Difference between revisions of "Parallelism"

From HaskellWiki
Jump to navigation Jump to search
(pure /= non-IO)
(Relevant content from "Parallel" transferred to here; various other changes)
 
(2 intermediate revisions by 2 users not shown)
Line 1: Line 1:
  +
''Note: you may want to read [[Parallelism vs. Concurrency]], as the terms have historically been conflated.''
−
Parallelism is about speeding up a program by using multiple processors.
  
   
  +
This page contains notes and information about how to write parallel programs in Haskell.
−
In Haskell we provide two ways to achieve parallelism:
  
−
* Pure parallelism, which can be used to speed up non-IO parts of the program.
  
−
* Concurrency, which can be used for parallelising IO.
  
   
  +
For practicality, the content is GHC-centric at the moment, although this may change as Haskell evolves.
−
Pure Parallelism (Control.Parallel): Speeding up a pure computation using multiple processors. Pure parallelism has these advantages:
  
−
* Guaranteed deterministic (same result every time)
  
−
* no [[race conditions]] or [[deadlocks]]
  
   
  +
== Overview ==
−
[[Concurrency]] (Control.Concurrent): Multiple threads of control that execute "at the same time".
  
−
* Threads are in the IO monad
  
−
* IO operations from multiple threads are interleaved non-deterministically
  
−
* communication between threads must be explicitly programmed
  
−
* Threads may execute on multiple processors simultaneously
  
−
* Dangers: [[race conditions]] and [[deadlocks]]
  
   
  +
GHC provides multi-scale support for parallel programming, from very fine-grained, small "sparks" to nested data parallelism and strategies.
−
Rule of thumb: use Pure Parallelism if you can, Concurrency otherwise.
  
   
−
== Starting points ==
 +
== Getting started ==
   
  +
Speed up your code by making it run on multicore:
−
* '''Control.Parallel'''. The first thing to start with parallel programming in Haskell is the use of par/pseq from the parallel library. Try the Real World Haskell [http://book.realworldhaskell.org/read/concurrent-and-multicore-programming.html chapter on parallelism and concurrency]. The parallelism-specific parts are in the second half of the chapter.
  
  +
: Start with Control.Parallel ('''par''', '''pseq''')
−
* If you need more control, try Strategies or perhaps the Par monad
  
  +
: [[Parallel/Reading|Learn more about parallelism]], then try using Strategies
   
−
== Multicore GHC ==
 +
== GHC parallelism specifics ==
  +
  +
You get access to parallelism operations by importing the library [http://www.haskell.org/ghc/docs/latest/html/libraries/base/Control-Parallel.html Control.Parallel].
   
 
{{GHC/Multicore}}
  
{{GHC/Multicore}}
  +
  +
Support for low-level parallelism features of modern processors continues to be added. As of version 7.8, GHC includes the ability to [https://ghc.haskell.org/trac/ghc/wiki/SIMD emit SIMD instructions], and also has a rudimentary ability to use [[AtomicMemoryOps|atomic memory operations]].
  +
  +
== Community ==
  +
  +
* Ask questions on [[Mailing lists|Haskell Cafe]]
  +
* See what [https://groups.google.com/group/parallel-haskell parallel-haskell] researchers and developers are working on
  +
* Get the latest from the [[Parallel GHC Project]]
  +
* Follow [http://twitter.com/#!/parallelhaskell @parallelhaskell] on Twitter [[image:Twitter-mini.png]]
  +
* StackOverflow on Haskell [http://stackoverflow.com/questions/tagged/haskell%2bparallel parallelism]
  +
  +
== News ==
  +
  +
* 2012-04-20 [http://www.well-typed.com/blog/65 Parallel Haskell Digest 9]
  +
* 2012-03-02 [http://www.well-typed.com/blog/64 Parallel Haskell Digest 8]
  +
* 2011-12-24 [http://www.well-typed.com/blog/62 Parallel Haskell Digest 7]
  +
* 2011-11-21 [http://www.haskell.org/pipermail/haskell-cafe/2011-November/097008.html Job Opportunity at Parallel Scientific]
  +
  +
== Tools ==
  +
  +
* [[ThreadScope]] - parallel programs not getting faster? Use the ThreadScope debugger and watch those sparks fly.
  +
* Various [[Applications and libraries/Concurrency and parallelism|libraries]], including those for parallelism.
  +
  +
== Documentation ==
  +
  +
* [[Parallel/Glossary|Glossary]]
  +
* [[Parallel/Reading|Learning to use parallelism in Haskell]]
  ⚫
* [[Parallel/Research|Current research]]
  +
* [http://chimera.labs.oreilly.com/books/1230000000929/index.html Parallel ... Programming in Haskell] (online book)
   
