2 % (c) The GRASP/AQUA Project, Glasgow University, 1993-1998
4 \section[Specialise]{Stamping out overloading, and (optionally) polymorphism}
7 module Specialise ( specProgram ) where
9 #include "HsVersions.h"
11 import CmdLineOpts ( DynFlags, DynFlag(..), dopt )
12 import Id ( Id, idName, idType, mkTemplateLocals, mkUserLocal,
13 idSpecialisation, setIdNoDiscard, isExportedId,
14 modifyIdInfo, idUnfolding
16 import IdInfo ( zapSpecPragInfo )
20 import Type ( Type, mkTyVarTy, splitSigmaTy, splitFunTysN,
21 tyVarsOfType, tyVarsOfTypes, tyVarsOfTheta, applyTys,
22 mkForAllTys, boxedTypeKind
24 import Subst ( Subst, mkSubst, substTy, mkSubst, substBndrs, extendSubstList, mkInScopeSet,
25 substId, substAndCloneId, substAndCloneIds, lookupIdSubst, substInScope
27 import Var ( TyVar, mkSysTyVar, setVarUnique )
31 import CoreUtils ( applyTypeToArgs )
32 import CoreUnfold ( certainlyWillInline )
33 import CoreFVs ( exprFreeVars, exprsFreeVars )
34 import CoreLint ( beginPass, endPass )
35 import PprCore ( pprCoreRules )
36 import Rules ( addIdSpecialisations, lookupRule )
38 import UniqSupply ( UniqSupply,
39 UniqSM, initUs_, thenUs, thenUs_, returnUs, getUniqueUs,
40 getUs, setUs, uniqFromSupply, splitUniqSupply, mapUs
42 import Name ( nameOccName, mkSpecOcc, getSrcLoc )
44 import Maybes ( MaybeErr(..), catMaybes, maybeToBool )
45 import ErrUtils ( dumpIfSet_dyn )
47 import List ( partition )
48 import Util ( zipEqual, zipWithEqual, mapAccumL )
55 %************************************************************************
57 \subsection[notes-Specialise]{Implementation notes [SLPJ, Aug 18 1993]}
59 %************************************************************************
61 These notes describe how we implement specialisation to eliminate
64 The specialisation pass works on Core
65 syntax, complete with all the explicit dictionary application,
66 abstraction and construction as added by the type checker. The
67 existing type checker remains largely as it is.
69 One important thought: the {\em types} passed to an overloaded
70 function, and the {\em dictionaries} passed are mutually redundant.
71 If the same function is applied to the same type(s) then it is sure to
72 be applied to the same dictionary(s)---or rather to the same {\em
73 values}. (The arguments might look different but they will evaluate
76 Second important thought: we know that we can make progress by
77 treating dictionary arguments as static and worth specialising on. So
78 we can do without binding-time analysis, and instead specialise on
79 dictionary arguments and no others.
88 and suppose f is overloaded.
90 STEP 1: CALL-INSTANCE COLLECTION
92 We traverse <body>, accumulating all applications of f to types and
95 (Might there be partial applications, to just some of its types and
96 dictionaries? In principle yes, but in practice the type checker only
97 builds applications of f to all its types and dictionaries, so partial
98 applications could only arise as a result of transformation, and even
99 then I think it's unlikely. In any case, we simply don't accumulate such
100 partial applications.)
105 So now we have a collection of calls to f:
109 Notice that f may take several type arguments. To avoid ambiguity, we
110 say that f is called at type t1/t2 and t3/t4.
112 We take equivalence classes using equality of the *types* (ignoring
113 the dictionary args, which as mentioned previously are redundant).
115 STEP 3: SPECIALISATION
117 For each equivalence class, choose a representative (f t1 t2 d1 d2),
118 and create a local instance of f, defined thus:
120 f@t1/t2 = <f_rhs> t1 t2 d1 d2
122 f_rhs presumably has some big lambdas and dictionary lambdas, so lots
123 of simplification will now result. However we don't actually *do* that
124 simplification. Rather, we leave it for the simplifier to do. If we
125 *did* do it, though, we'd get more call instances from the specialised
126 RHS. We can work out what they are by instantiating the call-instance
127 set from f's RHS with the types t1, t2.
129 Add this new id to f's IdInfo, to record that f has a specialised version.
131 Before doing any of this, check that f's IdInfo doesn't already
132 tell us about an existing instance of f at the required type/s.
133 (This might happen if specialisation was applied more than once, or
134 it might arise from user SPECIALIZE pragmas.)
138 Wait a minute! What if f is recursive? Then we can't just plug in
139 its right-hand side, can we?
141 But it's ok. The type checker *always* creates non-recursive definitions
142 for overloaded recursive functions. For example:
144 f x = f (x+x) -- Yes I know its silly
148 f a (d::Num a) = let p = +.sel a d
150 letrec fl (y::a) = fl (p y y)
154 We still have recusion for non-overloaded functions which we
155 speciailise, but the recursive call should get specialised to the
156 same recursive version.
