李怀明, 邓仰东. 基于图形处理器的高性能跳表(Skiplist)数据结构[J]. 微电子学与计算机, 2014, 31(12): 1-5.
引用本文: 李怀明, 邓仰东. 基于图形处理器的高性能跳表(Skiplist)数据结构[J]. 微电子学与计算机, 2014, 31(12): 1-5.
LI Huai-ming, DENG Yang-dong. High Performance Skiplist for GPU Computing[J]. Microelectronics & Computer, 2014, 31(12): 1-5.
Citation: LI Huai-ming, DENG Yang-dong. High Performance Skiplist for GPU Computing[J]. Microelectronics & Computer, 2014, 31(12): 1-5.

基于图形处理器的高性能跳表(Skiplist)数据结构

High Performance Skiplist for GPU Computing

  • 摘要: 提出了一种高效率、适合GPU的跳表结构及其相应例程,核心思想是将包含指针的操作转化为数组操作,从而充分发挥GPU的计算能力.实验结果证明,该数据结构的插入和删除操作相对目前最好的GPU结果分别改善6.8倍和9.6倍.

     

    Abstract: This work proposes a high performance skiplist data structure that can be efficiently implemented on graphics processors.The key idea is to convert the pointer operations into arrayed manipulations so as to unleash the computing power of GPUs.Experimental results prove that the proposed data structure outperform the best previous work by 6.8Xand 9.6Xin insertion and delete operations,respectively.

     

/

返回文章
返回