RandomGen

Random Number Generator using settable Basic RNG interface for future NumPy RandomState evolution.

This is a library and generic interface for alternative random generators in Python and NumPy.

Compatibility Warning

RandomGenerator does not support Box-Muller normal variates and so it not 100% compatible with NumPy (or randomstate). Box-Muller normals are slow to generate and all functions which previously relied on Box-Muller normals now use the faster Ziggurat implementation. If you require backward compatibility, a legacy generator, LegacyGenerator, has been created which can fully reproduce the sequence produced by NumPy.

Features

• Replacement for NumPy's RandomState

  from randomgen import RandomGenerator, MT19937
  rnd = RandomGenerator(MT19937())
  x = rnd.standard_normal(100)
  y = rnd.random_sample(100)
  z = rnd.randn(10,10)

• Default random generator is a fast generator called Xoroshiro128plus

• Support for random number generators that support independent streams and jumping ahead so that sub-streams can be generated

• Faster random number generation, especially for normal, standard exponential and standard gamma using the Ziggurat method

  from randomgen import RandomGenerator
  # Default basic PRNG is Xoroshiro128
  rnd = RandomGenerator()
  w = rnd.standard_normal(10000, method='zig')
  x = rnd.standard_exponential(10000, method='zig')
  y = rnd.standard_gamma(5.5, 10000, method='zig')

• Support for 32-bit floating randoms for core generators. Currently supported:

  • Uniforms (random_sample)
  • Exponentials (standard_exponential, both Inverse CDF and Ziggurat)
  • Normals (standard_normal)
  • Standard Gammas (via standard_gamma)

  WARNING: The 32-bit generators are experimental and subject to change.

  Note: There are no plans to extend the alternative precision generation to all distributions.

• Support for filling existing arrays using out keyword argument. Currently supported in (both 32- and 64-bit outputs)

  • Uniforms (random_sample)
  • Exponentials (standard_exponential)
  • Normals (standard_normal)
  • Standard Gammas (via standard_gamma)

• Support for Lemire's method of generating uniform integers on an arbitrary interval by setting use_masked=True.

Included Pseudo Random Number Generators

This module includes a number of alternative random number generators in addition to the MT19937 that is included in NumPy. The RNGs include:

• MT19937, the NumPy rng
• dSFMT a SSE2-aware version of the MT19937 generator that is especially fast at generating doubles
• xoroshiro128+, xorshift1024*φ, xoshiro256**, and xoshiro512**
• PCG64
• ThreeFry and Philox from Random123

Differences from numpy.random.RandomState

New Features

• standard_normal, normal, randn and multivariate_normal all use the much faster (100%+) Ziggurat method.
• standard_gamma and gamma both use the much faster Ziggurat method.
• standard_exponential exponential both support an additional method keyword argument which can be inv or zig where inv corresponds to the current method using the inverse CDF and zig uses the much faster (100%+) Ziggurat method.
• Core random number generators can produce either single precision (np.float32) or double precision (np.float64, the default) using the optional keyword argument dtype
• Core random number generators can fill existing arrays using the out keyword argument
• Standardizes integer-values random values as int64 for all platforms.
• randint supports generating using rejection sampling on masked values (the default) or Lemire's method. Lemire's method can be much faster when the required interval length is much smaller than the closes power of 2.

New Functions

• random_entropy - Read from the system entropy provider, which is commonly used in cryptographic applications
• random_raw - Direct access to the values produced by the underlying PRNG. The range of the values returned depends on the specifics of the PRNG implementation.
• random_uintegers - unsigned integers, either 32- ([0, 2**32-1]) or 64-bit ([0, 2**64-1])
• jump - Jumps RNGs that support it. jump moves the state a great distance. Only available if supported by the RNG.
• advance - Advanced the RNG 'as-if' a number of draws were made, without actually drawing the numbers. Only available if supported by the RNG.

Status

• Builds and passes all tests on:
  • Linux 32/64 bit, Python 2.7, 3.4, 3.5, 3.6
  • PC-BSD (FreeBSD) 64-bit, Python 2.7
  • OSX 64-bit, Python 3.6
  • Windows 32/64 bit, Python 2.7, 3.5 and 3.6

Version

The version matched the latest version of NumPy where RandomGenerator(MT19937()) passes all NumPy test.

Documentation

Documentation for the latest release is available on my GitHub pages. Documentation for the latest commit (unreleased) is available under devel.

Plans

This module is essentially complete. There are a few rough edges that need to be smoothed.

• Creation of additional streams from where supported (i.e. a next_stream() method)

Requirements

Building requires:

• Python (2.7, 3.5, 3.6, 3.7)
• NumPy (1.13, 1.14, 1.15)
• Cython (0.26+)
• tempita (0.5+), if not provided by Cython

Testing requires pytest (3.0+).

Note: it might work with other versions but only tested with these versions.

Development and Testing

All development has been on 64-bit Linux, and it is regularly tested on Travis-CI (Linux/OSX) and Appveyor (Windows). The library is occasionally tested on Linux 32-bit and Free BSD 11.1.

Basic tests are in place for all RNGs. The MT19937 is tested against NumPy's implementation for identical results. It also passes NumPy's test suite where still relevant.

Installing

Either install from PyPi using

pip install randomgen

or, if you want the latest version,

pip install git+https://github.com/bashtage/randomgen.git

or from a cloned repo,

python setup.py install

SSE2

dSFTM makes use of SSE2 by default. If you have a very old computer or are building on non-x86, you can install using:

python setup.py install --no-sse2

Windows

Either use a binary installer, or if building from scratch, use Python 3.6 with Visual Studio 2015/2017 Community Edition. It can also be build using Microsoft Visual C++ Compiler for Python 2.7 and Python 2.7.

Using

The separate generators are importable from randomgen

from randomgen import RandomGenerator, ThreeFry, PCG64, MT19937
rg = RandomGenerator(ThreeFry())
rg.random_sample(100)

rg = RandomGenerator(PCG64())
rg.random_sample(100)

# Identical to NumPy
rg = RandomGenerator(MT19937())
rg.random_sample(100)

License

Standard NCSA, plus sub licenses for components.

Performance

Performance is promising, and even the mt19937 seems to be faster than NumPy's mt19937.

Speed-up relative to NumPy (Uniform Doubles)
************************************************************
DSFMT                 184.9%
MT19937                17.3%
PCG32                  83.3%
PCG64                 108.3%
Philox                 -4.9%
ThreeFry              -12.0%
ThreeFry32            -63.9%
Xoroshiro128          159.5%
Xorshift1024          150.4%
Xoshiro256StarStar    145.7%
Xoshiro512StarStar    113.1%

Speed-up relative to NumPy (64-bit unsigned integers)
************************************************************
DSFMT                  17.4%
MT19937                 7.8%
PCG32                  60.3%
PCG64                  73.5%
Philox                -25.5%
ThreeFry              -30.5%
ThreeFry32            -67.8%
Xoroshiro128          124.0%
Xorshift1024          109.4%
Xoshiro256StarStar    100.3%
Xoshiro512StarStar     63.5%

Speed-up relative to NumPy (Standard normals)
************************************************************
DSFMT                 183.0%
MT19937               169.0%
PCG32                 240.7%
PCG64                 231.6%
Philox                131.3%
ThreeFry              118.3%
ThreeFry32             21.6%
Xoroshiro128          332.1%
Xorshift1024          232.4%
Xoshiro256StarStar    306.6%
Xoshiro512StarStar    274.6%