162 All this is crystal clear when the function is applied to *constant
163 types*; that is, types which have no type variables inside. But what if
164 it is applied to non-constant types? Suppose we find a call of f at type
165 t1/t2. There are two possibilities:
167 (a) The free type variables of t1, t2 are in scope at the definition point
168 of f. In this case there's no problem, we proceed just as before. A common
169 example is as follows. Here's the Haskell:
174 After typechecking we have
176 g a (d::Num a) (y::a) = let f b (d'::Num b) (x::b) = +.sel b d' x x
177 in +.sel a d (f a d y) (f a d y)
179 Notice that the call to f is at type type "a"; a non-constant type.
180 Both calls to f are at the same type, so we can specialise to give:
182 g a (d::Num a) (y::a) = let f@a (x::a) = +.sel a d x x
183 in +.sel a d (f@a y) (f@a y)
186 (b) The other case is when the type variables in the instance types
187 are *not* in scope at the definition point of f. The example we are
188 working with above is a good case. There are two instances of (+.sel a d),
189 but "a" is not in scope at the definition of +.sel. Can we do anything?
190 Yes, we can "common them up", a sort of limited common sub-expression deal.
193 g a (d::Num a) (y::a) = let +.sel@a = +.sel a d
194 f@a (x::a) = +.sel@a x x
195 in +.sel@a (f@a y) (f@a y)
197 This can save work, and can't be spotted by the type checker, because
198 the two instances of +.sel weren't originally at the same type.
202 * There are quite a few variations here. For example, the defn of
203 +.sel could be floated ouside the \y, to attempt to gain laziness.
204 It certainly mustn't be floated outside the \d because the d has to
207 * We don't want to inline f_rhs in this case, because
208 that will duplicate code. Just commoning up the call is the point.
210 * Nothing gets added to +.sel's IdInfo.
212 * Don't bother unless the equivalence class has more than one item!
214 Not clear whether this is all worth it. It is of course OK to
215 simply discard call-instances when passing a big lambda.
217 Polymorphism 2 -- Overloading
219 Consider a function whose most general type is
221 f :: forall a b. Ord a => [a] -> b -> b
223 There is really no point in making a version of g at Int/Int and another
224 at Int/Bool, because it's only instancing the type variable "a" which
225 buys us any efficiency. Since g is completely polymorphic in b there
226 ain't much point in making separate versions of g for the different
229 That suggests that we should identify which of g's type variables
230 are constrained (like "a") and which are unconstrained (like "b").
231 Then when taking equivalence classes in STEP 2, we ignore the type args
232 corresponding to unconstrained type variable. In STEP 3 we make
233 polymorphic versions. Thus:
235 f@t1/ = /\b -> <f_rhs> t1 b d1 d2
244 f a (d::Num a) = let g = ...
246 ...(let d1::Ord a = Num.Ord.sel a d in g a d1)...
248 Here, g is only called at one type, but the dictionary isn't in scope at the
249 definition point for g. Usually the type checker would build a
250 definition for d1 which enclosed g, but the transformation system
251 might have moved d1's defn inward. Solution: float dictionary bindings
252 outwards along with call instances.
256 f x = let g p q = p==q
262 Before specialisation, leaving out type abstractions we have
264 f df x = let g :: Eq a => a -> a -> Bool
266 h :: Num a => a -> a -> (a, Bool)
267 h dh r s = let deq = eqFromNum dh
268 in (+ dh r s, g deq r s)
272 After specialising h we get a specialised version of h, like this:
274 h' r s = let deq = eqFromNum df
275 in (+ df r s, g deq r s)
277 But we can't naively make an instance for g from this, because deq is not in scope
278 at the defn of g. Instead, we have to float out the (new) defn of deq
279 to widen its scope. Notice that this floating can't be done in advance -- it only
280 shows up when specialisation is done.
282 User SPECIALIZE pragmas
283 ~~~~~~~~~~~~~~~~~~~~~~~
284 Specialisation pragmas can be digested by the type checker, and implemented
285 by adding extra definitions along with that of f, in the same way as before
287 f@t1/t2 = <f_rhs> t1 t2 d1 d2
289 Indeed the pragmas *have* to be dealt with by the type checker, because
290 only it knows how to build the dictionaries d1 and d2! For example
292 g :: Ord a => [a] -> [a]
293 {-# SPECIALIZE f :: [Tree Int] -> [Tree Int] #-}
295 Here, the specialised version of g is an application of g's rhs to the
296 Ord dictionary for (Tree Int), which only the type checker can conjure
297 up. There might not even *be* one, if (Tree Int) is not an instance of
298 Ord! (All the other specialision has suitable dictionaries to hand
301 Problem. The type checker doesn't have to hand a convenient <f_rhs>, because
302 it is buried in a complex (as-yet-un-desugared) binding group.
305 f@t1/t2 = f* t1 t2 d1 d2
307 where f* is the Id f with an IdInfo which says "inline me regardless!".
308 Indeed all the specialisation could be done in this way.
309 That in turn means that the simplifier has to be prepared to inline absolutely
310 any in-scope let-bound thing.
313 Again, the pragma should permit polymorphism in unconstrained variables:
315 h :: Ord a => [a] -> b -> b
316 {-# SPECIALIZE h :: [Int] -> b -> b #-}
318 We *insist* that all overloaded type variables are specialised to ground types,
319 (and hence there can be no context inside a SPECIALIZE pragma).
320 We *permit* unconstrained type variables to be specialised to
322 - or left as a polymorphic type variable
323 but nothing in between. So
325 {-# SPECIALIZE h :: [Int] -> [c] -> [c] #-}
327 is *illegal*. (It can be handled, but it adds complication, and gains the
331 SPECIALISING INSTANCE DECLARATIONS
332 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
335 instance Foo a => Foo [a] where
337 {-# SPECIALIZE instance Foo [Int] #-}
339 The original instance decl creates a dictionary-function
342 dfun.Foo.List :: forall a. Foo a -> Foo [a]
344 The SPECIALIZE pragma just makes a specialised copy, just as for
345 ordinary function definitions:
347 dfun.Foo.List@Int :: Foo [Int]
348 dfun.Foo.List@Int = dfun.Foo.List Int dFooInt
350 The information about what instance of the dfun exist gets added to
351 the dfun's IdInfo in the same way as a user-defined function too.
