期刊文献+

基于混合精度算法的改进HPL软件包 被引量:2

Improved HPL Software Package Based on Mixed Precision Algorithm
下载PDF
导出
摘要 利用求解线性方程组的混合精度算法,对HPL软件包进行改进。从性能与加速比、迭代时间与迭代次数以及误差分析3个方面,在四路AMD Opteron870双核处理器平台上,对原HPL与改进的HPL软件包进行对比测试。实验结果表明,改进的HPL软件包在保证双精度浮点精度要求的前提下,计算性能大约提高1倍,并具有良好的可扩展性。 This paper improves a High Performance Linpack(HPL) software package by using mixed precision algorithm to solve linear equations set.The performance test which including performance and speedup,the number and time of iteration and error analysis for original HPL and improved HPL software package are conducted on the platform of four AMD Opteron870 dual-core processors.Experimental results show the computing performance of improved software package enhances almost twice compared with the original HPL while keeping double floating point precision and it also has good scalability.
出处 《计算机工程》 CAS CSCD 北大核心 2010年第19期47-49,共3页 Computer Engineering
基金 国家自然科学基金资助项目(60303020) 国家自然科学基金资助重点项目(60533020) 国家"863"计划基金资助项目(2006AA01A102 2006AA01A125)
关键词 混合精度算法 HPL软件包 加速比 mixed precision algorithm High Performance Linpack(HPL) software package speedup ratio
  • 相关文献

参考文献7

  • 1Petitet A, Whaley R C, Dongarra J. HPL----A Portable Implementation of the High-performance Linpack Benchmark for Distributed-memory Computers[EB/OL]. (2008-09-10). http://www. netlib.org/benchmark/hpl/.
  • 2Baboulin M, Butlari A, Dongarra J, et al. Accerlating Scientific Computations with Mixed Precision Algorithms[J]. Computer Physics Communications, 2009, 180(12): 2526-2533.
  • 3Demmel J W. Applied Numerical Linear Algebra[M]. 1st ed. [S. l.]: SIAM, 1997.
  • 4Wilkinson J H. Rounding Errors in Algebraic Processes[M]. [S. l.]: Prentice Hall Press, 1963.
  • 5Dongarra J, Luszczek P. How Elegant Code Evolves with Hardware: the Case of Gaussian Elimination[M]. [S. 1.]: O'Reilly & Associates Inc., 2007.
  • 6Langou J, Luszczek P, Kurzak J. Exploiting the Performance of 32 bit Floating Point Arithmetic in Obtaining 64 bit Accuracy[C]// Proc. of 2006 ACM/IEEE Conference on Supercomputing. Tampa, Florida, USA: [s. n.], 2006.
  • 7Arioli M, Duff I S, Using FGMRES to Obtain Backward Stability in Mixed Precision[R]. [S. 1.]: Rutherford Appleton Laboratory. Technical Report: RAL-TR-2008-006, 2008.

同被引文献12

  • 1王晓英,都志辉.基于HPL测试的集群系统性能分析与优化[J].计算机科学,2005,32(11):231-234. 被引量:10
  • 2张文力,陈明宇,樊建平.HPL测试性能仿真与预测[J].计算机研究与发展,2006,43(3):557-562. 被引量:13
  • 3王俊,文延华,漆锋滨.计算机浮点功能测试方法[J].计算机应用与软件,2006,23(6):68-70. 被引量:3
  • 4白中英.计算机组成原理[M].4版.北京:科学出版社,2008.
  • 5IEEE Computer Society. IEEE standard for floating-point arithmetic [ S/OL]. [ 2012-10-20]. http://ieeexplore, ieee. org/servlet/opac? punumber = 4610933.
  • 6BABOULIN M, BUTrARI A, DONGARRA J, et al. Accelerating scientific computations with mixed precision algorithms[ J]. Comput- er Physics Communications, 2009, 180(12) : 2526 -2533.
  • 7DEMMEL J, HIDA Y. Accurate and eflqcient floating point summation[ J]. SIAM Journal on Scientific Computing, 2003, 25(4) : 1214 - 1248.
  • 8李怀刚.计算机组成原理[M].北京:机械工业出版社,2012.
  • 9STALLINGSW.计算机组织与体系结构[M].7版.张昆藏,译.北京:清华大学出版社,2006.
  • 10BOLDO S, MELQUIOND G. Emulation of a FMA and correctly- rounded sums: proved algorithms using rounding to odd[ J]. IEEE Transactions on Computers, 2008, 57(4):462-471.

引证文献2

二级引证文献8

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部