X-Git-Url: http://git.megacz.com/?a=blobdiff_plain;f=ghc%2Fcompiler%2FcoreSyn%2FCoreUnfold.lhs;h=69f7150a4d37d5105d520b03d7697aa85c5ad161;hb=f1080bc82f87317ffa59cffef08b322d7354bb29;hp=b59e9cf140189d7772f0b2fe7590105b4613de6e;hpb=438596897ebbe25a07e1c82085cfbc5bdb00f09e;p=ghc-hetmet.git diff --git a/ghc/compiler/coreSyn/CoreUnfold.lhs b/ghc/compiler/coreSyn/CoreUnfold.lhs index b59e9cf..69f7150 100644 --- a/ghc/compiler/coreSyn/CoreUnfold.lhs +++ b/ghc/compiler/coreSyn/CoreUnfold.lhs @@ -14,186 +14,165 @@ find, unsurprisingly, a Core expression. \begin{code} module CoreUnfold ( - Unfolding(..), UnfoldingGuidance(..), -- types + Unfolding, UnfoldingGuidance, -- Abstract types - noUnfolding, mkMagicUnfolding, mkUnfolding, getUnfoldingTemplate, - isEvaldUnfolding, hasUnfolding, + noUnfolding, mkTopUnfolding, mkUnfolding, mkCompulsoryUnfolding, seqUnfolding, + mkOtherCon, otherCons, + unfoldingTemplate, maybeUnfoldingTemplate, + isEvaldUnfolding, isValueUnfolding, isCheapUnfolding, isCompulsoryUnfolding, + hasUnfolding, hasSomeUnfolding, - smallEnoughToInline, couldBeSmallEnoughToInline, - certainlySmallEnoughToInline, + couldBeSmallEnoughToInline, + certainlyWillInline, okToUnfoldInHiFile, - calcUnfoldingGuidance + callSiteInline, blackListed ) where #include "HsVersions.h" -import {-# SOURCE #-} MagicUFs ( MagicUnfoldingFun, mkMagicUnfoldingFun ) - -import CmdLineOpts ( opt_UnfoldingCreationThreshold, - opt_UnfoldingUseThreshold, - opt_UnfoldingConDiscount, - opt_UnfoldingKeenessFactor, - opt_UnfoldCasms - ) -import Constants ( uNFOLDING_CHEAP_OP_COST, - uNFOLDING_DEAR_OP_COST, - uNFOLDING_NOREP_LIT_COST +import CmdLineOpts ( opt_UF_CreationThreshold, + opt_UF_UseThreshold, + opt_UF_FunAppDiscount, + opt_UF_KeenessFactor, + opt_UF_DearOp, opt_UnfoldCasms, + DynFlags, dopt_D_dump_inlinings ) import CoreSyn +import PprCore ( pprCoreExpr ) import OccurAnal ( occurAnalyseGlobalExpr ) -import CoreUtils ( coreExprType, exprIsTrivial, mkFormSummary, - FormSummary(..) ) -import Id ( Id, idType, isId ) -import Const ( Con(..), isLitLitLit ) -import PrimOp ( PrimOp(..), primOpOutOfLine ) -import IdInfo ( ArityInfo(..), InlinePragInfo(..) ) -import TyCon ( tyConFamilySize ) -import Type ( splitAlgTyConApp_maybe ) -import Const ( isNoRepLit ) -import Unique ( Unique ) -import Util ( isIn, panic ) +import CoreUtils ( exprIsValue, exprIsCheap, exprIsTrivial ) +import Id ( Id, idType, idFlavour, isId, + idSpecialisation, idInlinePragma, idUnfolding, + isPrimOpId_maybe + ) +import VarSet +import Literal ( isLitLitLit, litIsDupable ) +import PrimOp ( PrimOp(..), primOpIsDupable, primOpOutOfLine, ccallIsCasm ) +import IdInfo ( InlinePragInfo(..), OccInfo(..), IdFlavour(..), + isNeverInlinePrag + ) +import Type ( isUnLiftedType ) +import PrelNames ( hasKey, buildIdKey, augmentIdKey ) +import Bag +import FastTypes import Outputable + +#if __GLASGOW_HASKELL__ >= 404 +import GlaExts ( fromInt ) +#endif \end{code} + %************************************************************************ %* * -\subsection{@Unfolding@ and @UnfoldingGuidance@ types} +\subsection{Making unfoldings} %* * %************************************************************************ \begin{code} -data Unfolding - = NoUnfolding - - | OtherCon [Con] -- It ain't one of these - -- (OtherCon xs) also indicates that something has been evaluated - -- and hence there's no point in re-evaluating it. - -- OtherCon [] is used even for non-data-type values - -- to indicated evaluated-ness. Notably: - -- data C = C !