Pseudo-random numbers in [0, n), covering all numbers in n steps?

Joseph Rushton Wakeling via Digitalmars-d digitalmars-d at puremagic.com
Fri Feb 26 00:18:01 PST 2016


On Friday, 26 February 2016 at 08:12:18 UTC, Joseph Rushton 
Wakeling wrote:
> On Friday, 26 February 2016 at 08:05:09 UTC, Joseph Rushton 
> Wakeling wrote:
>> On Thursday, 25 February 2016 at 17:27:25 UTC, Andrei 
>> Alexandrescu wrote:
>>> So we have 
>>> https://dlang.org/phobos/std_random.html#.randomCover which 
>>> needs to awkwardly allocate memory to keep track of the 
>>> portions of the array already covered.
>>
>> Yes, this is definitely a standout in terms of being an 
>> unpleasant solution.  It means that you require o(N) memory 
>> even when you're dealing with a lazily-evaluated range -- it 
>> would probably be more efficient in practice to just write the 
>> input into an array and do an in-place shuffle. :-(

Also: I don't have the access to determine if this is really on 
the money, but looks like it could be promising: 
http://www.sciencedirect.com/science/article/pii/S0020019098001276


More information about the Digitalmars-d mailing list