[Issue 11598] std.random.uniform could be faster for integrals

[d-bugmail at puremagic.com]

https://d.puremagic.com/issues/show_bug.cgi?id=11598


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |joseph.wakeling at webdrake.ne
                   |                            |t


--- Comment #4 from Joseph Rushton Wakeling <joseph.wakeling at webdrake.net> 2013-12-01 04:20:47 PST ---
It should be simple enough to do a few different cases of rolling uniform(0, n)
a sufficient number of times (say, n * 10 million or n * 100 million) to
confirm that each rollable value has no systematic bias.

Do it for n = 2, 3, ..., 10, 20, 50, 100, and post the results here and in the
pull request.  This doesn't need to be part of formal unittests (too heavy) but
should be a documented part of your testing procedure.

There have been bugs in the past in std.random (I won't say which) which would
never have got there if the developer responsible had done a simple test like
this before submitting a pull request.

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