(Int -> Int) - -- case x of { C f -> ... } - -- Here, f gets an OtherCon [] unfolding. - - | CoreUnfolding -- An unfolding with redundant cached information - FormSummary -- Tells whether the template is a WHNF or bottom - UnfoldingGuidance -- Tells about the *size* of the template. - CoreExpr -- Template; binder-info is correct - - | MagicUnfolding - Unique -- Unique of the Id whose magic unfolding this is - MagicUnfoldingFun +mkTopUnfolding expr = mkUnfolding True {- Top level -} expr + +mkUnfolding top_lvl expr + = CoreUnfolding (occurAnalyseGlobalExpr expr) + top_lvl + (exprIsValue expr) + -- Already evaluated + + (exprIsCheap expr) + -- OK to inline inside a lambda + + (calcUnfoldingGuidance opt_UF_CreationThreshold expr) + -- Sometimes during simplification, there's a large let-bound thing + -- which has been substituted, and so is now dead; so 'expr' contains + -- two copies of the thing while the occurrence-analysed expression doesn't + -- Nevertheless, we don't occ-analyse before computing the size because the + -- size computation bales out after a while, whereas occurrence analysis does not. + -- + -- This can occasionally mean that the guidance is very pessimistic; + -- it gets fixed up next round + +mkCompulsoryUnfolding expr -- Used for things that absolutely must be unfolded + = CompulsoryUnfolding (occurAnalyseGlobalExpr expr) \end{code} -\begin{code} -noUnfolding = NoUnfolding - -mkUnfolding expr - = let - -- strictness mangling (depends on there being no CSE) - ufg = calcUnfoldingGuidance opt_UnfoldingCreationThreshold expr - occ = occurAnalyseGlobalExpr expr - in - CoreUnfolding (mkFormSummary expr) ufg occ - -mkMagicUnfolding :: Unique -> Unfolding -mkMagicUnfolding tag = MagicUnfolding tag (mkMagicUnfoldingFun tag) - -getUnfoldingTemplate :: Unfolding -> CoreExpr -getUnfoldingTemplate (CoreUnfolding _ _ expr) = expr -getUnfoldingTemplate other = panic "getUnfoldingTemplate" - -isEvaldUnfolding :: Unfolding -> Bool -isEvaldUnfolding (OtherCon _) = True -isEvaldUnfolding (CoreUnfolding ValueForm _ expr) = True -isEvaldUnfolding other = False - -hasUnfolding :: Unfolding -> Bool -hasUnfolding NoUnfolding = False -hasUnfolding other = True - -data UnfoldingGuidance - = UnfoldNever - | UnfoldAlways -- There is no "original" definition, - -- so you'd better unfold. Or: something - -- so cheap to unfold (e.g., 1#) that - -- you should do it absolutely always. - - | UnfoldIfGoodArgs Int -- if "m" type args - Int -- and "n" value args - [Int] -- Discount if the argument is evaluated. - -- (i.e., a simplification will definitely - -- be possible). One elt of the list per *value* arg. - - Int -- The "size" of the unfolding; to be elaborated - -- later. ToDo - - Int -- Scrutinee discount: the discount to substract if the thing is in - -- a context (case (thing args) of ...), - -- (where there are the right number of arguments.) -\end{code} +%************************************************************************ +%* * +\subsection{The UnfoldingGuidance type} +%* * +%************************************************************************ \begin{code} instance Outputable UnfoldingGuidance where - ppr UnfoldAlways = ptext SLIT("_ALWAYS_") - ppr (UnfoldIfGoodArgs t v cs size discount) - = hsep [ptext SLIT("_IF_ARGS_"), int t, int v, - if null cs -- always print *something* - then char 'X' - else hcat (map (text . show) cs), + ppr UnfoldNever = ptext SLIT("NEVER") + ppr (UnfoldIfGoodArgs v cs size discount) + = hsep [ ptext SLIT("IF_ARGS"), int v, + brackets (hsep (map int cs)), int size, int discount ] \end{code} -%************************************************************************ -%* * -\subsection[calcUnfoldingGuidance]{Calculate ``unfolding guidance'' for an expression} -%* * -%************************************************************************ - \begin{code} calcUnfoldingGuidance :: Int -- bomb out if size gets bigger than this -> CoreExpr -- expression to look at -> UnfoldingGuidance calcUnfoldingGuidance bOMB_OUT_SIZE expr - | exprIsTrivial expr -- Often trivial expressions are never bound - -- to an expression, but it can happen. For - -- example, the Id for a nullary constructor has - -- a trivial expression as its unfolding, and - -- we want to make sure that we always unfold it. - = UnfoldAlways - - | otherwise - = case collectTyAndValBinders expr of { (ty_binders, val_binders, body) -> + = case collect_val_bndrs expr of { (inline, val_binders, body) -> + let + n_val_binders = length val_binders + + max_inline_size = n_val_binders+2 + -- The idea is that if there is an INLINE pragma (inline is True) + -- and there's a big body, we give a size of n_val_binders+2. This + -- This is just enough to fail the no-size-increase test in callSiteInline, + -- so that INLINE things don't get inlined into entirely boring contexts, + -- but no more. + + in case (sizeExpr bOMB_OUT_SIZE val_binders body) of - TooBig -> UnfoldNever + TooBig + | not inline -> UnfoldNever + -- A big function with an INLINE pragma must + -- have an UnfoldIfGoodArgs guidance + | inline -> UnfoldIfGoodArgs n_val_binders + (map (const 0) val_binders) + max_inline_size 0 SizeIs size cased_args scrut_discount -> UnfoldIfGoodArgs - (length ty_binders) - (length val_binders) + n_val_binders (map discount_for val_binders) - (I# size) - (I# scrut_discount) + final_size + (iBox scrut_discount) where - discount_for b - | num_cases == 0 = 0 - | otherwise - = if is_data - then tyConFamilySize tycon * num_cases - else num_cases -- prim cases are pretty cheap - - where - (is_data, tycon) - = case (splitAlgTyConApp_maybe (idType b)) of - Nothing -> (False, panic "discount") - Just (tc,_,_) -> (True, tc) - num_cases = length (filter (==b) cased_args) + boxed_size = iBox size + + final_size | inline = boxed_size `min` max_inline_size + | otherwise = boxed_size + + -- Sometimes an INLINE thing is smaller than n_val_binders+2. + -- A particular case in point is a constructor, which has size 1. + -- We want to inline this regardless, hence the `min` + + discount_for b = foldlBag (\acc (b',n) -> if b==b' then acc+n else acc) + 0 cased_args } + where + collect_val_bndrs e = go False [] e + -- We need to be a bit careful about how we collect the + -- value binders. In ptic, if we see + -- __inline_me (\x y -> e) + -- We want to say "2 value binders". Why? So that + -- we take account of information given for the arguments + + go inline rev_vbs (Note InlineMe e) = go True rev_vbs e + go inline rev_vbs (Lam b e) | isId b = go inline (b:rev_vbs) e + | otherwise = go inline rev_vbs e + go inline