文化大學機構典藏 CCUR:Item 987654321/53067
English  |  正體中文  |  简体中文  |  全文笔数/总笔数 : 46867/50733 (92%)
造访人次 : 11872169      在线人数 : 418
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
搜寻范围 查询小技巧:
  • 您可在西文检索词汇前后加上"双引号",以获取较精准的检索结果
  • 若欲以作者姓名搜寻,建议至进阶搜寻限定作者字段,可获得较完整数据
  • 进阶搜寻
    主页登入上传说明关于CCUR管理 到手机版


    jsp.display-item.identifier=請使用永久網址來引用或連結此文件: https://irlib.pccu.edu.tw/handle/987654321/53067


    题名: 利用GPU加速Skip Lists查詢效能之最佳化之研究
    Maximizing Query Performance for Skip Lists for GPU
    作者: 張新羽
    贡献者: 資訊工程學系
    关键词: 顯示卡
    查詢
    資料結構
    跳錶
    GPUs
    Tree
    Query
    SkipList
    日期: 2023
    上传时间: 2024-03-01 13:56:54 (UTC+8)
    摘要: 本論文介紹了一項研究,旨在改善(improve)GPU平台上跳表(Skip Lists)的查詢性能。Skip Lists是一種以鏈表為基礎、有額外層級的資料結構,它可以提升鏈表的查詢效能。同時,有許多的專案也使用了Skip Lists進行資料索引。但是,目前Skip Lists由於隨機化偶爾無法達到效能最大化,故我們實作了一個演算法模型,使Skip Lists無須改變其資料結構即可從中得到效能提升。我們通過在Skip Lists插入時進行資料取樣索引,以確定資料可能的層級,從而在查詢時間接提升Skip Lists的效能。同時,我們還進行了實驗,實驗結果表明相對於完全隨機化的Skip Lists,該改進能使其查詢的效能得到一定的提升。
    This essay presents a study aimed at improving the query performance of Skip Lists on GPU platforms. Skip Lists is a data structure that is based on linked lists and that has extra layers, it can improve the query performance of linked lists. Meanwhile, many projects also use Skip Lists for data indexing. However, at present, due to randomization, the performance of Skip Lists sometimes cannot be maximized. Therefore, we have implemented an algorithm model, so that the performance of Skip Lists can be improved without changing its data structure. We perform data sampling and indexing when Skip Lists is inserted to determine the possible levels of data, thereby indirectly improving the performance of Skip Lists at query time. Meanwhile, we also conduct experiments, and the experimental results show that compared with completely randomized Skip Lists, this improvement can improve the query performance to a certain extent.
    显示于类别:[資訊工程學系] 博碩士論文

    文件中的档案:

    档案 描述 大小格式浏览次数
    index.html0KbHTML76检视/开启


    在CCUR中所有的数据项都受到原著作权保护.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - 回馈