of the first applications to show the potential of GPUs in the area of high performance computing. Our interest is to find ways to effectively use GPUs for lattice calculations using the overlap operator. The large memory footprint of these codes requires the use of multiple GPUs in parallel. In this paper we show the methods we used to implement this operator efficiently. We run our codes both on a GPU cluster and a CPU cluster with similar interconnects. We find that to match performance the CPU cluster requires 20-30 times more CPU cores than GPUs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