README.md

Parallel enumeration with Ractors

An exploration into Ractors.

Set up

Run bundle install

Using the ParallelEnumerable

Simply require the parallel_enumerable, and call parallel on your original enumerable. It will parallelize as much as it can any operations made on that collection. For now only the map operation is parallelized, but more will come.

require 'parallel_enumerable'

collection.parallel.map { |n| cpu_intensive_operation(n) }

Testing

Run rake test

Benchmarking

You can benchmark the different ways to run a CPU intensive operation (factorial), using different strategies:

❯ bin/benchmark                                        
Rehearsal -------------------------------------------------------
sequential            3.700499   2.552881   6.253380 (  6.275990)
threads               3.551265   1.740886   5.292151 (  5.332419)
ractors               4.870103   1.153873   6.023976 (  1.774149)
parallel_enumerable   4.864191   1.229560   6.093751 (  1.896910)
--------------------------------------------- total: 23.663258sec

                          user     system      total        real
sequential            3.610238   3.002262   6.612500 (  6.679901)
threads               3.570723   2.033741   5.604464 (  5.643820)
ractors               4.801946   1.140257   5.942203 (  1.749837)
parallel_enumerable   4.816865   1.048911   5.865776 (  1.701564)