首页> 外文会议>ISCA International Conference on Parallel and Distributed Computing Systems >Memory Overhead for Directory-Based Cache Coherence Protocols in Multiprocessor Systems
【24h】

Memory Overhead for Directory-Based Cache Coherence Protocols in Multiprocessor Systems

机译:多处理器系统中基于目录的高速缓存同时协议协议的内存开销

获取原文

摘要

The problem of cache coherence in large-scale shared memory multiprocessor systems has been addressed using directory schemes. The memory overhead for various directory schemes expand as the number of processors becomes large, making the implementation cost unacceptable for some of these schemes. In this paper, the various directory schemes are compared on the basis of the memory overhead required for maintaining cache coherence in large-scale shared memory multiprocessor systems. The memory overhead for Singly Linked List is acceptable and grows indistinctly with the increase in number of processors.
机译:使用目录方案解决了大规模共享内存多处理系统中的缓存一致性问题。随着处理器的数量变大,各种目录方案的内存开销扩展,使得对于其中一些方案,实现成本不可接受。在本文中,基于在大规模共享内存多处理器系统中维护高速缓存相干的存储器开销的基础上进行各种目录方案。单链接列表的内存开销是可接受的,并且随着处理器数量的增加而毫无缩小。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号