Optimization of Digital Signal Transformation Functions in Multicluster VLIW DSP

doi:10.3969/j.issn.1000-3428.2016.03.017

Computer Engineering

Previous Articles Next Articles

Optimization of Digital Signal Transformation Functions in Multicluster VLIW DSP

ZHEN Yang ^a,b,c,GU Naijie ^a,b,c,YE Hong ^a,b,c

(a.School of Computer Science and Technology; b.Anhui Province Key Laboratory of Computing and Communication Software; c.Institute of Advanced Technology,University of Science and Technology of China,Hefei 230027,China)

Received:2015-03-12 Online:2016-03-15 Published:2016-03-15

数字信号变换函数在多簇VLIW DSP上的优化

甄扬^a,b,c,顾乃杰^a,b,c,叶鸿^a,b,c

(中国科学技术大学a.计算机科学与技术学院; b.安徽省计算与通信软件重点实验室; c.先进技术研究院,合肥 230027)

作者简介:甄扬(1991-),男,硕士研究生,主研方向为软件并行优化;顾乃杰(通讯作者),教授、博士生导师;叶鸿,博士研究生。
基金资助:
高等学校学科创新引智计划基金资助项目(B07033);安徽省自然科学基金资助项目“基于GPU集群的深度神经网络并行部署和优化策略研究”(1408085MKL06)。

Abstract

Abstract:

According to the characteristics of BWDSP100 processor’s architecture,this paper presents several practical ways to improve the performance of digital signal transformation functions in Digital Signal Processor(DSP) function library,including using special assembly instructions,instruction-level reordering,zero-overhead looping instruction,Instruction-level Parallelism(ILP),software vectorization and pipelining.It realizes parallel optimization version in library based on the original order version.Experimental results show that,in four-macro parallel mode,all digital signal transformation functions can achieve 9x speedup,90% functions can achieve 10x speedup,and 11.12x speedup is achieved on average.

Key words: Very Long Instruction Word(VLIW), Single Instruction Multiple Data(SIMD), Digital Signal Processor(DSP), loop unrolling, parallelization, multicluster

摘要：

针对BWDSP100体系结构特点,基于循环展开、指令调度以及软件流水等并行优化技术,结合多簇超长指令架构的特点,通过使用超算硬件指令、零开销循环、指令重新编排与并行等方法对BWDSP100数字信号处理函数库中的函数实施并行化,并基于库中原有的顺序版本实现并行优化版本。实验结果表明,在4宏并行化模式下,所有函数加速比达到9以上,90%的函数加速比超过10,平均加速比为11.12。

关键词: 超长指令字, 单指令流多数据流, 数字信号处理器, 循环展开, 并行化, 多簇

CLC Number:

TP391

ZHEN Yang,GU Naijie,YE Hong. Optimization of Digital Signal Transformation Functions in Multicluster VLIW DSP[J]. Computer Engineering, doi: 10.3969/j.issn.1000-3428.2016.03.017.

甄扬,顾乃杰,叶鸿. 数字信号变换函数在多簇VLIW DSP上的优化[J]. 计算机工程, doi: 10.3969/j.issn.1000-3428.2016.03.017.

/ / Recommend / Download Citations

URL: http://www.ecice06.com/EN/10.3969/j.issn.1000-3428.2016.03.017

http://www.ecice06.com/EN/Y2016/V42/I3/47

References

参考文献［1］中国电子科技集团公司第三十八研究所.BWDSP 100软件用户手册［Z］.2014. ［2］吴曼青.中国研发新型雷达系统打破西方技术垄断［EB/OL］.［20150211］.http://news.xinhuanet.com/18cpcnc/201211/13/c_113679376.htm. ［3］孟占红,赵保军.基于DSP的实时图像压缩软件优化技术研究［J］.电子学报,2006,34(9):15581561. ［4］李世军.JPEG图像压缩编码算法的DSP优化实现［J］.微计算机信息,2012,(9):193195. ［5］Kharin A,Vityazev S,Vityazev V,et al.Parallel FFT Implementation on TMS320c66x Multicore DSP［C］//Proceedings of the 6th European Embedded Design in Education and Research Conference.Washington D.C.,USA:IEEE Press,2014:4649. ［6］Qian Zhihong,Cao Lei,Su Weilian,et al.Recent Advances in Computer Science and Information Engineering［M］.Berlin,Germany:SpringerVerlag,2012. ［7］Sasanka R,Cook J J,Das A,et al.Analyzing Potential Benefits of Vectorization:USA,US20140258677_A1［P］.20140911. ［8］Turkington D A.Generalized Vectorization,Crossproducts,and Matrix Calculus［M］.Cambridge,UK:Cambridge University Press,2013. ［9］Yang Yangzhao,Gu Naijie,Zhao Zeng,et al.IPRAR:A DFGbased Approach to Instruction Clustering for Multicluster VLIW DSP Processor with SIMD Structure［J］.Journal of Computational Information Systems,2014,10(3):12571269. ［10］Cooley J W,Tukey J W.An Algorithm for the Machine Calculation of Complexes Fourier Series［J］.Mathematics of Computation,1965,19(19):297301. ［11］Lin T,Chen S,Kuo Yuting,et al.Design and Implementation of a Highperformance and Complexityeffective VLIW DSP for Multimedia Applications［J］.Journal of Signal Processing Systems,2007,51(3):209223. ［12］Rao K R,Yip P.Discrete Cosine Transform:Algorithms,Advantages,Applications［M］.［S.l.］:Academic Press,2014. ［13］葛珊,徐书文.TMS320C64x系列DSP的软件优化技术［J］.电视技术,2014,38(13):7071. ［14］Bryant R,O’Hallaron D R.Computer Systems:A Programmer’s Perspective［M］.［S.l.］:AddisonWesley,2010. ［15］顾乃杰,李凯,陈国良,等.基于龙芯2F体系结构的BLAS库优化［J］.中国科学技术大学学报,2008,38(7):854859. ［16］Yang Yangzhao,Gu Naijie,Ren Kaixin,et al.An Approach to Enhance Loop Performance for Multicluster VLIW DSP Processor［C］//Proceedings of the 27th International Conference on Architecture of Computing Systems.Luebeck,Germany:VDE VERLAG GMBH,2014:18. 编辑金胡考

