一种基于Kepler架构GPU的通信仿真加速方法

doi:10.11959/j.issn.1000-0801.2015248

电信科学 ›› 2015, Vol. 31 ›› Issue (10): 82-88.doi: 10.11959/j.issn.1000-0801.2015248

一种基于Kepler架构GPU的通信仿真加速方法

韩秉君¹,黄诗铭²,杜滢¹

¹ 中国信息通信研究院北京 100191
² 北京邮电大学北京 100876

出版日期:2015-10-20 发布日期:2017-07-21
基金资助:
国家科技重大专项基金资助项目

A Simulation Accelerating Method Based on CUDA with Kepler GPU

Bingjun Han¹,Shiming Huang²,Ying Du¹

¹ China Academy of Information and Communication Technology，Beijing 100191，China
² Beijing University of Posts and Telecommunications，Beijing 100876，China

Online:2015-10-20 Published:2017-07-21
Supported by:
The National Science and Technology Major Project

摘要/Abstract

摘要：

提出了一种在 Kepler 架构 GPU（graphics processing unit，图形处理器）上利用 CUDA（compute unified device architecture，统一计算设备架构）技术加速通信仿真中DFT（discrete Fourier transform，离散傅里叶变换）处理过程的方法。该方法的核心思想是利用线程级并行技术实现单条收发链路内部DFT运算的并行加速，并利用动态并行和Hyper-Q技术实现不同收发用户对之间链路处理过程的并行加速，从而最终达到加速仿真中DFT处理过程的目的。实验结果表明，相对单核单线程CPU程序和上一代Fermi架构GPU程序，该方法分别能够将DFT处理速度提升300倍和3倍，具有较好的加速效果。

关键词: CUDA, 仿真平台DFT, 加速方法, 并行处理

Abstract:

An accelerating method based on CUDA（compute unified device architecture）with Kepler GPU（graphics processing unit）was proposed to speed up the DFT（discrete Fourier transform）processing in the communication simulation platform.Based on this method，the whole DFT processing was split into subtasks named molecular-subtasks corresponding to communication links and a molecular-subtask was further split into smaller parallel subtasks named atomic-subtasks which correspond to the DFT processing in a link.Then，the atomic-subtasks were processed in parallel by the threads in a GPU kernel function，as well as the molecular-subtasks were processed in parallel via several GPU kernel functions to shorter the simulation time.Simulation results show this method can speed up the DFT processing more than 300 times compared with single thread CPU program and 3 times compared with traditional GPU program.

Key words: compute unified device architecture, simulation platform DFT, accelerating method, processed in parallel

韩秉君,黄诗铭,杜滢. 一种基于Kepler架构GPU的通信仿真加速方法[J]. 电信科学, 2015, 31(10): 82-88.

Bingjun Han,Shiming Huang,Ying Du. A Simulation Accelerating Method Based on CUDA with Kepler GPU[J]. Telecommunications Science, 2015, 31(10): 82-88.

图/表 11

图1

图2

图3

图4

图5

图6

表1

表2

图7

图8

图9

参考文献 12

1	NVIDIA Corporation CUDA toolkit documentation v7.5. ， 2015
2	NVIDIA Corporation Nvidia kepler GK110 next-generation CUDA compute architecture. ， 2012
3	Abdelrazek A F ， Kaschub M ， Blankenhorn C ， et al. A novel architecture using NVIDIA CUDA to speed up simulation of multi-path fast fading channels. Proceedings of the 69th IEEE Vehicular Technology Conference， Barcelona，Spain， 2009
4	Laguna-Sanchez G A ， Prieto-Guerrero A ， Rodriguez-Colina E . Speedup simulation for OFDM over PLC channel using a multithreading GPU. Proceedings of IEEE Latin-American Conference on Communications （LATINCOM）， Belem，Brazil， 2011
5	Potluri S ， Wang H ， Bureddy D ， et al. Optimizing MPI communication on multi-GPU systems using CUDA inter-process communication. Proceedings of the 26th IEEE International on Parallel and Distributed Processing Symposium Workshops ＆ phD Forum（IPDPSW）， Shanghai，China， 2012： 1848～1857
6	Wu J ， JaJa J ， Balaras E . An optimized FFT-based direct Poisson solver on CUDA GPUs. IEEE Transactions on Parallel and Distributed Systems， 2014（1）： 550～559
7	Beermann M ， Monro E ， Schmalen H ， et al. High speed decoding of non-binary irregular LDPC codes using GPUs. Proceedings of IEEE Workshop on Signal Processing System （SiPS）， Taipei，China， 2013
8	Rodriguez A ， Valverde J ， Torre E ， et al. Dynamic management of multikernel multithread accelerators using dynamic partial reconfiguration. Proceedings of the 9th International Symposium on Reconfigurable and Communication-Cenric Systems-on-Chip （ReCoSoC）， Montpellier，France， 2014
9	Proakis J G . Digital Signal Processing， 4th Revised Edition London：Pearson Prentice Hall， 2009： 105～129
10	Noga A ， Topa T . Kernel execution strategies for GPU-accelerated version of method of moments. Proceedings of the 20th International Conference on Microwaves，Radar，and Wireless Communication（MIKON）， Gdansk，Poland， 2014
11	Wilt N ， The CUDA Handbook. Upper Saddle River：Addison-Wesley， 2013
12	Bilel B R ， Navid N . Cunetsim：a GPU based simulation testbed for large scale mobile networks. Proceedings of International Conference on Communications and Information Technology （ICCIT）， Hammamet，Tunisia， 2012

条目	取值
系统拓扑	57扇区/570用户
发送天线数/根	4
接收天线数/根	2
最小频域采样点/个	50（50 RB，1个/RB）
最大频域采样点/个	1 200（100 RB，12个/RB）
时域采样点	20

设备型号	Intel Q9500	Tesla C2050	Tesla K20c
硬件架构	Core2	Fermi	Kepler
运行时间/s	32	0.31	0.10

一种基于Kepler架构GPU的通信仿真加速方法

A Simulation Accelerating Method Based on CUDA with Kepler GPU

在线阅读

PDF下载

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 12

相关文章 1

Metrics

推荐阅读 0