----------------------------------------------------------------------------- -- | -- Module : Control.Parallel.Strategies -- Copyright : (c) The University of Glasgow 2001-2009 -- License : BSD-style (see the file libraries/base/LICENSE) -- -- Maintainer : libraries@haskell.org -- Stability : experimental -- Portability : portable -- -- Parallel Evaluation Strategies, or Strategies for short, specify a -- way to evaluate a structure with components in sequence or in -- parallel. -- -- Strategies are for expressing /deterministic parallelism/: -- the result of the program is unaffected by evaluating in parallel. -- For non-deterministic parallel programming, see -- "Control.Concurrent". -- -- Strategies let you separate the description of parallelism from the -- logic of your program, enabling modular parallelism. -- -- Version 1.x -- -- The original Strategies design is described in -- <http://www.macs.hw.ac.uk/~dsg/gph/papers/html/Strategies/strategies.html> -- and the code was written by -- Phil Trinder, Hans-Wolfgang Loidl, Kevin Hammond et al. -- -- Version 2.x -- -- Later, during work on the shared-memory implementation of -- parallelism in GHC, we discovered that the original formulation of -- Strategies had some problems, in particular it lead to space leaks -- and difficulties expressing speculative parallelism. Details are in -- the paper \"Runtime Support for Multicore Haskell\" <http://www.haskell.org/~simonmar/papers/multicore-ghc.pdf>. -- -- This module has been rewritten in version 2. The main change is to -- the 'Strategy a' type synonym, which was previously @a -> Done@ and -- is now @a -> a@. This change helps to fix the space leak described -- in \"Runtime Support for Multicore Haskell\". The problem is that -- the runtime will currently retain the memory referenced by all -- sparks, until they are evaluated. Hence, we must arrange to -- evaluate all the sparks eventually, just in case they aren't -- evaluated in parallel, so that they don't cause a space leak. This -- is why we must return a \"new\" value after applying a 'Strategy', -- so that the application can evaluate each spark created by the -- 'Strategy'. -- -- The simple rule is this: you /must/ use the result of applying -- a 'Strategy' if the strategy creates parallel sparks, and you -- should probably discard the the original value. If you don't -- do this, currently it may result in a space leak. In the -- future (GHC 6.14), it will probably result in lost parallelism -- instead, as we plan to change GHC so that unreferenced sparks -- are discarded rather than retained (we can't make this change -- until most code is switched over to this new version of -- Strategies, because code using the old verison of Strategies -- would be broken by the change in policy). -- -- The other changes in version 2.x are: -- -- * Strategies can now be defined using a convenient Applicative -- type Eval. e.g. parList s = unEval $ traverse (Par . s) -- -- * 'parList' has been generalised to 'parTraverse', which works on -- any 'Traversable' type. -- -- * 'parList' and 'parBuffer' have versions specialised to 'rwhnf', -- and there are transformation rules that automatically translate -- e.g. @parList rwnhf@ into a call to the optimised version. -- -- * 'NFData' is deprecated; please use the @DeepSeq@ class in the @deepseq@ -- package instead. Note that since the 'Strategy' type changed, 'rnf' -- is no longer a 'Strategy': use 'rdeepseq' instead. ----------------------------------------------------------------------------- module Control.Parallel.Strategies ( -- * Strategy type and basic operations Strategy, using, withStrategy, rwhnf, rdeepseq, -- * Tuple strategies seqPair, parPair, seqTriple, parTriple, -- * General traversals seqTraverse, parTraverse, -- * List strategies parList, seqList, parListN, parListChunk, parMap, parBuffer, -- * Simple list strategies parListWHNF, parBufferWHNF, -- * Strategy composition operators ($|), ($||), (.|), (.