Glen wrote in message ... >"Alan Miller" <[EMAIL PROTECTED]> wrote in message news:<K1Fa8.25709$[EMAIL PROTECTED]>... >> The fastest way to generate random normals and exponentials is to use George >> Marsaglia's ziggurat algorithm. > >I've seen both ziggurat and Monty Python approaches claimed as being >"about the fastest" or "close to the fastest" among reasonably general >algorithms (not restricted to a single distribution), and they are >both nice and easy to understand and reasonably easy to code. > >But in the case of gaussian distributions, which is faster? > >I don't yet have the CACM article on the Monty Python for the gaussian >case, presumably it has some timing information. But maybe I don't >even need to look if the ziggurat approach is faster. I haven't seen >anything which directly discusses how they compare. > >Glen
First - the reference to George's paper on the ziggurat, and the code: The Journal of Statistical Software (2000) at: http://www.jstatsoft.org/v05/i08 Surprisingly, there is no reference to the Marsaglia & Tsang paper in TOMS in 1998, even though 7 of the 10 references are to Marsaglia & someone. If G & T have done speed comparisons, which you would expect, with the Monty Python method, they are not in the ziggurat paper, though there are other speed comparisons. -- Alan Miller (Honorary Research Fellow, CSIRO Mathematical & Information Sciences) http://www.ozemail.com.au/~milleraj http://users.bigpond.net.au/amiller/ ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =================================================================