354 Automatic instance decl specialisation?
355 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
356 Can instance decls be specialised automatically? It's tricky.
357 We could collect call-instance information for each dfun, but
358 then when we specialised their bodies we'd get new call-instances
359 for ordinary functions; and when we specialised their bodies, we might get
360 new call-instances of the dfuns, and so on. This all arises because of
361 the unrestricted mutual recursion between instance decls and value decls.
363 Still, there's no actual problem; it just means that we may not do all
364 the specialisation we could theoretically do.
366 Furthermore, instance decls are usually exported and used non-locally,
367 so we'll want to compile enough to get those specialisations done.
369 Lastly, there's no such thing as a local instance decl, so we can
370 survive solely by spitting out *usage* information, and then reading that
371 back in as a pragma when next compiling the file. So for now,
372 we only specialise instance decls in response to pragmas.
375 SPITTING OUT USAGE INFORMATION
376 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
378 To spit out usage information we need to traverse the code collecting
379 call-instance information for all imported (non-prelude?) functions
380 and data types. Then we equivalence-class it and spit it out.
382 This is done at the top-level when all the call instances which escape
383 must be for imported functions and data types.
385 *** Not currently done ***
388 Partial specialisation by pragmas
389 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
390 What about partial specialisation:
392 k :: (Ord a, Eq b) => [a] -> b -> b -> [a]
393 {-# SPECIALIZE k :: Eq b => [Int] -> b -> b -> [a] #-}
397 {-# SPECIALIZE k :: Eq b => [Int] -> [b] -> [b] -> [a] #-}
399 Seems quite reasonable. Similar things could be done with instance decls:
401 instance (Foo a, Foo b) => Foo (a,b) where
403 {-# SPECIALIZE instance Foo a => Foo (a,Int) #-}
404 {-# SPECIALIZE instance Foo b => Foo (Int,b) #-}
406 Ho hum. Things are complex enough without this. I pass.
409 Requirements for the simplifer
410 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
411 The simplifier has to be able to take advantage of the specialisation.
413 * When the simplifier finds an application of a polymorphic f, it looks in
414 f's IdInfo in case there is a suitable instance to call instead. This converts
416 f t1 t2 d1 d2 ===> f_t1_t2
418 Note that the dictionaries get eaten up too!
420 * Dictionary selection operations on constant dictionaries must be
423 +.sel Int d ===> +Int
425 The obvious way to do this is in the same way as other specialised
426 calls: +.sel has inside it some IdInfo which tells that if it's applied
427 to the type Int then it should eat a dictionary and transform to +Int.
429 In short, dictionary selectors need IdInfo inside them for constant
432 * Exactly the same applies if a superclass dictionary is being
435 Eq.sel Int d ===> dEqInt
437 * Something similar applies to dictionary construction too. Suppose
438 dfun.Eq.List is the function taking a dictionary for (Eq a) to
439 one for (Eq [a]). Then we want
441 dfun.Eq.List Int d ===> dEq.List_Int
443 Where does the Eq [Int] dictionary come from? It is built in
444 response to a SPECIALIZE pragma on the Eq [a] instance decl.
446 In short, dfun Ids need IdInfo with a specialisation for each
447 constant instance of their instance declaration.
449 All this uses a single mechanism: the SpecEnv inside an Id
452 What does the specialisation IdInfo look like?
453 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
455 The SpecEnv of an Id maps a list of types (the template) to an expression
459 For example, if f has this SpecInfo:
461 [Int, a] -> \d:Ord Int. f' a
463 it means that we can replace the call
465 f Int t ===> (\d. f' t)
467 This chucks one dictionary away and proceeds with the
468 specialised version of f, namely f'.
471 What can't be done this way?
472 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
473 There is no way, post-typechecker, to get a dictionary for (say)
474 Eq a from a dictionary for Eq [a]. So if we find
478 we can't transform to
483 eqList :: (a->a->Bool) -> [a] -> [a] -> Bool
485 Of course, we currently have no way to automatically derive
486 eqList, nor to connect it to the Eq [a] instance decl, but you
487 can imagine that it might somehow be possible. Taking advantage
488 of this is permanently ruled out.
490 Still, this is no great hardship, because we intend to eliminate
491 overloading altogether anyway!
495 A note about non-tyvar dictionaries
496 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
497 Some Ids have types like
499 forall a,b,c. Eq a -> Ord [a] -> tau
501 This seems curious at first, because we usually only have dictionary
502 args whose types are of the form (C a) where a is a type variable.
503 But this doesn't hold for the functions arising from instance decls,
504 which sometimes get arguements with types of form (C (T a)) for some
507 Should we specialise wrt this compound-type dictionary? We used to say
509 "This is a heuristic judgement, as indeed is the fact that we
510 specialise wrt only dictionaries. We choose *not* to specialise
511 wrt compound dictionaries because at the moment the only place
512 they show up is in instance decls, where they are simply plugged
513 into a returned dictionary. So nothing is gained by specialising
516 But it is simpler and more uniform to specialise wrt these dicts too;
517 and in future GHC is likely to support full fledged type signatures
519 f ;: Eq [(a,b)] => ...
