speculation-0.2.0: A framework for safe, programmable, speculative parallelism




specSTM :: Eq a => a -> (a -> STM b) -> a -> STM bSource

specSTM g f a evaluates f g while forcing a, if g == a then f g is returned. Otherwise the side-effects of the current STM transaction are rolled back and f a is evaluated.

If the argument a is already evaluated, we don't bother to perform f g 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.

 specSTM a f a = f $! a

The best-case timeline looks like:

 [------ f g ------]
     [------- a -------]
 [--- specSTM g f a ---]

The worst-case timeline looks like:

 [------ f g ------] 
     [------- a -------]
                       [-- rollback --]
                                      [------ f a ------]     
 [------------------ spec g f a ------------------------]

Compare these to the timeline of f $! a:

 [------- a -------]
                   [------ f a ------]

specSTM' :: Eq a => a -> (a -> STM b) -> a -> STM bSource

Unlike specSTM, specSTM' doesn't check if the argument has already been evaluated.