rev_vbs e = (inline, reverse rev_vbs, e) \end{code} \begin{code} @@ -203,25 +182,29 @@ sizeExpr :: Int -- Bomb out if it gets bigger than this -> CoreExpr -> ExprSize -sizeExpr (I# bOMB_OUT_SIZE) args expr +sizeExpr bOMB_OUT_SIZE top_args expr = size_up expr where - size_up (Type t) = sizeZero -- Types cost nothing - size_up (Note _ body) = size_up body -- Notes cost nothing - size_up (Var v) = sizeOne - size_up (App fun arg) = size_up fun `addSize` size_up arg + size_up (Type t) = sizeZero -- Types cost nothing + size_up (Var v) = sizeOne + + size_up (Note _ body) = size_up body -- Notes cost nothing + + size_up (App fun (Type t)) = size_up fun + size_up (App fun arg) = size_up_app fun [arg] - size_up (Con con args) = foldr (addSize . size_up) - (size_up_con con (valArgCount args)) - args + size_up (Lit lit) | litIsDupable lit = sizeOne + | otherwise = sizeN opt_UF_DearOp -- For lack of anything better - size_up (Lam b e) | isId b = size_up e `addSizeN` 1 + size_up (Lam b e) | isId b = lamScrutDiscount (size_up e `addSizeN` 1) | otherwise = size_up e size_up (Let (NonRec binder rhs) body) = nukeScrutDiscount (size_up rhs) `addSize` size_up body `addSizeN` - 1 -- For the allocation + (if isUnLiftedType (idType binder) then 0 else 1) + -- For the allocation + -- If the binder has an unlifted type there is no allocation size_up (Let (Rec pairs) body) = nukeScrutDiscount rhs_size `addSize` @@ -230,62 +213,136 @@ sizeExpr (I# bOMB_OUT_SIZE) args expr where rhs_size = foldr (addSize . size_up . snd) sizeZero pairs - size_up (Case scrut _ alts) - = nukeScrutDiscount (size_up scrut) `addSize` - arg_discount scrut `addSize` - foldr (addSize . size_up_alt) sizeZero alts `addSizeN` - case (splitAlgTyConApp_maybe (coreExprType scrut)) of - Nothing -> 1 - Just (tc,_,_) -> tyConFamilySize tc + size_up (Case (Var v) _ alts) + | v `elem` top_args -- We are scrutinising an argument variable + = +{- I'm nuking this special case; BUT see the comment with case alternatives. + + (a) It's too eager. We don't want to inline a wrapper into a + context with no benefit. + E.g. \ x. f (x+x) o point in inlining (+) here! + + (b) It's ineffective. Once g's wrapper is inlined, its case-expressions + aren't scrutinising arguments any more + + case alts of + + [alt] -> size_up_alt alt `addSize` SizeIs 0# (unitBag (v, 1)) 0# + -- We want to make wrapper-style evaluation look cheap, so that + -- when we inline a wrapper it doesn't make call site (much) bigger + -- Otherwise we get nasty phase ordering stuff: + -- f x = g x x + -- h y = ...(f e)... + -- If we inline g's wrapper, f looks big, and doesn't get inlined + -- into h; if we inline f first, while it looks small, then g's + -- wrapper will get inlined later anyway. To avoid this nasty + -- ordering difference, we make (case a of (x,y) -> ...), + -- *where a is one of the arguments* look free. + + other -> +-} + alts_size (foldr addSize sizeOne alt_sizes) -- The 1 is for the scrutinee + (foldr1 maxSize alt_sizes) + + -- Good to inline if an arg is scrutinised, because + -- that may eliminate allocation in the caller + -- And it eliminates the case itself + + where + alt_sizes = map size_up_alt alts + + -- alts_size tries to compute a good discount for + -- the case when we are scrutinising an argument variable + alts_size (SizeIs tot tot_disc tot_scrut) -- Size of all alternatives + (SizeIs