Gaussian Random Number Generation
Authors:
Open Access
- 1 November 2016
- journal article
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
- Vol. 49 (3) , 1-37
Abstract
Some excellent surveys of the Gaussian random number generators (GRNGs) from the algorithmic perspective exist in the published literature to date (e.g., Thomas et al. [2007]). In the last decade, however, advancements in digital hardware have resulted in an ever-decreasing hardware cost and increased design flexibility. Additionally, recent advances in applications like gaming, weather forecasting, and simulations in physics and astronomy require faster, cheaper, and statistically accurate GRNGs. These two trends have contributed toward the development of a number of novel GRNG architectures optimized for hardware design. A detailed comparative study of these hardware architectures has been somewhat missing in the published literature. This work provides the potential user a capsulization of the published hardware GRNG architectures. We have provided the method and theory, pros and cons, and a comparative summary of the speed, statistical accuracy, and hardware resource utilization of these architectures. Finally, we have complemented this work by describing two novel hardware GRNG architectures, namely, the CLT-inversion and the multihat algorithm, respectively. These new architectures provide high tail accuracy (6σand8σ, respectively) at a low hardware cost.Keywords
This publication has 31 references indexed in Scilit:
- Some Comments on C. S. Wallace's Random Number GeneratorsThe Computer Journal, 2008
- Gaussian random number generatorsACM Computing Surveys, 2007
- The Quantum Separability Problem for Gaussian StatesElectronic Notes in Theoretical Computer Science, 2007
- Design and Evaluation of Hardware Pseudo-Random Number Generator MT19937IEICE Transactions on Information and Systems, 2005
- A Gaussian noise generator for hardware-based simulationsIEEE Transactions on Computers, 2004
- The Monty Python method for generating random variablesACM Transactions on Mathematical Software, 1998
- Fast pseudorandom generators for normal and exponential variatesACM Transactions on Mathematical Software, 1996
- A Fast, Easily Implemented Method for Sampling from Decreasing or Symmetric Unimodal Density FunctionsSIAM Journal on Scientific and Statistical Computing, 1984
- An Inverse Method for The Generation of Random Normal Deviates on Large-Scale ComputersMathematical Tables and Other Aids to Computation, 1958
- A Note on the Generation of Random Normal DeviatesThe Annals of Mathematical Statistics, 1958