Parallelizing with reducers

In the last recipe, Combining function calls with reducers, we looked at the ability of reducers to compose multiple sequence processing functions into one function. This saves the effort of creating intermediate data structures.

Another feature of reducers is that they can automatically partition and parallelize the processing of tree-based data structures. This includes Clojure's native vectors and hash maps.

For this recipe, we'll continue the Monte Carlo simulation example that we started in the Partitioning Monte Carlo simulations for better pmap performance recipe. In this case, we'll write a version that uses reducers and see how it performs.

Getting ready

From the Partitioning Monte Carlo simulations for better pmap performance recipe, we'll use the same imports, as well as the rand-point, center-dist, and mc-pi functions. Along with these, we also need to require the reducers and Criterium libraries:

(require '[clojure.core.reducers :as r])
(use 'criterium...