max max_disc max_scrut) -- Size of biggest alternative + = SizeIs tot (unitBag (v, iBox (_ILIT 1 +# tot -# max)) `unionBags` max_disc) max_scrut + -- If the variable is known, we produce a discount that + -- will take us back to 'max', the size of rh largest alternative + -- The 1+ is a little discount for reduced allocation in the caller + alts_size tot_size _ = tot_size + + + size_up (Case e _ alts) = nukeScrutDiscount (size_up e) `addSize` + foldr (addSize . size_up_alt) sizeZero alts + -- We don't charge for the case itself + -- It's a strict thing, and the price of the call + -- is paid by scrut. Also consider + -- case f x of DEFAULT -> e + -- This is just ';'! Don't charge for it. ------------ - size_up_alt (con, bndrs, rhs) = size_up rhs - -- Don't charge for args, so that wrappers look cheap - - ------------ - size_up_con (Literal lit) nv | isNoRepLit lit = sizeN uNFOLDING_NOREP_LIT_COST - | otherwise = sizeOne + size_up_app (App fun arg) args + | isTypeArg arg = size_up_app fun args + | otherwise = size_up_app fun (arg:args) + size_up_app fun args = foldr (addSize . nukeScrutDiscount . size_up) + (size_up_fun fun args) + args + + -- A function application with at least one value argument + -- so if the function is an argument give it an arg-discount + -- + -- Also behave specially if the function is a build + -- + -- Also if the function is a constant Id (constr or primop) + -- compute discounts specially + size_up_fun (Var fun) args + | fun `hasKey` buildIdKey = buildSize + | fun `hasKey` augmentIdKey = augmentSize + | otherwise + = case idFlavour fun of + DataConId dc -> conSizeN (valArgCount args) + + PrimOpId op -> primOpSize op (valArgCount args) + -- foldr addSize (primOpSize op) (map arg_discount args) + -- At one time I tried giving an arg-discount if a primop + -- is applied to one of the function's arguments, but it's + -- not good. At the moment, any unlifted-type arg gets a + -- 'True' for 'yes I'm evald', so we collect the discount even + -- if we know nothing about it. And just having it in a primop + -- doesn't help at all if we don't know something more. + + other -> fun_discount fun `addSizeN` + (1 + length (filter (not . exprIsTrivial) args)) + -- The 1+ is for the function itself + -- Add 1 for each non-trivial arg; + -- the allocation cost, as in let(rec) + -- Slight hack here: for constructors the args are almost always + -- trivial; and for primops they are almost always prim typed + -- We should really only count for non-prim-typed args in the + -- general case, but that seems too much like hard work + + size_up_fun other args = size_up other - size_up_con (DataCon dc) n_val_args = conSizeN n_val_args - - size_up_con (PrimOp op) nv = sizeN op_cost - where - op_cost = if primOpOutOfLine op - then uNFOLDING_DEAR_OP_COST - -- these *tend* to be more expensive; - -- number chosen to avoid unfolding (HACK) - else uNFOLDING_CHEAP_OP_COST + ------------ + size_up_alt (con, bndrs, rhs) = size_up rhs + -- Don't charge for args, so that wrappers look cheap + -- (See comments about wrappers with Case) ------------ - -- We want to record if we're case'ing an argument - arg_discount (Var v) | v `is_elem` args = scrutArg v - arg_discount other = sizeZero - - is_elem :: Id -> [Id] -> Bool - is_elem = isIn "size_up_scrut" + -- We want to record if we're case'ing, or applying, an argument + fun_discount v | v `elem` top_args = SizeIs 0# (unitBag (v, opt_UF_FunAppDiscount)) 0# + fun_discount other = sizeZero ------------ -- These addSize things have to be here because -- I don't want to give them bOMB_OUT_SIZE as an argument - addSizeN TooBig _ = TooBig - addSizeN (SizeIs n xs d) (I# m) - | n_tot -# d <# bOMB_OUT_SIZE = SizeIs n_tot xs d - | otherwise = TooBig + addSizeN TooBig _ = TooBig + addSizeN (SizeIs n xs d) m + | n_tot ># (iUnbox bOMB_OUT_SIZE) = TooBig + | otherwise = SizeIs n_tot xs d where - n_tot = n +# m + n_tot = n +# iUnbox m addSize TooBig _ = TooBig addSize _ TooBig = TooBig addSize (SizeIs n1 xs d1) (SizeIs n2 ys d2) - | (n_tot -# d_tot) <# bOMB_OUT_SIZE = SizeIs n_tot xys d_tot - | otherwise = TooBig + | n_tot ># (iUnbox bOMB_OUT_SIZE) = TooBig + | otherwise = SizeIs n_tot xys d_tot where n_tot = n1 +# n2 d_tot = d1 +# d2 - xys = xs ++ ys - - + xys = xs `unionBags` ys \end{code} Code for manipulating sizes @@ -293,22 +350,56 @@ Code for manipulating sizes \begin{code} data ExprSize = TooBig - | SizeIs Int# -- Size found - [Id] -- Arguments cased herein - Int# -- Size to subtract if result is scrutinised - -- by a case expression - -sizeZero = SizeIs 0# [] 0# -sizeOne = SizeIs 1# [] 0# -sizeN (I# n) = SizeIs n [] 0# -conSizeN (I# n) = SizeIs 0# [] n -- We don't count 1 for the constructor because we're - -- quite keen to get constructors into the open -scrutArg v = SizeIs 0# [v] 0# - + | SizeIs FastInt -- Size found + (Bag (Id,Int)) -- Arguments cased herein, and discount for each such + FastInt -- Size to subtract if result is scrutinised + -- by a case expression + +isTooBig TooBig = True +isTooBig _ = False + +maxSize TooBig _ = TooBig +maxSize _ TooBig = TooBig +maxSize s1@(SizeIs n1 _ _) s2@(SizeIs n2 _ _) | n1 ># n2 = s1 + | otherwise = s2 + +sizeZero = SizeIs (_ILIT 0) emptyBag (_ILIT 0) +sizeOne = SizeIs (_ILIT 1) emptyBag (_ILIT 0) +sizeTwo = SizeIs (_ILIT 2) emptyBag (_ILIT 0) +sizeN n = SizeIs (iUnbox n) emptyBag (_ILIT 0) +conSizeN n = SizeIs (_ILIT 1) emptyBag (iUnbox n +# _ILIT 1) + -- Treat constructors as size 1; we are keen to expose them + -- (and we charge separately for their args). We can't treat + -- them as size zero, else we find that (iBox x) has size 1, + -- which is the same as a lone variable; and hence 'v' will + -- always be replaced by (iBox x), where v is bound to iBox x. + +primOpSize op n_args + | not (primOpIsDupable op) = sizeN opt_UF_DearOp + | not (primOpOutOfLine op) = sizeZero -- These are good to inline + | otherwise = sizeOne + +buildSize = SizeIs (-2#) emptyBag 4# + -- We really want to inline applications of build + -- build t (\cn -> e) should cost only the cost of e (because build will be inlined later) + -- Indeed, we should add a result_discount becuause build is + -- very like a constructor. We don't bother to check that the + -- build is saturated (it usually is). The "-2" discounts for the \c n, + -- The "4" is rather arbitrary. + +augmentSize = SizeIs (-2#) emptyBag 4# + -- Ditto (augment t (\cn -> e) ys) should cost only the cost of + -- e plus ys. The -2 accounts for the \cn + nukeScrutDiscount (SizeIs n vs d) = SizeIs n vs 0# nukeScrutDiscount TooBig = TooBig + +-- When we return a lambda, give a discount if it's used (applied) +lamScrutDiscount (SizeIs n vs d) = case opt_UF_FunAppDiscount of { d -> SizeIs n vs (iUnbox