 
== Alternative approaches ==
  
== Alternative approaches ==
   
  +
* [https://www.cse.chalmers.se/edu/year/2012/course/pfp/Papers/strategies10.pdf Seq no more: Better Strategies for Parallel Haskell]
−
* Nested data parallelism: a parallel programming model based on bulk data parallelism, in the form of the [http://www.haskell.org/haskellwiki/GHC/Data_Parallel_Haskell DPH] and [http://hackage.haskell.org/package/repa Repa] libraries for transparently parallel arrays.
  
−
* [https://hackage.haskell.org/package/monad-par monad-par] and [https://hackage.haskell.org/package/lvish LVish] provide Par monads that can structure parallel computations over "monotonic" data structures, which in turn can be used from within purely functional programs.
  
−
* [OLD] Intel [http://software.intel.com/en-us/blogs/2010/05/27/announcing-intel-concurrent-collections-for-haskell-01/ Concurrent Collections for Haskell]: a graph-oriented parallel programming model.
  
   
 
== See also ==
  
== See also ==
   
−
* The [[Parallel|parallelism and concurrency portal]]
 +
* [[:Category:Parallel|Parallel]] category
−
* Parallel [[Parallel/Reading|reading list]]
 +
* Parallelism [[Parallel/Research|research]]
  +
⚫
* [[Parallel/Research|Ongoing research in Parallel Haskell]]
  
  +
  +
[[Category:GHC]]
  +
[[Category:Parallel]]

Latest revision as of 12:02, 9 May 2024

Note: you may want to read Parallelism vs. Concurrency, as the terms have historically been conflated.

This page contains notes and information about how to write parallel programs in Haskell.

For practicality, the content is GHC-centric at the moment, although this may change as Haskell evolves.

Overview

GHC provides multi-scale support for parallel programming, from very fine-grained, small "sparks" to nested data parallelism and strategies.

Getting started

Speed up your code by making it run on multicore:

Start with Control.Parallel (par, pseq)
Learn more about parallelism, then try using Strategies

GHC parallelism specifics

You get access to parallelism operations by importing the library Control.Parallel.

Since 2004, GHC supports running programs in parallel on an SMP or multi-core machine. How to do it:

  • Compile your program using the -threaded switch.
  • Run the program with +RTS -N2 to use 2 threads, for example (RTS stands for runtime system; see the GHC users' guide). You should use a -N value equal to the number of CPU cores on your machine (not including Hyper-threading cores). As of GHC v6.12, you can leave off the number of cores and all available cores will be used (you still need to pass -N however, like so: +RTS -N).
  • Concurrent threads (forkIO) will run in parallel, and you can also use the par combinator and Strategies from the Control.Parallel.Strategies module to create parallelism.
  • Use +RTS -sstderr for timing stats.
  • To debug parallel program performance, use ThreadScope.

Support for low-level parallelism features of modern processors continues to be added. As of version 7.8, GHC includes the ability to emit SIMD instructions, and also has a rudimentary ability to use atomic memory operations.

Community

News

Tools

  • ThreadScope - parallel programs not getting faster? Use the ThreadScope debugger and watch those sparks fly.
  • Various libraries, including those for parallelism.

Documentation

Alternative approaches

See also

Retrieved from "https://wiki.haskell.org/index.php?title=Parallelism&oldid=66640"