一种可重构多核系统的仿真加速方案

牛云鹏; 张多利; 唐旭; 宋宇鲲

Chinese

您当前的位置：

首页 >

文章列表页 >

一种可重构多核系统的仿真加速方案

更新时间：2026-09-04

- 一种可重构多核系统的仿真加速方案
- An acceleration scheme for reconfigurable multi-core system
- 微电子学与计算机 2021年38卷第2期页码：19-24
- 作者机构：
  
  1. 合肥工业大学微电子设计研究所,安徽,合肥,230601
  2. 教育部IC设计网上合作研究中心,安徽,合肥,230601
- 作者简介：
- 基金信息：
  
  国家自然科学基金 61874156国家重点研发计划 2018YFB2202604
- DOI：
  中图分类号：
- 纸质出版：2021
- 稿件说明：
移动端阅览
牛云鹏, 张多利, 唐旭, 宋宇鲲. 一种可重构多核系统的仿真加速方案[J]. 微电子学与计算机, 2021, 38(2): 19-24.

NIU Yun-peng, ZHANG Duo-li, TANG Xu, et al. An acceleration scheme for reconfigurable multi-core system[J]. 2021, 38(2): 19-24.
牛云鹏, 张多利, 唐旭, 宋宇鲲. 一种可重构多核系统的仿真加速方案[J]. 微电子学与计算机, 2021, 38(2): 19-24. DOI：

NIU Yun-peng, ZHANG Duo-li, TANG Xu, et al. An acceleration scheme for reconfigurable multi-core system[J]. 2021, 38(2): 19-24. DOI：

摘要

针对可重构多核系统多任务仿真耗时长、并发性差的问题，提供了一种分层的、以资源请求异步响应为特点的松耦合多精度多线程系统建模方案.采用此方案的系统模型在任务仿真时，不仅能够提供寄存器级的实时数据反馈，同时也具有优异的任务并发性能.实验显示，相对于传统上位机软件数据收集和计算，该模拟器周期预测相对误差小于3%，模拟计算相对误差小于2e-5，仿真耗时最大缩短73.52%；在启用任务异步响应优化后，任务并行度平均提升28.21%.该设计在硬件目标系统的快速性能评估方面具有良好的实用价值.

Abstract

This paper provides a layered

loosely coupled

multi-precision

multi-threaded system modeling for solving the problem of time-consuming of multi-task simulation and poor concurrency in reconfigurable multi-core system. The system model using this scheme can not only provide register-level real-time data feedback during task simulation

but also have excellent task concurrency performance. Experiments show that

compared with the data collection and calculation of traditional host computer software

the relative error of the simulator′s period prediction is less than 3%

the relative error of simulation calculation is less than 2e-5

and the simulation time is shortened by a maximum of 73.52%; after enabling the task asynchronous response optimization

Task parallelism increased by an average of 28.21%. The design has good practical value in the rapid performance evaluation of the hardware target system.

关键词

Keywords

references

NOLTING S, PAYÁ-VAYÁ G, GIESEMANN F, et al. Dynamic self-reconfiguration of a MIPS-based soft-core processor architecture[J]. Journal of Parallel and Distributed Computing, 2019, 133: 391-406. DOI: 10.1016/j.jpdc.2017.09.013.

WEI Y. Construct multi-coreDSP software simulation platform based on SystemC[J]. Microelectronics% Computer, 2016, 33(4): 86-90. DOI: 10.19304/j.cnki.issn1000-7180.2016.04.019.

HE W Q, YANG L, LU Q. Cycle-accurate DSP processor modelingwithSystemC[J]. Microelectronics % Computer, 2013, 30(4): 107-110. DOI: 10.19304/j.cnki.issn1000-7180.2013.04.026.

KIRCHSTEIGER C M, SCHWEITZER H, TRUMMER C, et al. A software performance simulation methodology for rapid system architecture exploration[C]//Proceedings ofthe 15th IEEE International Conference on Electronics, Circuits and Systems. St. Julien's, Malta: IEEE, 2008: 494-497. DOI: 10.1109/ICECS.2008.4674898.

CHEN H, LUO W L, GAN Y Y. Multi_MINT: A multiprocessor simulator basedonMINT[J]. ComputerEngineering% Science, 2011, 33(12): 37-43. DOI: 10.3969/j.issn.1007-130X.2011.12.007.

YU J, LI X, ZHOU X H, et al. Study on retargetable cycle-accurate simulator generation environment[J]. Journal of Chinese Computer Systems, 2007, 28(1): 166-171. DOI: 10.3969/j.issn.1000-1220.2007.01.036.

EGGER B, SONG E, LEE H, et al. Random test program generation for verification and validation of the Samsung Reconfigurable Processor[J]. Journal of Systems Architecture, 2019, 97: 219-238. DOI: 10.1016/j.sysarc.2019.05.007.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

一种基于特征上下文理解与切片分析的肺结核检测算法

基于区块链的AI生成内容溯源框架设计研究

面向处理器分支预测单元的Spectre攻击轻量化防御方法

V型沟槽SiC中子探测器的蒙特卡洛分析

AlN/GaN混合缓冲层双异质结p-GaN HEMT直流特性仿真