"The generation of random numbers is too important to be left to chance." Robert R. Coveyou
An alternate random number generator. You can import Random.Pcg as Random and everything will continue to
work (except generate for use with Cmds, andelm-random-extra).
This library offers two main improvements over core, without a large loss of performance (see test/benchmark):
Better statistical properties. If you use any seed less than 53668 and generate one bool, it will be True – if
you're using the core library. This library produces far less predictable output, especially if you use thousands of
random numbers. The core library fails after as little as five seconds of scrutiny. See test/dieharder for more
details.
More features. This library exports constant and andMap, which are conspicuously absent from core, along with
other helpful functions for composing generators. Especially interesting is the independentSeed function which allows
for lazy lists and isolated components to generate as much randomness as they need, when they need it. There is also a
fastForward function that allows random access (rather than sequential access) into, effectively, an infinite array of
random numbers. This can be used instead of reading a finite amount of random values into a large data structure.
This is an implementation of PCG by M. E. O'Neil, based on the JS port by Thom Chiovoloni (MIT license). The generator is not cryptographically secure.
Please report bugs, feature requests, and other issues on GitHub.
generate renamed step to match core 4.0.0 API.Random.Pcg from Random.PCG.split has been removed; use independentSeed.minInt and maxInt values changed to match core.