522 %************************************************************************
524 \subsubsection{The new specialiser}
526 %************************************************************************
528 Our basic game plan is this. For let(rec) bound function
529 f :: (C a, D c) => (a,b,c,d) -> Bool
531 * Find any specialised calls of f, (f ts ds), where
532 ts are the type arguments t1 .. t4, and
533 ds are the dictionary arguments d1 .. d2.
535 * Add a new definition for f1 (say):
537 f1 = /\ b d -> (..body of f..) t1 b t3 d d1 d2
539 Note that we abstract over the unconstrained type arguments.
543 [t1,b,t3,d] |-> \d1 d2 -> f1 b d
545 to the specialisations of f. This will be used by the
546 simplifier to replace calls
547 (f t1 t2 t3 t4) da db
549 (\d1 d1 -> f1 t2 t4) da db
551 All the stuff about how many dictionaries to discard, and what types
552 to apply the specialised function to, are handled by the fact that the
553 SpecEnv contains a template for the result of the specialisation.
555 We don't build *partial* specialisations for f. For example:
557 f :: Eq a => a -> a -> Bool
558 {-# SPECIALISE f :: (Eq b, Eq c) => (b,c) -> (b,c) -> Bool #-}
560 Here, little is gained by making a specialised copy of f.
561 There's a distinct danger that the specialised version would
562 first build a dictionary for (Eq b, Eq c), and then select the (==)
563 method from it! Even if it didn't, not a great deal is saved.
565 We do, however, generate polymorphic, but not overloaded, specialisations:
567 f :: Eq a => [a] -> b -> b -> b
568 {#- SPECIALISE f :: [Int] -> b -> b -> b #-}
570 Hence, the invariant is this:
572 *** no specialised version is overloaded ***
575 %************************************************************************
577 \subsubsection{The exported function}
579 %************************************************************************
582 specProgram :: DynFlags -> UniqSupply -> [CoreBind] -> IO [CoreBind]
583 specProgram dflags us binds
585 beginPass dflags "Specialise"
587 let binds' = initSM us (go binds `thenSM` \ (binds', uds') ->
588 returnSM (dumpAllDictBinds uds' binds'))
590 endPass dflags "Specialise"
591 (dopt Opt_D_dump_spec dflags
592 || dopt Opt_D_verbose_core2core dflags) binds'
594 dumpIfSet_dyn dflags Opt_D_dump_rules "Top-level specialisations"
595 (vcat (map dump_specs (concat (map bindersOf binds'))))
599 -- We need to start with a Subst that knows all the things
600 -- that are in scope, so that the substitution engine doesn't
601 -- accidentally re-use a unique that's already in use
602 -- Easiest thing is to do it all at once, as if all the top-level
603 -- decls were mutually recursive
604 top_subst = mkSubst (mkInScopeSet (mkVarSet (bindersOfBinds binds))) emptySubstEnv
606 go [] = returnSM ([], emptyUDs)
607 go (bind:binds) = go binds `thenSM` \ (binds', uds) ->
608 specBind top_subst bind uds `thenSM` \ (bind', uds') ->
609 returnSM (bind' ++ binds', uds')
611 dump_specs var = pprCoreRules var (idSpecialisation var)
614 %************************************************************************
616 \subsubsection{@specExpr@: the main function}
618 %************************************************************************
621 specVar :: Subst -> Id -> CoreExpr
622 specVar subst v = case lookupIdSubst subst v of
626 specExpr :: Subst -> CoreExpr -> SpecM (CoreExpr, UsageDetails)
627 -- We carry a substitution down:
628 -- a) we must clone any binding that might flaot outwards,
629 -- to avoid name clashes
630 -- b) we carry a type substitution to use when analysing
631 -- the RHS of specialised bindings (no type-let!)
633 ---------------- First the easy cases --------------------
634 specExpr subst (Type ty) = returnSM (Type (substTy subst ty), emptyUDs)
635 specExpr subst (Var v) = returnSM (specVar subst v, emptyUDs)
636 specExpr subst (Lit lit) = returnSM (Lit lit, emptyUDs)
638 specExpr subst (Note note body)
639 = specExpr subst body `thenSM` \ (body', uds) ->
640 returnSM (Note (specNote subst note) body', uds)
643 ---------------- Applications might generate a call instance --------------------
644 specExpr subst expr@(App fun arg)
647 go (App fun arg) args = specExpr subst arg `thenSM` \ (arg', uds_arg) ->
648 go fun (arg':args) `thenSM` \ (fun', uds_app) ->
649 returnSM (App fun' arg', uds_arg `plusUDs` uds_app)
651 go (Var f) args = case specVar subst f of
652 Var f' -> returnSM (Var f', mkCallUDs subst f' args)
653 e' -> returnSM (e', emptyUDs) -- I don't expect this!
