RomainFranceschini

require "quartz"
require "statistics"
require "tablo"

include Statistics::Distributions
include Quartz

class Arrival < AtomicModel
  output :user
  precision :milli

RomainFranceschini

Keybase proof

I hereby claim:

• I am rumenzu on github.
• I am rumenzu (https://keybase.io/rumenzu) on keybase.
• I have a public key ASCtLGxO5vxNuWWjPxuceTmolSSxvMDC5DzV0D7U-6px3go

To claim this, I am signing this object:

RomainFranceschini

A list of Crystal shards for science

Linear algebra

• crystal-gsl - GNU Scientific Library bindings
• crystalla - library for Numerical Methods. It binds to LAPACK and is unashamedly inspired by Numpy
• linalg - Linear algebra library inspired by MATLAB and SciPy.linalg
• linprog - Wrapper for Coin-OR SYMPHONY library - LP and MILP solver
• crystaledge - A pure Vector Math library

Optimization

RomainFranceschini

The following results and those shown on the chart are obtained using OpenJDK7 on Ubuntu
Models Ruby 2.1.2  jruby-1.7.13  rbx-2.2.10
3000   64.379486   37.682        29.618011

JRuby becomes fastest when I run the benchmark on my MacBook with HotSpot 1.8
Models Ruby 2.1.2  jruby-1.7.13  rbx-2.2.10
3000   77.023007   31.381        33.223119

Obviously, something was wrong with my previous chart (https://twitter.com/romain1189/status/486443279589056512). As the results were generated with my MacBook, something skewed the result. Maybe TimeMachine or something. Apologizes for this.

RomainFranceschini

# To run this benchmark
# $ git clone https://github.com/devs-ruby/devs .
# $ gem build devs.gemspec && gem install devs-0.5.1.gem
# $ gem install devs-models
# $ git clone https://github.com/devs-ruby/devstone .
# $ gem build devs.gemspec && gem install devstone-0.0.1.gem

require 'devstone'

opt_base = {