1、 萬方數(shù)據(jù) 萬方數(shù)據(jù) 萬方數(shù)據(jù) 萬方數(shù)據(jù) 萬方數(shù)據(jù) 萬方數(shù)據(jù) 萬方數(shù)據(jù)172計算機學報2011缶6總結(jié)本文首先應用矩陣的形式分析了CombinedTausworthe隨機數(shù)生成器中的核心部件狀態(tài)轉(zhuǎn)換邏輯A5的特征,提出了一種只需要使用s個二輸人異或門來實現(xiàn)A的硬件結(jié)構(gòu).而后,給出了Combined Tausworthe隨機數(shù)生成器中周期和輸出位寬的配置方法和配置原則.在Box Muller算法的硬件實現(xiàn)部分,本文給出了應用“最值分析法”和“靜態(tài)誤差分析法”確定各操作數(shù)位寬的具體分析過程,并計算出了均勻分布隨機數(shù)的位寬L與高斯分布隨機數(shù)輸出范圍的關(guān)系.在Xilinx Vertex6FPGA上的硬

2、件實現(xiàn)結(jié)果表明:應用本文提出的轉(zhuǎn)換邏輯A所設(shè)計實現(xiàn)的Combined Tausworthe隨機數(shù)生成器可以在消耗較少硬件資源的前提下獲得較高的工作速度.通過與其它幾種高斯隨機數(shù)生成器的硬件實現(xiàn)結(jié)果進行比對,本文設(shè)計的高斯隨機數(shù)生成器可以實現(xiàn)更高的工作速度和獲得更優(yōu)的硬件資源使用效率.本文的硬件設(shè)計所產(chǎn)生的高斯隨機數(shù)可以通過X2和KS測試方法的檢驗.1 23參考文獻Jason C et a1.Accelerating Monte Carlo based SSTA using FPGA/Proceedings of the18th Annual ACM/SIGDA International Sy

3、mposium on Field Programmable Gate Arrays. Monterey。California,USA.2010I111114Luu J et a1.FPGAbased Monte Carlo computation of light absorption for photodynamic cancer therapy/Proceedings ofGU Xiao-Chen。bom in1980,Ph.13.His research interests include mixed signal IC design。RFIC design and reconfigur

4、able computing.BackgroundThis paper is supported by NationaI High Technology Research and Development Program(863Programof China (2009AA012124。2009AA0lZl04,2009AA0lZl02.The research work of this paper is part of the project that is being carried on by Institute for Integrated Micro andthe2009l7th IE

5、EE Symposium on Field Programmable Custom Computing Machines.Napa,CA,2009:1571643Thomas D B,Howes L。Luk W.A comparison of CPUs,GPUs。FPGAs,and massively parallel processor arrays forrandom number generation/Proceedings of the ACM/SIGDA International Symposium on Field ProgrammableGate Arrays.Monterey

6、。California,USA.2009:6372 4Cheung R C C et a1.Hardware generation of arbitrary random number distributions from uniform distributions via theinversion method.IEEE Transactions on Very Large ScaleIntegration(VLSlSystems,2007,15(8:9529625Lee D U et a1.A Hardware Gaussian noise generator usingthe box-m

7、uUer method and its error analysis.IEEE Transac-tions on Computers,2006,55(6l65967163Pierre L E.Maximally equidistributed combined tausworthe.Mathematics of Computation。1996,65(213:20321373Box G E P,Muller M E.A note on the generation of random normaldeviates.Annals of Mathematical Statistics,1958。2

8、9(2:610一61183Gu X.Zhang M.Uniform random number generator using leap-ahead LFSR architecture/Proceedings of the International Conference on Computer and Communications Security(ICCCS2009.Hong Kong。China,2009l150一1549Dalai I L。Stefan D.A hardware framework for the fast gener-afion of multiple longper

9、iod random number streams/Pro。ceedings ofthe16th International ACM/SIGDA Symposium on Field Programmable Gate Arrays(FPGA2008.Monterey.California,USA,2008:245254IOPierre L E,Richard S.TestU01l A C library for empirical testing of random number generators.ACM Transactions onMathematical Software(TOMS

10、。2007,33(4:Atrtical22 11Agostino R B D.Stephens M A.Goodness-of-Fit Techniques.New York:Marcel Dekker Inc.1986ZHANG Min-Xuan。born in1954,professor。Ph.D. supervisor.His current research interests include high performance computer architecture and micro-electronics.Nano Systems(IMNSof Edinburgh Univer

11、sity.The aim of the project is to explore the performance of FPGA based accelerators which have Monte Carlo based applications running on them.MonteCarlo method is computation intensive and could be well accelerated on FPGAs.There is a FPGA based 萬方數(shù)據(jù) 萬方數(shù)據(jù) 一種基于FPGA的高斯隨機數(shù)生成器的設(shè)計與實現(xiàn)作者:谷曉忱, 張民選, GU Xiao

12、-Chen, ZHANG Min-Xuan作者單位:國防科學技術(shù)大學計算機學院PDL重點實驗室,長沙,410073刊名: 計算機學報英文刊名:CHINESE JOURNAL OF COMPUTERS年,卷(期:2011,34(1參考文獻(11條1.Agostino R B D;Stephens M A Goodness-of-Fit Techniques 19862.Pierre L E;Richard S TestU01:A C library for empirical testing of random number generators2007(043.Dalal I L;Stefan

13、 D A hardware framework for the fast generation of multiple long-period random number streams 20084.Gu X;Zhang M Uniform random number generator using leap-ahead LFSR architecture 20095.Box G E P;Muller M E A note on the generation of random normal deviates 1958(026.Pierre L E Maximally equidistribu

14、ted combined tausworthe外文期刊 1996(2137.Lee D U A Hardware Gaussian noise generator using the box-muller method and its error analysis外文期刊 2006(068.Cheung R C C Hardware generation of arbitrary random number distributions from uniformdistributions via the inversion method外文期刊 2007(089.Thomas D B;Howes L;Luk W A comparison of CPUs,GPUs,FPGAs,and massivel