654 go other args = specExpr subst other
656 ---------------- Lambda/case require dumping of usage details --------------------
657 specExpr subst e@(Lam _ _)
658 = specExpr subst' body `thenSM` \ (body', uds) ->
660 (filtered_uds, body'') = dumpUDs bndrs' uds body'
662 returnSM (mkLams bndrs' body'', filtered_uds)
664 (bndrs, body) = collectBinders e
665 (subst', bndrs') = substBndrs subst bndrs
666 -- More efficient to collect a group of binders together all at once
667 -- and we don't want to split a lambda group with dumped bindings
669 specExpr subst (Case scrut case_bndr alts)
670 = specExpr subst scrut `thenSM` \ (scrut', uds_scrut) ->
671 mapAndCombineSM spec_alt alts `thenSM` \ (alts', uds_alts) ->
672 returnSM (Case scrut' case_bndr' alts', uds_scrut `plusUDs` uds_alts)
674 (subst_alt, case_bndr') = substId subst case_bndr
675 -- No need to clone case binder; it can't float like a let(rec)
677 spec_alt (con, args, rhs)
678 = specExpr subst_rhs rhs `thenSM` \ (rhs', uds) ->
680 (uds', rhs'') = dumpUDs args uds rhs'
682 returnSM ((con, args', rhs''), uds')
684 (subst_rhs, args') = substBndrs subst_alt args
686 ---------------- Finally, let is the interesting case --------------------
687 specExpr subst (Let bind body)
689 cloneBindSM subst bind `thenSM` \ (rhs_subst, body_subst, bind') ->
691 -- Deal with the body
692 specExpr body_subst body `thenSM` \ (body', body_uds) ->
694 -- Deal with the bindings
695 specBind rhs_subst bind' body_uds `thenSM` \ (binds', uds) ->
698 returnSM (foldr Let body' binds', uds)
700 -- Must apply the type substitution to coerceions
701 specNote subst (Coerce t1 t2) = Coerce (substTy subst t1) (substTy subst t2)
702 specNote subst note = note
705 %************************************************************************
707 \subsubsection{Dealing with a binding}
709 %************************************************************************
712 specBind :: Subst -- Use this for RHSs
714 -> UsageDetails -- Info on how the scope of the binding
715 -> SpecM ([CoreBind], -- New bindings
716 UsageDetails) -- And info to pass upstream
718 specBind rhs_subst bind body_uds
719 = specBindItself rhs_subst bind (calls body_uds) `thenSM` \ (bind', bind_uds) ->
721 bndrs = bindersOf bind
722 all_uds = zapCalls bndrs (body_uds `plusUDs` bind_uds)
723 -- It's important that the `plusUDs` is this way round,
724 -- because body_uds may bind dictionaries that are
725 -- used in the calls passed to specDefn. So the
726 -- dictionary bindings in bind_uds may mention
727 -- dictionaries bound in body_uds.
729 case splitUDs bndrs all_uds of
731 (_, ([],[])) -- This binding doesn't bind anything needed
732 -- in the UDs, so put the binding here
733 -- This is the case for most non-dict bindings, except
734 -- for the few that are mentioned in a dict binding
735 -- that is floating upwards in body_uds
736 -> returnSM ([bind'], all_uds)
738 (float_uds, (dict_binds, calls)) -- This binding is needed in the UDs, so float it out
739 -> returnSM ([], float_uds `plusUDs` mkBigUD bind' dict_binds calls)
742 -- A truly gruesome function
743 mkBigUD bind@(NonRec _ _) dbs calls
744 = -- Common case: non-recursive and no specialisations
745 -- (if there were any specialistions it would have been made recursive)
746 MkUD { dict_binds = listToBag (mkDB bind : dbs),
747 calls = listToCallDetails calls }
749 mkBigUD bind dbs calls
751 MkUD { dict_binds = unitBag (mkDB (Rec (bind_prs bind ++ dbsToPairs dbs))),
753 calls = listToCallDetails calls }
755 bind_prs (NonRec b r) = [(b,r)]
756 bind_prs (Rec prs) = prs
759 dbsToPairs ((bind,_):dbs) = bind_prs bind ++ dbsToPairs dbs
761 -- specBindItself deals with the RHS, specialising it according
762 -- to the calls found in the body (if any)
763 specBindItself rhs_subst (NonRec bndr rhs) call_info
764 = specDefn rhs_subst call_info (bndr,rhs) `thenSM` \ ((bndr',rhs'), spec_defns, spec_uds) ->
766 new_bind | null spec_defns = NonRec bndr' rhs'
767 | otherwise = Rec ((bndr',rhs'):spec_defns)
768 -- bndr' mentions the spec_defns in its SpecEnv
769 -- Not sure why we couln't just put the spec_defns first
771 returnSM (new_bind, spec_uds)
773 specBindItself rhs_subst (Rec pairs) call_info
774 = mapSM (specDefn rhs_subst call_info) pairs `thenSM` \ stuff ->
776 (pairs', spec_defns_s, spec_uds_s) = unzip3 stuff
777 spec_defns = concat spec_defns_s
778 spec_uds = plusUDList spec_uds_s
779 new_bind = Rec (spec_defns ++ pairs')
781 returnSM (new_bind, spec_uds)
784 specDefn :: Subst -- Subst to use for RHS
785 -> CallDetails -- Info on how it is used in its scope
786 -> (Id, CoreExpr) -- The thing being bound and its un-processed RHS
787 -> SpecM ((Id, CoreExpr), -- The thing and its processed RHS
788 -- the Id may now have specialisations attached
789 [(Id,CoreExpr)], -- Extra, specialised bindings
790 UsageDetails -- Stuff to fling upwards from the RHS and its
791 ) -- specialised versions
793 specDefn subst calls (fn, rhs)
794 -- The first case is the interesting one
795 | n_tyvars == length rhs_tyvars -- Rhs of fn's defn has right number of big lambdas
796 && n_dicts <= length rhs_bndrs -- and enough dict args
797 && not (null calls_for_me) -- And there are some calls to specialise
798 && not (certainlyWillInline fn) -- And it's not small
799 -- If it's small, it's better just to inline
800 -- it than to construct lots of specialisations
801 = -- Specialise the body of the function
802 specExpr subst rhs `thenSM` \ (rhs', rhs_uds) ->
804 -- Make a specialised version for each call in calls_for_me
805 mapSM spec_call calls_for_me `thenSM` \ stuff ->
807 (spec_defns, spec_uds, spec_env_stuff) = unzip3 stuff
809 fn' = addIdSpecialisations zapped_fn spec_env_stuff
811 returnSM ((fn',rhs'),
813 rhs_uds `plusUDs` plusUDList spec_uds)
815 | otherwise -- No calls or RHS doesn't fit our preconceptions
816 = specExpr subst rhs `thenSM` \ (rhs', rhs_uds) ->
817 returnSM ((zapped_fn, rhs'), [], rhs_uds)
820 zapped_fn = modifyIdInfo zapSpecPragInfo fn
821 -- If the fn is a SpecPragmaId, make it discardable
822 -- It's role as a holder for a call instance is o'er
823 -- But it might be alive for some other reason by now.
