ACORN: Difference between revisions

From SklogWiki
Jump to navigation Jump to search
No edit summary
(clarification: performance, references)
Line 3: Line 3:


Advantages of ACORN (from http://ACORN.wikramaratna.org/critique.html):
Advantages of ACORN (from http://ACORN.wikramaratna.org/critique.html):
<blockquote>
   <li>extremely light-weight code (a few lines) with reproducible results in any high-level language and on any platform; </li>
   <li>extremely light-weight code (a few lines) with reproducible results in any high-level language and on any platform; </li>
   <li>theoretical convergence is mathematically proven; </li>
  <li>computational and statistical performance comparable to the best currently available methods</li>
   <li>all current empirical test suites for PRNGs are passed (i.e TestU01 current version, in 2019); </li>
   <li>theoretical convergence is mathematically proven </li>
   <li>all current empirical test suites for PRNGs are passed (i.e TestU01 current version, in 2019); this is better than the widely-used Mersenne Twister which consistently fails on two of the tests</li>
   <li>can be easily extended to give sequences with longer period length, and improved statistical performance over higher dimensions and with higher precision.</li>
   <li>can be easily extended to give sequences with longer period length, and improved statistical performance over higher dimensions and with higher precision.</li>
 
</blockquote>


==Web site==  
==Web site==  

Revision as of 10:55, 5 July 2019

ACORN is the Additive Congruential Random Number generator [1] introduced by Roy Wikramaratna in 1989.


Advantages of ACORN (from http://ACORN.wikramaratna.org/critique.html):

  • extremely light-weight code (a few lines) with reproducible results in any high-level language and on any platform;
  • computational and statistical performance comparable to the best currently available methods
  • theoretical convergence is mathematically proven
  • all current empirical test suites for PRNGs are passed (i.e TestU01 current version, in 2019); this is better than the widely-used Mersenne Twister which consistently fails on two of the tests
  • can be easily extended to give sequences with longer period length, and improved statistical performance over higher dimensions and with higher precision.
  • Web site

    The web site ACORN.wikramaratna.org describes ACORN in detail and provides complete references.


    References

    see also ACORN.wikramaratna.org/references.html