[Issue 10550] Xorshift32 and Xorshift160 do not generate uniformly-distributed random numbers

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=10550


Joseph Rushton Wakeling <joseph.wakeling at webdrake.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |major


--- Comment #6 from Joseph Rushton Wakeling <joseph.wakeling at webdrake.net> 2013-07-08 02:46:40 PDT ---
(In reply to comment #5)
> Maybe this bug should have a "major" importance. And maybe a warning note in
> site ddocs should be added in the meantime.

Agree with the "major" importance, tweaked accordingly.

> These tests can help:
> http://en.wikipedia.org/wiki/Diehard_tests

These are now a little out of date, but there is the "dieharder" suite which is
actually available as a utility in many Linux distros.  I plan on incorporating
that into my test suite -- I think it should be as simple as just getting D to
generate random variates using whatever method and then piping them through to
dieharder.

> Some people should live longer. Probably some of his collaborators or people
> that have used his generators have some errata list or some suggestions to
> help.

I did not find anything yet, but I'll keep looking.  I assumed that his
homepage would still exist with remarks like this on it, but didn't track down
anything useful so far.

There are other people who've written follow-up papers on Xorshift who could be
worth contacting.  I may do that if I can't find obvious documentary material.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------