826 (tyvars, theta, _) = splitSigmaTy fn_type
827 n_tyvars = length tyvars
828 n_dicts = length theta
830 (rhs_tyvars, rhs_ids, rhs_body) = collectTyAndValBinders rhs
831 rhs_dicts = take n_dicts rhs_ids
832 rhs_bndrs = rhs_tyvars ++ rhs_dicts
833 body = mkLams (drop n_dicts rhs_ids) rhs_body
834 -- Glue back on the non-dict lambdas
836 calls_for_me = case lookupFM calls fn of
838 Just cs -> fmToList cs
840 ----------------------------------------------------------
841 -- Specialise to one particular call pattern
842 spec_call :: ([Maybe Type], ([DictExpr], VarSet)) -- Call instance
843 -> SpecM ((Id,CoreExpr), -- Specialised definition
844 UsageDetails, -- Usage details from specialised body
845 ([CoreBndr], [CoreExpr], CoreExpr)) -- Info for the Id's SpecEnv
846 spec_call (call_ts, (call_ds, call_fvs))
847 = ASSERT( length call_ts == n_tyvars && length call_ds == n_dicts )
848 -- Calls are only recorded for properly-saturated applications
850 -- Suppose f's defn is f = /\ a b c d -> \ d1 d2 -> rhs
851 -- Supppose the call is for f [Just t1, Nothing, Just t3, Nothing] [dx1, dx2]
853 -- Construct the new binding
854 -- f1 = SUBST[a->t1,c->t3, d1->d1', d2->d2'] (/\ b d -> rhs)
855 -- PLUS the usage-details
856 -- { d1' = dx1; d2' = dx2 }
857 -- where d1', d2' are cloned versions of d1,d2, with the type substitution applied.
859 -- Note that the substitution is applied to the whole thing.
860 -- This is convenient, but just slightly fragile. Notably:
861 -- * There had better be no name clashes in a/b/c/d
864 -- poly_tyvars = [b,d] in the example above
865 -- spec_tyvars = [a,c]
866 -- ty_args = [t1,b,t3,d]
867 poly_tyvars = [tv | (tv, Nothing) <- rhs_tyvars `zip` call_ts]
868 spec_tyvars = [tv | (tv, Just _) <- rhs_tyvars `zip` call_ts]
869 ty_args = zipWithEqual "spec_call" mk_ty_arg rhs_tyvars call_ts
871 mk_ty_arg rhs_tyvar Nothing = Type (mkTyVarTy rhs_tyvar)
872 mk_ty_arg rhs_tyvar (Just ty) = Type ty
873 rhs_subst = extendSubstList subst spec_tyvars [DoneTy ty | Just ty <- call_ts]
875 cloneBinders rhs_subst rhs_dicts `thenSM` \ (rhs_subst', rhs_dicts') ->
877 inst_args = ty_args ++ map Var rhs_dicts'
879 -- Figure out the type of the specialised function
880 spec_id_ty = mkForAllTys poly_tyvars (applyTypeToArgs rhs fn_type inst_args)
882 newIdSM fn spec_id_ty `thenSM` \ spec_f ->
883 specExpr rhs_subst' (mkLams poly_tyvars body) `thenSM` \ (spec_rhs, rhs_uds) ->
885 -- The rule to put in the function's specialisation is:
886 -- forall b,d, d1',d2'. f t1 b t3 d d1' d2' = f1 b d
887 spec_env_rule = (poly_tyvars ++ rhs_dicts',
889 mkTyApps (Var spec_f) (map mkTyVarTy poly_tyvars))
891 -- Add the { d1' = dx1; d2' = dx2 } usage stuff
892 final_uds = foldr addDictBind rhs_uds (my_zipEqual "spec_call" rhs_dicts' call_ds)
894 returnSM ((spec_f, spec_rhs),
899 my_zipEqual doc xs ys
900 | length xs /= length ys = pprPanic "my_zipEqual" (ppr xs $$ ppr ys $$ (ppr fn <+> ppr call_ts) $$ ppr rhs)
901 | otherwise = zipEqual doc xs ys
904 %************************************************************************
906 \subsubsection{UsageDetails and suchlike}
908 %************************************************************************
913 dict_binds :: !(Bag DictBind),
914 -- Floated dictionary bindings
915 -- The order is important;
916 -- in ds1 `union` ds2, bindings in ds2 can depend on those in ds1
917 -- (Remember, Bags preserve order in GHC.)
