【24h】

Storage Class Memory

机译:存储类内存

获取原文
       

摘要

Advances in processor technology have created myriads of related problems for system designers, chief amongst which is the need to improve on the latency between storage, memory and the processor itself. Significant data transfer rates between the processor, L1 cache, L2 cache and main memory are possible but it is at this point that the scenario changes and it is this problem which must be addressed in the years to come if the gains made in processor technology are to be further translated into overall system performance improvements. Storage Class Memory (SCM) offers just the bridge needed to get over the ever widening gap between CPU processing speeds, the need to move large blocks of data quickly and the read/write speeds offered by traditional disk reliant systems. In terms of their predicted cost, speed, retentive abilities(data persistence) and power requirements, the technologies that can be classed as SCM are the most likely candidates to keep Moore’s Law true well into the next decade and beyond. This paper provides an overview of SCM and its role in the ‘greening’ of data centres around the world. The makeup of data centres with their vast banks of HDD’s all-consuming power whether idle or in use is not a sustainable model as the world faces into a future of increasing energy costs and decreasing supply of traditional fuel sources.
机译:处理器技术的进步为系统设计人员带来了无数相关问题,其中最主要的是需要改善存储,内存和处理器本身之间的延迟。处理器,L1高速缓存,L2高速缓存和主存储器之间的数据传输速率可能很高,但是此时情况发生了变化,如果处理器技术取得了丰硕的成果,则是今后几年必须解决的问题。进一步转化为整体系统性能的提高。存储类内存(SCM)只是克服CPU处理速度,快速移动大数据块以及传统磁盘相关系统提供的读/写速度之间不断扩大的差距所需的桥梁。就其预计成本,速度,保持能力(数据持久性)和电源要求而言,可以归类为SCM的技术最有可能在未来十年乃至更长时间内保持摩尔定律的正确执行。本文概述了SCM及其在全球数据中心“绿色化”中的作用。随着世界面临着能源成本增加和传统燃料供应减少的未来,数据中心以其大量的闲置或使用中的HDD所有消耗功率组成的数据中心不是可持续的模型。

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