||), (-|), (-||), -- * Building strategies Eval(..), unEval, -- * Deprecated functionality NFData(..), Done, demanding, sparking, (>|), (>||), r0, ) where import Data.Traversable import Control.Applicative import Control.Parallel import Control.DeepSeq -- imports for deprecated APIs import Data.Int import Data.Word import Data.Complex import qualified Data.Map import qualified Data.IntMap import qualified Data.Set import qualified Data.IntSet import Data.Tree import Data.Array import GHC.Real -- ----------------------------------------------------------------------------- -- Eval -- | `Eval` is an Applicative Functor that makes it easier to define -- parallel strategies that involve traversing structures. -- -- a 'Seq' value will be evaluated strictly in sequence in its context, -- whereas a 'Par' value wraps an expression that may be evaluated in -- parallel. The Applicative instance allows sequential composition, -- making it possible to describe an evaluateion strategy by composing -- 'Par' and 'Seq' with '<*>'. -- -- For example, -- -- > parList :: Strategy a -> Strategy [a] -- > parList strat = unEval . traverse (Par . strat) -- -- > seqPair :: Strategy a -> Strategy b -> Strategy (a,b) -- > seqPair f g (a,b) = unEval $ (,) <$> Seq (f a) <*> Seq (g b) -- data Eval a = Seq a | Par a | Lazy a unEval :: Eval a -> a unEval (Seq a) = a unEval (Par a) = a unEval (Lazy a) = a instance Functor Eval where fmap f (Par a) = Seq (a `par` f a) -- only Par once: context becomes Seq fmap f (Seq a) = Seq (a `pseq` f a) fmap f (Lazy a) = Lazy (f a) instance Applicative Eval where pure a = Lazy a Par f <*> x = f `par` fmap f x Seq f <*> x = f `pseq` fmap f x Lazy f <*> x = fmap f x -- ----------------------------------------------------------------------------- -- Strategies -- | A 'Strategy' is a function that embodies a parallel evaluation strategy. -- The function traverses (parts of) its argument, evaluating subexpressions -- in parallel or in sequence. -- -- A 'Strategy' may do an arbitrary amount of evaluation of its -- argument, but should not return a value different from the one it -- was passed. -- -- Parallel computations may be discarded by the runtime system if the -- program no longer requires their result, which is why a 'Strategy' -- function returns a new value equivalent to the old value. The -- intention is that the program applies the 'Strategy' to a -- structure, and then uses the returned value, discarding the old -- value. This idiom is expressed by the 'using' function. -- type Strategy a = a -> a -- | evaluate a value using the given 'Strategy'. -- -- > using x s = s x -- using :: a -> Strategy a -> a using x s = s x -- | evaluate a value using the given 'Strategy'. This is simply -- 'using' with the arguments reversed, and is equal to '($)'. -- withStrategy :: Strategy a -> a -> a withStrategy = ($) -- | A 'Strategy' that simply evaluates its argument to Weak Head Normal -- Form (i.e. evaluates it as far as the topmost constructor). rwhnf :: Strategy a rwhnf a = a -- | A 'Strategy' that fully evaluates its argument -- -- > rdeepseq a = deepseq a `pseq` a -- rdeepseq :: DeepSeq a => Strategy a rdeepseq a = deepseq a `pseq` a -- ----------------------------------------------------------------------------- -- Tuples seqPair :: Strategy a -> Strategy b -> Strategy (a,b) seqPair f g (a,b) = unEval $ (,) <$> Seq (f a) <*> Seq (g b) parPair :: Strategy a -> Strategy b -> Strategy (a,b) parPair f g (a,b) = unEval $ (,) <$> Par (f a) <*> Par (g b) seqTriple :: Strategy a -> Strategy b -> Strategy c -> Strategy (a,b,c) seqTriple f g h (a,b,c) = unEval $ (,,) <$> Seq (f a) <*> Seq (g b) <*> Seq (h c) parTriple :: Strategy a -> Strategy b -> Strategy c -> Strategy (a,b,c) parTriple f g h (a,b,c) = unEval $ (,,) <$> Par (f a) <*> Par (g b) <*> Par (h c) -- ----------------------------------------------------------------------------- -- General sequential/parallel traversals -- | A strategy that traverses a container data type with an instance -- of 'Traversable', and sparks each of the elements using the supplied -- strategy. parTraverse :: Traversable t => Strategy a -> Strategy (t a) parTraverse strat = unEval . traverse (Par . strat) -- | A strategy that traverses a container data type with an instance -- of 'Traversable', and evaluates each of the elements in left-to-right -- sequence using the supplied strategy. seqTraverse :: Traversable t => Strategy a -> Strategy (t a) seqTraverse strat = unEval . traverse (Seq . strat) {-# SPECIALISE parTraverse :: Strategy a -> Strategy [a] #-} {-# SPECIALISE seqTraverse :: Strategy a -> Strategy [a] #-} -- ----------------------------------------------------------------------------- -- Lists -- | Spark each of the elements of a list using the given strategy. -- Equivalent to 'parTraverse' at the list type. parList :: Strategy a -> Strategy [a] parList = parTraverse -- | Evaluate each of the elements of a list sequentially from left to right -- using the given