919 calls :: !CallDetails
922 type DictBind = (CoreBind, VarSet)
923 -- The set is the free vars of the binding
924 -- both tyvars and dicts
926 type DictExpr = CoreExpr
928 emptyUDs = MkUD { dict_binds = emptyBag, calls = emptyFM }
930 type ProtoUsageDetails = ([DictBind],
931 [(Id, [Maybe Type], ([DictExpr], VarSet))]
934 ------------------------------------------------------------
935 type CallDetails = FiniteMap Id CallInfo
936 type CallInfo = FiniteMap [Maybe Type] -- Nothing => unconstrained type argument
937 ([DictExpr], VarSet) -- Dict args and the vars of the whole
938 -- call (including tyvars)
939 -- [*not* include the main id itself, of course]
940 -- The finite maps eliminate duplicates
941 -- The list of types and dictionaries is guaranteed to
942 -- match the type of f
944 unionCalls :: CallDetails -> CallDetails -> CallDetails
945 unionCalls c1 c2 = plusFM_C plusFM c1 c2
947 singleCall :: Id -> [Maybe Type] -> [DictExpr] -> CallDetails
948 singleCall id tys dicts
949 = unitFM id (unitFM tys (dicts, call_fvs))
951 call_fvs = exprsFreeVars dicts `unionVarSet` tys_fvs
952 tys_fvs = tyVarsOfTypes (catMaybes tys)
953 -- The type args (tys) are guaranteed to be part of the dictionary
954 -- types, because they are just the constrained types,
955 -- and the dictionary is therefore sure to be bound
956 -- inside the binding for any type variables free in the type;
957 -- hence it's safe to neglect tyvars free in tys when making
958 -- the free-var set for this call
959 -- BUT I don't trust this reasoning; play safe and include tys_fvs
961 -- We don't include the 'id' itself.
963 listToCallDetails calls
964 = foldr (unionCalls . mk_call) emptyFM calls
966 mk_call (id, tys, dicts_w_fvs) = unitFM id (unitFM tys dicts_w_fvs)
967 -- NB: the free vars of the call are provided
969 callDetailsToList calls = [ (id,tys,dicts)
970 | (id,fm) <- fmToList calls,
971 (tys,dicts) <- fmToList fm
974 mkCallUDs subst f args
976 || length spec_tys /= n_tyvars
977 || length dicts /= n_dicts
978 || maybeToBool (lookupRule (substInScope subst) f args)
979 -- There's already a rule covering this call. A typical case
980 -- is where there's an explicit user-provided rule. Then
981 -- we don't want to create a specialised version
982 -- of the function that overlaps.
983 = emptyUDs -- Not overloaded, or no specialisation wanted
986 = MkUD {dict_binds = emptyBag,
987 calls = singleCall f spec_tys dicts
990 (tyvars, theta, _) = splitSigmaTy (idType f)
991 constrained_tyvars = tyVarsOfTheta theta
992 n_tyvars = length tyvars
993 n_dicts = length theta
995 spec_tys = [mk_spec_ty tv ty | (tv, Type ty) <- tyvars `zip` args]
996 dicts = [dict_expr | (_, dict_expr) <- theta `zip` (drop n_tyvars args)]
998 mk_spec_ty tyvar ty | tyvar `elemVarSet` constrained_tyvars
1003 ------------------------------------------------------------
1004 plusUDs :: UsageDetails -> UsageDetails -> UsageDetails
1005 plusUDs (MkUD {dict_binds = db1, calls = calls1})
1006 (MkUD {dict_binds = db2, calls = calls2})
1007 = MkUD {dict_binds = d, calls = c}
1009 d = db1 `unionBags` db2
1010 c = calls1 `unionCalls` calls2
1012 plusUDList = foldr plusUDs emptyUDs
1014 -- zapCalls deletes calls to ids from uds
1015 zapCalls ids uds = uds {calls = delListFromFM (calls uds) ids}
1017 mkDB bind = (bind, bind_fvs bind)
1019 bind_fvs (NonRec bndr rhs) = exprFreeVars rhs
1020 bind_fvs (Rec prs) = foldl delVarSet rhs_fvs bndrs
1023 rhs_fvs = unionVarSets [exprFreeVars rhs | (bndr,rhs) <- prs]
1025 addDictBind (dict,rhs) uds = uds { dict_binds = mkDB (NonRec dict rhs) `consBag` dict_binds uds }
1027 dumpAllDictBinds (MkUD {dict_binds = dbs}) binds
1028 = foldrBag add binds dbs
1030 add (bind,_) binds = bind : binds
1032 dumpUDs :: [CoreBndr]
1033 -> UsageDetails -> CoreExpr
1034 -> (UsageDetails, CoreExpr)
1035 dumpUDs bndrs uds body
1036 = (free_uds, foldr add_let body dict_binds)
1038 (free_uds, (dict_binds, _)) = splitUDs bndrs uds
1039 add_let (bind,_) body = Let bind body
1041 splitUDs :: [CoreBndr]
1043 -> (UsageDetails, -- These don't mention the binders
1044 ProtoUsageDetails) -- These do
1046 splitUDs bndrs uds@(MkUD {dict_binds = orig_dbs,
1047 calls = orig_calls})
1049 = if isEmptyBag dump_dbs && null dump_calls then
1050 -- Common case: binder doesn't affect floats
1054 -- Binders bind some of the fvs of the floats
1055 (MkUD {dict_binds = free_dbs,
1056 calls = listToCallDetails free_calls},
1057 (bagToList dump_dbs, dump_calls)
1061 bndr_set = mkVarSet bndrs
1063 (free_dbs, dump_dbs, dump_idset)
