Edit section Single Machine Concurrency

[nushio3]

Add parallel-io to the list of the Single Machine Concurrency libraries.

[nushio3]

parallel-io is the first-aid-cure-all medicine I use whenever I find my Haskell program slow and my many CPU cores vacant except one. For really compute-heavy task we Haskellers have more professional tools such as accelerate in our arsenal, but for any other casual programming parallel is just satisfactory. It is a Haskeller's joy to see all your program suddenly parallelize, since the code you write is 100% pure, and all the CPU cores utilized.

p.s. don't forget to deepseq the results of the individual computations!

[Gabriella439]

Normally I would use the async library for this purpose, specifically the mapConcurrently function or the Concurrently applicative. For example, the parallel function can be implemented either as:

parallel = mapConcurrently id

-- or:

parallel = runConcurrently . sequenceA . Concurrently

Also, both libraries are implemented in roughly the same way.