strategy. Equivalent to 'seqTraverse' at the list type. seqList :: Strategy a -> Strategy [a] seqList = seqTraverse parListN :: Int -> Strategy a -> Strategy [a] parListN n strat xs = parList strat as ++ bs where (as,bs) = splitAt n xs parListChunk :: Int -> Strategy a -> Strategy [a] parListChunk n strat xs = concat (parList (seqList strat) (chunk n xs)) chunk :: Int -> [a] -> [[a]] chunk _ [] = [] chunk n xs = as : chunk n bs where (as,bs) = splitAt n xs parMap :: Strategy b -> (a -> b) -> [a] -> [b] parMap strat f = parList strat . map f -- ----------------------------------------------------------------------------- -- parBuffer -- | Applies a strategy to the nth element of list when the head is demanded. -- More precisely: -- -- * semantics: @parBuffer n s = id :: [a] -> [a]@ -- -- * dynamic behaviour: evalutates the nth element of the list when the -- head is demanded. -- -- The idea is to provide a `rolling buffer' of length n. It is a -- better than 'parList' for a lazy stream, because 'parList' will -- evaluate the entire list, whereas 'parBuffer' will only evaluate a -- fixed number of elements ahead. parBuffer :: Int -> Strategy a -> [a] -> [a] parBuffer n strat xs = parBufferWHNF n (map strat xs) -- ----------------------------------------------------------------------------- -- Simple strategies -- These are non-compositional strategies that might be more efficient -- than their more general counterparts. We use RULES to do the -- specialisation. {-# RULES "parList/rwhnf" parList rwhnf = parListWHNF "parList/id" parList id = parListWHNF "parBuffer/rwhnf" forall n . parBuffer n rwhnf = parBufferWHNF n "parBuffer/id" forall n . parBuffer n id = parBufferWHNF n #-} -- | version of 'parList' specialised to 'rwhnf'. This version is -- much simpler, and may be faster than 'parList rwhnf'. You should -- never need to use this directly, since 'parList rwhnf' is -- automatically optimised to 'parListWHNF'. It is here for -- experimentation purposes only. parListWHNF :: Strategy [a] parListWHNF [] = [] parListWHNF (x:xs) = x `par` parListWHNF xs -- | version of 'parBuffer' specialised to 'rwhnf'. You should -- never need to use this directly, since 'parBuffer rwhnf' is -- automatically optimised to 'parBufferWHNF'. It is here for -- experimentation purposes only. parBufferWHNF :: Int -> Strategy [a] parBufferWHNF n0 xs0 = ret xs0 (start n0 xs0) where ret (x:xs) (y:ys) = y `par` (x : ret xs ys) ret xs _ = xs start _ [] = [] start 0 ys = ys start n (y:ys) = y `par` start (n-1) ys ------------------------------------------------------------------------------ -- * Strategic Function Application ------------------------------------------------------------------------------ {- These are very handy when writing pipeline parallelism asa sequence of @$@, @$|@ and @$||@'s. There is no need of naming intermediate values in this case. The separation of algorithm from strategy is achieved by allowing strategies only as second arguments to @$|@ and @$||@. -} -- | Sequential function application. The argument is evaluated using -- the given strategy before it is given to the function. ($|) :: (a -> b) -> Strategy a -> a -> b f $| s = \ x -> let z = x `using` s in z `pseq` f z -- | Parallel function application. The argument is evaluated using -- the given strategy, in parallel with the function application. ($||) :: (a -> b) -> Strategy a -> a -> b f $|| s = \ x -> let z = x `using` s in z `par` f z -- | Sequential function composition. The result of -- the second function is evaluated using the given strategy, -- and then given to the first function. (.|) :: (b -> c) -> Strategy b -> (a -> b) -> (a -> c) (.|) f s g = \ x -> let z = s (g x) in z `pseq` f z -- | Parallel function composition. The result of the second -- function is evaluated using the given strategy, -- in parallel with the application of the first function. (.||) :: (b -> c) -> Strategy b -> (a -> b) -> (a -> c) (.