1064 = foldlBag dump_db (emptyBag, emptyBag, bndr_set) orig_dbs
1065 -- Important that it's foldl not foldr;
1066 -- we're accumulating the set of dumped ids in dump_set
1068 -- Filter out any calls that mention things that are being dumped
1069 orig_call_list = callDetailsToList orig_calls
1070 (dump_calls, free_calls) = partition captured orig_call_list
1071 captured (id,tys,(dicts, fvs)) = fvs `intersectsVarSet` dump_idset
1072 || id `elemVarSet` dump_idset
1074 dump_db (free_dbs, dump_dbs, dump_idset) db@(bind, fvs)
1075 | dump_idset `intersectsVarSet` fvs -- Dump it
1076 = (free_dbs, dump_dbs `snocBag` db,
1077 dump_idset `unionVarSet` mkVarSet (bindersOf bind))
1079 | otherwise -- Don't dump it
1080 = (free_dbs `snocBag` db, dump_dbs, dump_idset)
1084 %************************************************************************
1086 \subsubsection{Boring helper functions}
1088 %************************************************************************
1091 lookupId:: IdEnv Id -> Id -> Id
1092 lookupId env id = case lookupVarEnv env id of
1096 ----------------------------------------
1097 type SpecM a = UniqSM a
1102 getUniqSM = getUniqueUs
1103 getUniqSupplySM = getUs
1104 setUniqSupplySM = setUs
1108 mapAndCombineSM f [] = returnSM ([], emptyUDs)
1109 mapAndCombineSM f (x:xs) = f x `thenSM` \ (y, uds1) ->
1110 mapAndCombineSM f xs `thenSM` \ (ys, uds2) ->
1111 returnSM (y:ys, uds1 `plusUDs` uds2)
1113 cloneBindSM :: Subst -> CoreBind -> SpecM (Subst, Subst, CoreBind)
1114 -- Clone the binders of the bind; return new bind with the cloned binders
1115 -- Return the substitution to use for RHSs, and the one to use for the body
1116 cloneBindSM subst (NonRec bndr rhs)
1117 = getUs `thenUs` \ us ->
1119 (subst', us', bndr') = substAndCloneId subst us bndr
1122 returnUs (subst, subst', NonRec bndr' rhs)
1124 cloneBindSM subst (Rec pairs)
1125 = getUs `thenUs` \ us ->
1127 (subst', us', bndrs') = substAndCloneIds subst us (map fst pairs)
1130 returnUs (subst', subst', Rec (bndrs' `zip` map snd pairs))
1132 cloneBinders subst bndrs
1133 = getUs `thenUs` \ us ->
1135 (subst', us', bndrs') = substAndCloneIds subst us bndrs
1138 returnUs (subst', bndrs')
1141 newIdSM old_id new_ty
1142 = getUniqSM `thenSM` \ uniq ->
1144 -- Give the new Id a similar occurrence name to the old one
1145 -- We used to add setIdNoDiscard if the old id was exported, to
1146 -- avoid it being dropped as dead code, but that's not necessary any more.
1147 name = idName old_id
1148 new_id = mkUserLocal (mkSpecOcc (nameOccName name)) uniq new_ty (getSrcLoc name)
1153 = getUniqSM `thenSM` \ uniq ->
1154 returnSM (mkSysTyVar uniq boxedTypeKind)
1158 Old (but interesting) stuff about unboxed bindings
1159 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1161 What should we do when a value is specialised to a *strict* unboxed value?
1163 map_*_* f (x:xs) = let h = f x
1167 Could convert let to case:
1169 map_*_Int# f (x:xs) = case f x of h# ->
1173 This may be undesirable since it forces evaluation here, but the value
1174 may not be used in all branches of the body. In the general case this
1175 transformation is impossible since the mutual recursion in a letrec
1176 cannot be expressed as a case.
1178 There is also a problem with top-level unboxed values, since our
1179 implementation cannot handle unboxed values at the top level.
1181 Solution: Lift the binding of the unboxed value and extract it when it
1184 map_*_Int# f (x:xs) = let h = case (f x) of h# -> _Lift h#
1189 Now give it to the simplifier and the _Lifting will be optimised away.
1191 The benfit is that we have given the specialised "unboxed" values a
1192 very simplep lifted semantics and then leave it up to the simplifier to
1193 optimise it --- knowing that the overheads will be removed in nearly
1196 In particular, the value will only be evaluted in the branches of the
1197 program which use it, rather than being forced at the point where the
1198 value is bound. For example:
1200 filtermap_*_* p f (x:xs)
1207 filtermap_*_Int# p f (x:xs)
1208 = let h = case (f x) of h# -> _Lift h#
1211 True -> case h of _Lift h#
1215 The binding for h can still be inlined in the one branch and the
1216 _Lifting eliminated.
1219 Question: When won't the _Lifting be eliminated?
1221 Answer: When they at the top-level (where it is necessary) or when
1222 inlining would duplicate work (or possibly code depending on
1223 options). However, the _Lifting will still be eliminated if the
1224 strictness analyser deems the lifted binding strict.