Linear Congruential Generators for Parallel Monte Carlo: the Leap-Frog Case.

DOI10.1515/MCMA.1998.4.1.1MaRDI QIDQ4392294zbMATH OpenOpenAlexFDO

Authors K. Entacher, Andreas Uhl, Stefan Wegenkittl

Publication date 18 February 1999

Published in Monte Carlo Methods and Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1515/mcma.1998.4.1.1

zbMATH Keywords

leap frog methodlinear congruential generatorsMonte Carlo integrationparallel computationparallel generationspectral test

Mathematics Subject Classification ID

Pseudo-random numbers; Monte Carlo methods (11K45) Monte Carlo methods (65C05) Random number generation in numerical analysis (65C10) Numerical quadrature and cubature formulas (65D32) Parallel numerical computation (65Y05)

Review text

Three linear congruential generators having good results of spectral test are evaluated. The generators are tested using the results of Monte Carlo integration as a test statistic. The assumption of randomness can not be rejected on 1\% level in the case, when the generators are used in the standard way. It is not true in the case when the generators are used to generate parallel streams of random numbers using the leap frog technique. In this case the values of test statistics even seem not to converge to the value of the integral for growing sample sizes. The practical importance of the finding follows from the fact that the leap frog technique is used to generate random sequences on parallel computers. The paper contains results of many tests in an instructive graphical form.

Recommendations

• Parallel use of multiplicative congruential random number generators
• Parallel streams of linear random numbers in the spectral test
• Parallel linear congruential generators with prime moduli
• Testing random number generators by numerical modeling of an exactly solvable problem
• Parallel realization of statistical simulation and random number generators

Cited in

(8)

• Long range correlations in linear congruential generators
• On the CRAY-System Random Number Generator
• Defects in parallel Monte Carlo and quasi-Monte Carlo integration using the leap-frog technique
• Efficient lattice assessment for LCG and GLP parameter searches
• Techniques for parallel quasi-Monte Carlo integration with digital sequences and associated problems
• Parallel use of multiplicative congruential random number generators
• Testing random number generators by numerical modeling of an exactly solvable problem
• A non-hybrid method for the PDF equations of turbulent flows on unstructured grids

This page was built for publication: Linear Congruential Generators for Parallel Monte Carlo: the Leap-Frog Case.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4392294)