Comparison and analysis of matrix multiplications on GPU and CPU

Computer Engineering and Applications ›› 2011, Vol. 47 ›› Issue (19): 9-11.

• 博士论坛 • Previous Articles Next Articles

Comparison and analysis of matrix multiplications on GPU and CPU

LIU Jinfeng，GUO Lei

School of Automation，Northwestern Polytechnical University，Xi’an 710129，China

Received:1900-01-01 Revised:1900-01-01 Online:2011-07-01 Published:2011-07-01

CPU与GPU上几种矩阵乘法的比较与分析

刘进锋，郭雷

西北工业大学自动化学院，西安 710129

Abstract

Abstract: Three matrix multiplications on CPU and four CUDA-based matrix multiplications on GPU are described，the causes of high performance are analyzed and the common characteristic of efficient algorithm is that data are properly organized and rationally utilized，and therefore the access cost effectively reduced and the speed is greatly improved.The best optimized implementation on CPU gain more 200 times fast than the common one，the best optimized implementation on GPU gain about 6 times fast than the best one on CPU.

Key words: matrix multiplication, Compute Unified Device Architecture（CUDA）, Graphic Processing Unit（GPU）, storage pattern

摘要： 描述了矩阵乘法在CPU上的三种实现方法和在GPU上基于CUDA架构的四种实现方法，分析了高性能方法的原由，发现它们的共同特点都是合理地组织数据并加以利用，这样能有效地减少存取开销，极大地提高算法的速度。其中CPU上的最优实现方法比普通算法快了200多倍，GPU上的最优实现方法又比CPU上的最优实现方法快了约6倍。

关键词: 矩阵乘法, 统一计算设备架构, 图形处理器, 存储模式

LIU Jinfeng，GUO Lei. Comparison and analysis of matrix multiplications on GPU and CPU[J]. Computer Engineering and Applications, 2011, 47(19): 9-11.

刘进锋，郭雷. CPU与GPU上几种矩阵乘法的比较与分析[J]. 计算机工程与应用, 2011, 47(19): 9-11.

[1]	LIU Gaoming1，2, RONG Kui2, ZHU Hui1, TAN Huailiang2. Design of graphics display system for embedded RTOS [J]. Computer Engineering and Applications, 2016, 52(9): 190-195.
[2]	WEI Hongchang, ZHU Zhengdong, DONG Xiaoshe, NING Jie. Asymptotic fitting optimization technology for source-to-source compile system on CPU-GPU architecture [J]. Computer Engineering and Applications, 2016, 52(21): 30-35.
[3]	LI Zhengfu1，2, WANG Xicheng3, LI Keqiu1, YAO Xiang3, DONG Yueli2. Information entropy multi-population genetic algorithm based on CUDA [J]. Computer Engineering and Applications, 2016, 52(1): 12-16.
[4]	ZHONG Jiyuan, MEI Kuizhi, WEN Zhexi. Parallel stream computing implementation of GIST algorithm on heterogeneous platform [J]. Computer Engineering and Applications, 2015, 51(6): 139-144.
[5]	DONG Lili, DONG Wei, ZHANG Xiang. Research for memory data clustering efficiency with CUDA [J]. Computer Engineering and Applications, 2015, 51(22): 243-251.
[6]	LIN Min, ZHONG Yiwen. Three GPU-based parallel simulated annealing algorithm with adaptive neighborhood [J]. Computer Engineering and Applications, 2015, 51(22): 70-76.
[7]	ZHU Zhengdong1, LIU Yuan1, ＷEI Hongchang1, YAN Kang1, WANG Yinfeng2, DONG Xiaoshe1. Novel automatic mapping technology on CPU-GPU heterogeneous systems [J]. Computer Engineering and Applications, 2015, 51(21): 41-47.
[8]	WANG Hong1，2, WANG Peng1，2. Improved AC pattern matching algorithm based on GPU [J]. Computer Engineering and Applications, 2015, 51(18): 7-12.
[9]	ZENG Bo1，2, LEI Youcheng1, WANG Congzhi2, QIU Weibao2, FENG Ge2, ZENG Chengzhi2, YANG Ge2, ZHENG Hairong2. CUDA-based acoustic radiation force imaging algorithm [J]. Computer Engineering and Applications, 2015, 51(18): 249-254.
[10]	XU Liang1, TIAN Zheng2, WANG Zhen2. Parallel stereo matching using variable window based disparity refinement [J]. Computer Engineering and Applications, 2015, 51(15): 193-197.
[11]	ZHOU Bingyuan1, CHEN Qingkui1，2, GAO Liping1, QIN Chuan1，2. Image matching algorithm based on CUDA [J]. Computer Engineering and Applications, 2015, 51(12): 165-170.
[12]	ZHAO Jiawei, FANG Jiuling, SU Ming. Fast implementation of Legendre sequence generation by CUDA [J]. Computer Engineering and Applications, 2014, 50(8): 66-71.
[13]	CHEN Hua, SHI Yuerong. Study on restarted PGMRES parallel algorithm with GPU [J]. Computer Engineering and Applications, 2014, 50(7): 35-40.
[14]	CAI Yunlong, XIAO Sumei, QI Long. Locking paralleled GPU-based method research for unstructured mesh generation [J]. Computer Engineering and Applications, 2014, 50(6): 56-60.
[15]	TANG Shaohua. Parallelizing network coding on manycore GPU-accelerated system with optimization [J]. Computer Engineering and Applications, 2014, 50(21): 79-84.

Comparison and analysis of matrix multiplications on GPU and CPU

CPU与GPU上几种矩阵乘法的比较与分析

PDF

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 15

Recommended Articles

Metrics