遇见数据集

申威物理机CPU的整形性能数据

收藏
官方服务:

资源简介:

基于面向国产处理器的虚拟机构建、调度和管理技术,针对国产多核处理器进行虚拟执行环境构建,基于开源虚拟化软件提出高效能、低开销与可扩展虚拟机支撑技术;;支持申威CPU、内存虚拟化,虚拟化开销小于3%。该项通过SPECCPU 2006对申威物理机进行CPU性能测试,最终得到申威物理机的CPU的浮点型的相关负载的Base Rate数值,并与虚拟机进行对比,以达到虚拟化的CPU开销小于3%的指标。

Based on virtual machine construction, scheduling and management technologies targeting domestic processors, this work constructs virtual execution environments for domestic multi-core processors and proposes high-efficiency, low-overhead and scalable virtual machine support technologies based on open-source virtualization software. It supports Shenwei CPU and memory virtualization, with a virtualization overhead of less than 3%. This study conducts CPU performance testing on Shenwei-based physical machines using SPECCPU 2006, derives the Base Rate metrics for floating-point workloads of the Shenwei physical machine's CPU, and compares these metrics with those of the corresponding virtual machine to confirm that the virtualization-induced CPU overhead is less than 3%.

提供机构:
上海交通大学
搜集汇总
数据集介绍
申威物理机CPU的整形性能数据 数据集图片
背景与挑战
背景概述
该数据集基于面向国产处理器的虚拟化技术,通过SPECCPU 2006对申威物理机进行CPU性能测试,获取了浮点型负载的Base Rate数值。数据旨在对比物理机与虚拟机性能,以验证虚拟化开销小于3%的指标。
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务