||) f s g = \ x -> let z = s (g x) in z `par` f z -- | Sequential inverse function composition, -- for those who read their programs from left to right. -- The result of the first function is evaluated using the -- given strategy, and then given to the second function. (-|) :: (a -> b) -> Strategy b -> (b -> c) -> (a -> c) (-|) f s g = \ x -> let z = s (f x) in z `pseq` g z -- | Parallel inverse function composition, -- for those who read their programs from left to right. -- The result of the first function is evaluated using the -- given strategy, in parallel with the application of the -- second function. (-||) :: (a -> b) -> Strategy b -> (b -> c) -> (a -> c) (-||) f s g = \ x -> let z = s (f x) in z `par` g z -- ----------------------------------------------------------------------------- -- Old/deprecated stuff {-# DEPRECATED Done "The Strategies type is now a -> a, not a -> Done" #-} type Done = () {-# DEPRECATED demanding "Use pseq or $| instead" #-} demanding :: a -> Done -> a demanding = flip pseq {-# DEPRECATED sparking "Use par or $|| instead" #-} sparking :: a -> Done -> a sparking = flip par {-# DEPRECATED (>|) "Use pseq or $| instead" #-} (>|) :: Done -> Done -> Done (>|) = Prelude.seq {-# DEPRECATED (>||) "Use par or $|| instead" #-} (>||) :: Done -> Done -> Done (>||) = par {-# DEPRECATED r0 "Strategies must return a result, there is no r0 any more" #-} r0 :: a -> () r0 _ = () {-# DEPRECATED NFData "Use DeepSeq from deepseq:Control.DeepSeq instead" #-} {-# DEPRECATED rnf "Use rdeepseq instead" #-} class NFData a where -- | Reduces its argument to (head) normal form. rnf :: a -> () -- Default method. Useful for base types. A specific method is necessay for -- constructed types rnf a = a `seq` () {-# DEPRECATED NFDataIntegral "Use DeepSeqIntegral from deepseq:Control.DeepSeq instead" #-} {-# DEPRECATED NFDataOrd "Use DeepSeqOrd from deepseq:Control.DeepSeq instead" #-} class (NFData a, Integral a) => NFDataIntegral a class (NFData a, Ord a) => NFDataOrd a instance (NFData a, NFData b) => NFData (a,b) where rnf (x,y) = rnf x `seq` rnf y instance (NFData a, NFData b, NFData c) => NFData (a,b,c) where rnf (x,y,z) = rnf x `seq` rnf y `seq` rnf z instance (NFData a, NFData b, NFData c, NFData d) => NFData (a,b,c,d) where rnf (x1,x2,x3,x4) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 instance (NFData a1, NFData a2, NFData a3, NFData a4, NFData a5) => NFData (a1, a2, a3, a4, a5) where rnf (x1, x2, x3, x4, x5) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 `seq` rnf x5 instance (NFData a1, NFData a2, NFData a3, NFData a4, NFData a5, NFData a6) => NFData (a1, a2, a3, a4, a5, a6) where rnf (x1, x2, x3, x4, x5, x6) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 `seq` rnf x5 `seq` rnf x6 instance (NFData a1, NFData a2, NFData a3, NFData a4, NFData a5, NFData a6, NFData a7) => NFData (a1, a2, a3, a4, a5, a6, a7) where rnf (x1, x2, x3, x4, x5, x6, x7) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 `seq` rnf x5 `seq` rnf x6 `seq` rnf x7 instance (NFData a1, NFData a2, NFData a3, NFData a4, NFData a5, NFData a6, NFData a7, NFData a8) => NFData (a1, a2, a3, a4, a5, a6, a7, a8) where rnf (x1, x2, x3, x4, x5, x6, x7, x8) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 `seq` rnf x5 `seq` rnf x6 `seq` rnf x7 `seq` rnf x8 instance (NFData a1, NFData a2, NFData a3, NFData a4, NFData a5, NFData a6, NFData a7, NFData a8, NFData a9) => NFData (a1, a2, a3, a4, a5, a6, a7, a8, a9) where rnf (x1, x2, x3, x4, x5, x6, x7, x8, x9) = rnf x1 `seq` rnf x2 `seq` rnf x3 `seq` rnf x4 `seq` rnf x5 `seq` rnf x6 `seq` rnf x7 `seq` rnf x8 `seq` rnf x9 instance NFData Int instance NFData Integer instance NFData Float instance NFData Double instance NFData Int8 instance NFData Int16 instance NFData Int32 instance NFData Int64 instance NFData Word8 instance NFData Word16 instance NFData Word32 instance NFData Word64 instance NFDataIntegral Int instance NFDataOrd Int --Rational and complex numbers. instance (Integral a, NFData a) => NFData (Ratio a) where rnf (x:%y) = rnf x `seq` rnf y `seq` () instance (RealFloat a, NFData a) => NFData (Complex a) where rnf (x:+y) = rnf x `seq` rnf y `seq` () instance NFData Char instance NFData Bool instance NFData () instance NFData a => NFData (Maybe a) where rnf Nothing = () rnf (Just x) = rnf x instance (NFData a, NFData b) => NFData (Either a b) where rnf (Left x) = rnf x rnf (Right y) = rnf y instance (NFData k, NFData a) => NFData (Data.Map.Map k a) where rnf = rnf . Data.Map.toList instance NFData a => NFData (Data.Set.Set a) where rnf = rnf . Data.Set.toList instance NFData a => NFData (Data.Tree.Tree a) where rnf (Data.Tree.Node r f) = rnf r `seq` rnf f instance NFData a => NFData (Data.IntMap.IntMap a) where rnf = rnf . Data.IntMap.toList instance NFData Data.IntSet.IntSet where rnf = rnf . Data.IntSet.toList instance NFData a => NFData [a] where rnf [] = () rnf (x:xs) = rnf x `seq` rnf xs instance (Ix a, NFData a, NFData b) => NFData (Array a b) where rnf x = rnf (bounds x) `seq` rnf (elems x) `seq` ()