[1]	LIN Lin, ZHU Aiqi, ZHAO Mingcan, ZHANG Shuai, YE Yanhao, XU Ji, HAN Lin, ZHAO Rongcai, HOU Chaofeng. GPU-Accelerated Algorithm Optimization for Molecular Dynamics Simulation of Crystalline Silicon [J]. Computer Engineering, 2023, 49(4): 166-173.
[2]	LI Bo, HUANG Dongqiang, JIA Jinfang, WU Li, WANG Xiaoying, HUANG Jianqiang. Research on Optimization of Heterogeneous Stencil Computing Based on CPU and GPU [J]. Computer Engineering, 2023, 49(4): 131-137.
[3]	FAN Mingliang, GUO Zihan, CHAI Xiaonan, SHANG Jiandong. Optimized Realization of Sobel Edge Detection Algorithm for FT-M7002 [J]. Computer Engineering, 2022, 48(6): 193-199.
[4]	DI Xinkai, YANG Haigang. FPGA-based Accelerator for Sparse Convolutional Neutral Network [J]. Computer Engineering, 2021, 47(7): 189-195,204.
[5]	GUO Yuluo, BIAN Haodong, DONG Runting, TANG Jiahao, WANG Xiaoying, HUANG Jianqiang. Parallel Fourier Space Image Similarity Calculation Based on SIMD [J]. Computer Engineering, 2021, 47(11): 247-253.
[6]	LU Jiajia, DU Yusong. Constant-time Implementation Method for Discrete Gaussian Sampling on Integer [J]. Computer Engineering, 2020, 46(8): 119-123.
[7]	TAO Fei, CHENG Keyang, ZHANG Jianming, TANG Yuhao. Pedestrian Reidentification Method Based on Posture and Parallel Attribute Learning [J]. Computer Engineering, 2020, 46(3): 246-253.
[8]	XUE Jianwei, JIANG Aimin. Computation and Control System for Principle Prototype of Optical Synthetic Aperture [J]. Computer Engineering, 2019, 45(8): 107-112.
[9]	HE Aixiang,GU Naijie,SU Junjie. Optimization Method of Basis Function Based on Multi-core ARM Architecture [J]. Computer Engineering, 2018, 44(5): 47-52,59.
[10]	GUO Baolong,WANG Jian,YAN Yunyi,NING Weikang. Optimal Design of DSP Protection Based on Multi-target PSO Algorithm [J]. Computer Engineering, 2018, 44(4): 74-80.
[11]	NING Ke,SUN Tongjing,XU Jiejie. Improved Nearest Neighbor Absorption First Clustering Algorithm for Massive Data [J]. Computer Engineering, 2018, 44(4): 35-40.
[12]	SUN Zongxin,ZHANG Guiyun. CPU Parallelization Eclat Algorithm Based on Bit Storage Tid [J]. Computer Engineering, 2018, 44(12): 79-84.
[13]	PENG Zhen,WU Baifeng. Collision Detection Based on Data Parallelism [J]. Computer Engineering, 2017, 43(9): 1-6.
[14]	HAN Lin,GAO Wei,WANG Dong,WANG Pengxiang,LI Yingying. A Single Instruction Multiple Data Vectorization Reduction Method [J]. Computer Engineering, 2017, 43(7): 9-14.
[15]	YU Zijian,MA De,YAN Xiaolang,SHEN Juncheng. FPGA-based Accelerator for Convolutional Neural Network [J]. Computer Engineering, 2017, 43(1): 109-114,119.

Please choose a citation manager

Content to export

Optimization of Digital Signal Transformation Functions in Multicluster VLIW DSP

数字信号变换函数在多簇VLIW DSP上的优化

PDF

Knowledge

Cited

Abstract

Cite this article

share this article

References

Related Articles 15

Recommended Articles

Metrics

Comments

模态框（Modal）标题

Please choose a citation manager

Content to export

Optimization of Digital Signal Transformation Functions in Multicluster VLIW DSP

数字信号变换函数在多簇VLIW DSP上的优化

PDF

Knowledge

Cited

Abstract

Cite this article

share this article

References

Related Articles 15

Recommended Articles

Metrics

Comments