{-# LANGUAGE CPP, BangPatterns, DeriveDataTypeable #-} module Control.Concurrent.Speculation ( -- * Speculative application spec , spec' , specBy , specBy' , specOn , specOn' -- * Speculative application with transactional rollback , specSTM , specSTM' , specOnSTM , specOnSTM' , specBySTM , specBySTM' -- * Determining if a closure is evaluated , unsafeGetTagBits , unsafeIsEvaluated ) where import Control.Concurrent.STM import Control.Concurrent.Speculation.Internal (returning) import Control.Parallel (par) import Control.Monad (liftM2, unless) import Data.Function (on) #if __GLASGOW_HASKELL__ >= 608 import Data.Bits ((.&.)) import Foreign (sizeOf) import Unsafe.Coerce (unsafeCoerce) -- dynamic pointer tagging is present on this platform #define TAGGED #endif -- * Basic speculation -- | @'spec' g f a@ evaluates @f g@ while forcing @a@, if @g == a@ then @f g@ is returned, otherwise @f a@ is evaluated and returned. Furthermore, if the argument has already been evaluated, we skip the @f g@ computation entirely. If a good guess at the value of @a@ is available, this is one way to induce parallelism in an otherwise sequential task. However, if the guess isn\'t available more cheaply than the actual answer, then this saves no work and if the guess is wrong, you risk evaluating the function twice. Under high load, since 'f g' is computed via the spark queue, the speculation will be skipped and you will obtain the same answer as 'f $! a'. -- --The best-case timeline looks like: -- -- > foreground: [----- a -----] -- > foreground: [-] (check g == a) -- > spark: [----- f g -----] -- > overall: [--- spec g f a ---] -- -- The worst-case timeline looks like: -- -- > foreground: [----- a -----] -- > foreground: [-] (check g == a) -- > foreground: [---- f a ----] -- > spark: [----- f g -----] -- > overall: [-------- spec g f a ---------] -- -- Note that, if @f g@ takes longer than a to compute, in the HEAD release of GHC, @f g@ will be collected and killed during garbage collection. -- -- > foreground: [----- a -----] -- > foreground: [-] (check g == a) -- > foreground: [---- f a ----] -- > spark: [---- f g ----###### (#'s mark when this spark is collectable) -- > overall: [--------- spec g f a --------] -- -- Under high load: -- -- > foreground: [----- a -----] -- > foreground: [-] (check g == a) -- > foreground: [---- f a ----] -- > overall: [-------- spec g f a ---------] -- -- Compare these to the timeline of @f $! a@: -- -- > foreground: [----- a -----] -- > foreground: [---- f a ----] -- > orverall: [---------- f $! a ---------] spec :: Eq a => a -> (a -> b) -> a -> b spec = specBy (==) {-# INLINE spec #-} -- | Unlike 'spec', this version does not check to see if the argument has already been evaluated. This can save -- a small amount of work when you know the argument will always require computation. spec' :: Eq a => a -> (a -> b) -> a -> b spec' = specBy' (==) {-# INLINE spec' #-} -- | 'spec' with a user defined comparison function specBy :: (a -> a -> Bool) -> a -> (a -> b) -> a -> b specBy cmp guess f a | unsafeIsEvaluated a = f a | otherwise = specBy' cmp guess f a {-# INLINE specBy #-} -- | 'spec'' with a user defined comparison function specBy' :: (a -> a -> Bool) -> a -> (a -> b) -> a -> b specBy' cmp guess f a = speculation `par` if cmp guess a then speculation else f a where speculation = f guess {-# INLINE specBy' #-} -- | 'spec' comparing by projection onto another type specOn :: Eq c => (a -> c) -> a -> (a -> b) -> a -> b specOn = specBy . on (==) {-# INLINE specOn #-} -- | 'spec'' comparing by projection onto another type specOn' :: Eq c => (a -> c) -> a -> (a -> b) -> a -> b specOn' = specBy' . on (==) {-# INLINE specOn' #-} -- * STM-based speculation -- | @'specSTM' g f a@ evaluates @fg = do g' <- g; f g'@, while forcing @a@, then if @g' == a@ then @fg@ is returned. Otherwise the side-effects of @fg@ are rolled back and @f a@ is evaluated. @g@ is allowed to be a monadic action, so that we can kickstart the computation of @a@ earlier. -- -- If the argument @a@ is already evaluated, we don\'t bother to perform @fg@ at all. -- -- If a good guess at the value of @a@ is available, this is one way to induce parallelism in an otherwise sequential task. -- -- However, if the guess isn\'t available more cheaply than the actual answer then this saves no work, and if the guess is -- wrong, you risk evaluating the function twice. -- -- The best-case timeline looks like: -- -- > foreground: [--- g >>= f ---] -- > spark: [------- a -------] -- > foreground: [-] (compare g' == a) -- > overall: [---- specSTM g f a ----] -- -- The worst-case timeline looks like: -- -- > foreground: [---- g >>= f ----] -- > spark: [------- a -------] -- > foreground: [-] (check if g' == a) -- > foreground: [--] (rollback) -- > foreground: [------ f a ------] -- > overall: [------------ specSTM g f a ----------------] -- -- Under high load, 'specSTM' degrades less gracefully than 'spec': -- -- > foreground: [---- g >>= f ----] -- > spark: [------- a -------] -- > foreground: [-] (check if g' == a) -- > foreground: [--] (rollback) -- > foreground: [------ f a ------] -- > overall: [--------------------specSTM g f a ------------------------] -- -- Compare these to the timeline of @f $! a@: -- -- > foreground: [------- a -------] -- > foreground: [------ f a ------] -- specSTM :: Eq a => STM a -> (a -> STM b) -> a -> STM b specSTM = specBySTM (returning (==)) {-# INLINE specSTM #-} -- | Unlike 'specSTM', 'specSTM'' doesn't check if the argument has already been evaluated. specSTM' :: Eq a => STM a -> (a -> STM b) -> a -> STM b specSTM' = specBySTM' (returning (==)) {-# INLINE specSTM' #-} -- | 'specSTM' using a user defined comparison function specBySTM :: (a -> a -> STM Bool) -> STM a -> (a -> STM b) -> a -> STM b specBySTM cmp guess f a | unsafeIsEvaluated a = f a | otherwise = specBySTM' cmp guess f a {-# INLINE specBySTM #-} -- | 'specSTM'' using a user defined comparison function specBySTM' :: (a -> a -> STM Bool) -> STM a -> (a -> STM b) -> a -> STM b specBySTM' cmp mguess f a = a `par` do guess <- mguess result <- f guess -- rendezvous with a matching <- cmp guess a unless matching retry return result `orElse` f a {-# INLINE specBySTM' #-} -- | @'specBySTM' . 'on' (==)@ specOnSTM :: Eq c => (a -> STM c) -> STM a -> (a -> STM b) -> a -> STM b specOnSTM = specBySTM . on (liftM2 (==)) {-# INLINE specOnSTM #-} -- | @'specBySTM'' . 'on' (==)@ specOnSTM' :: Eq c => (a -> STM c) -> STM a -> (a -> STM b) -> a -> STM b specOnSTM' = specBySTM' . on (liftM2 (==)) {-# INLINE specOnSTM' #-} -- | Inspect the dynamic pointer tagging bits of a closure. This is an impure function that relies on GHC internals and may falsely return 0, but should never give the wrong tag number if it returns a non-0 value. unsafeGetTagBits :: a -> Int {-# INLINE unsafeGetTagBits #-} #ifndef TAGGED unsafeGetTagBits _ = 0 #else unsafeGetTagBits a = unsafeCoerce (Box a) .&. (sizeOf (undefined :: Int) - 1) -- | Used to inspect tag bits data Box a = Box a #endif -- | Returns a guess as to whether or not a value has been evaluated. This is an impure function that relies on GHC internals and will return false negatives, but no false positives. This is unsafe as the value of this function will vary (from False to True) over the course of otherwise pure invocations! unsafeIsEvaluated :: a -> Bool unsafeIsEvaluated a = unsafeGetTagBits a /= 0 {-# INLINE unsafeIsEvaluated #-}