pcg_random

PCGRandom

 _____ _____ _____ _____           _           
|  _  |     |   __| __  |___ ___ _| |___ _____ 
|   __|   --|  |  |    -| .'|   | . | . |     |
|__|  |_____|_____|__|__|__,|_|_|___|___|_|_|_|

Ruby wrappers for the PCGRandom family of random number generators.

Installation

Prerequisites

• A C-compiler (any modern C compiler is supported on *nix, Windows isn't supported yet)
• Modified C source code for pcg_random available here for building libpcg_random.
• Follow the build instructions given here to build libpcg_random

Installing the gem:

• Rubygems:

gem install pcg_random

You can also configure the path to libpcg_random by specifying it's path explicitly:

# specify the custome include path
gem install pcg_random --with-pcg_random-include=directory/containing/pcg_variants.h

# or, specify the custom path to the libpcg_random binary
gem install pcg_random --with-pcg_random-lib=directory/containing/libpcg_random.a

# Or, specify both the lib/ and include/ together to be under path/to/some/directory
gem install pcg_random --with-pcg_random-dir=path/to/some/directory

Development

TODO: Create a CONTRIBTING.md

Install pcg_random from source using the above steps. Using the above links are important because the library must be compiled with additional flags to be used for gem-development. There are also some issues in the original repository that are addressed in my fork.

Using the extension

Either jump into your REPL of choice:

~$ irb
2.3.0 :001 > require 'pcg_random'
 => true

Or go ahead and try the API, it's built to be as close to Kernel::Random as possible!

2.3.0 :002 > seed = PCGRanom.new_seed
 => 315431527813695542666700950080169144774
2.3.0 :003 > rng = PCGRanom.new 1234
 => #<PCGRandom:0x00000001bf3eb0 @seed=1234>

For detailed instructions, keep checking this repo, I'm working on getting a wiki up.

Benchmarks

PCGRandom.new_seed

Benchmark of Random.new_seed vs PCGRandom.new_seed
==================================================

Benchmark created around: 2016-03-30 14:54:16 +0000

Running for: 100000 iterations

Rehearsal ------------------------------------------------------
Random.new_seed      0.090000   1.240000   1.330000 (  1.323970)
PCGRandom.new_seed   0.100000   1.150000   1.250000 (  1.254944)
--------------------------------------------- total: 2.580000sec

                         user     system      total        real
Random.new_seed      0.110000   1.200000   1.310000 (  1.321571)
PCGRandom.new_seed   0.100000   1.130000   1.230000 (  1.222634)

PCGRandom.raw_seed(16)

Benchmark of Random.raw_seed(16) vs PCGRandom.raw_seed(16)
==================================================

Benchmark created around: 2016-03-30 14:41:26 +0000

Running for: 100000 iterations

Rehearsal ----------------------------------------------------------
Random.raw_seed(16)      0.080000   1.440000   1.520000 (  1.515513)
PCGRandom.raw_seed(16)   0.110000   1.320000   1.430000 (  1.444519)
------------------------------------------------- total: 2.950000sec

                             user     system      total        real
Random.raw_seed(16)      0.080000   1.470000   1.550000 (  1.563351)
PCGRandom.raw_seed(16)   0.090000   1.340000   1.430000 (  1.441253)

Roadmap

• Achieve full parity with Kernel::Random
• Get a wiki explaining how to go about using this gem
• Showcase benchmarks in some other place