CAJ | 학술논문

以Tilera公司的TLR26480处理器开发板为硬件平台，采用SMP-LINUX操作系统，应用TILE64多核开发环境（ MDE）开展了并行数据压缩的研究。无损数据压缩采用并行的GZIP算法-mgz-ip，在获得很好压缩比的同时，可以大大缩短压缩时间。 mgzip是基于共享存储的并行压缩策略，流水线并行与数据并行混合的并行模式。研究表明，基于TILE64多核平台上的并行数据压缩为高能物理实验中的数据获取系统的数据压缩需求提供了可行性方案。
이Tilera공사적TLR26480처리기개발판위경건평태，채용SMP-LINUX조작계통，응용TILE64다핵개발배경（ MDE）개전료병행수거압축적연구。무손수거압축채용병행적GZIP산법-mgz-ip，재획득흔호압축비적동시，가이대대축단압축시간。 mgzip시기우공향존저적병행압축책략，류수선병행여수거병행혼합적병행모식。연구표명，기우TILE64다핵평태상적병행수거압축위고능물리실험중적수거획취계통적수거압축수구제공료가행성방안。
It presented the study of parallel data compression on a TILE 64 developing board TLR26480 produced by Tilera Company .Used SMP -LINUX as operating system and Tilera Multicore Development Environment ( MDE) as software platform .Lossless data compression mgzip , the parallel algorithm of GIZP in this study , can not only get a good compression ratio , but also reduce the compression time greatly .Mgzip is based on the strategy of shared memory parallel compression , and the hybrid model of pipeline parallel and data parallel .Ac-cording to the study , the parallel data compression based on the Tilera platform can be a feasibility scheme for data acquisition system of high energy physics experiment .