d) } +lamScrutDiscount TooBig = TooBig \end{code} + %************************************************************************ %* * \subsection[considerUnfolding]{Given all the info, do (not) do the unfolding} @@ -339,64 +430,26 @@ the expression is going to be taken apart, discounting its size is more accurate (see @sizeExpr@ above for how this discount size is computed). -\begin{code} -smallEnoughToInline :: Id -- The function (trace msg only) - -> [Bool] -- Evaluated-ness of value arguments - -> Bool -- Result is scrutinised - -> UnfoldingGuidance - -> Bool -- True => unfold it - -smallEnoughToInline _ _ _ UnfoldAlways = True -smallEnoughToInline _ _ _ UnfoldNever = False -smallEnoughToInline id arg_is_evald_s result_is_scruted - (UnfoldIfGoodArgs m_tys_wanted n_vals_wanted discount_vec size scrut_discount) - = if enough_args n_vals_wanted arg_is_evald_s && - size - discount <= opt_UnfoldingUseThreshold - then - True - else - False - where - - enough_args n [] | n > 0 = False -- A function with no value args => don't unfold - enough_args _ _ = True -- Otherwise it's ok to try - - -- We multiple the raw discounts (args_discount and result_discount) - -- ty opt_UnfoldingKeenessFactor because the former have to do with - -- *size* whereas the discounts imply that there's some extra - -- *efficiency* to be gained (e.g. beta reductions, case reductions) - -- by inlining. - - -- we also discount 1 for each argument passed, because these will - -- reduce with the lambdas in the function (we count 1 for a lambda - -- in size_up). - - discount :: Int - discount = length (take n_vals_wanted arg_is_evald_s) + - round ( - opt_UnfoldingKeenessFactor * - fromInt (args_discount + result_discount) - ) - - args_discount = sum (zipWith arg_discount discount_vec arg_is_evald_s) - result_discount | result_is_scruted = scrut_discount - | otherwise = 0 - - arg_discount no_of_constrs is_evald - | is_evald = no_of_constrs * opt_UnfoldingConDiscount - | otherwise = 0 -\end{code} - We use this one to avoid exporting inlinings that we ``couldn't possibly use'' on the other side. Can be overridden w/ flaggery. Just the same as smallEnoughToInline, except that it has no actual arguments. \begin{code} -couldBeSmallEnoughToInline :: Id -> UnfoldingGuidance -> Bool -couldBeSmallEnoughToInline id guidance = smallEnoughToInline id (repeat True) True guidance +couldBeSmallEnoughToInline :: Int -> CoreExpr -> Bool +couldBeSmallEnoughToInline threshold rhs = case calcUnfoldingGuidance threshold rhs of + UnfoldNever -> False + other -> True + +certainlyWillInline :: Id -> Bool + -- Sees if the Id is pretty certain to inline +certainlyWillInline v + = case idUnfolding v of -certainlySmallEnoughToInline :: Id -> UnfoldingGuidance -> Bool -certainlySmallEnoughToInline id guidance = smallEnoughToInline id (repeat False) False guidance + CoreUnfolding _ _ is_value _ g@(UnfoldIfGoodArgs n_vals _ size _) + -> is_value + && size - (n_vals +1) <= opt_UF_UseThreshold + + other -> False \end{code} @okToUnfoldInHifile@ is used when emitting unfolding info into an interface @@ -415,18 +468,287 @@ okToUnfoldInHiFile :: CoreExpr -> Bool okToUnfoldInHiFile e = opt_UnfoldCasms || go e where -- Race over an expression looking for CCalls.. - go (Var _) = True - go (Con (Literal lit) _) = not (isLitLitLit lit) - go (Con (PrimOp op) args) = okToUnfoldPrimOp op && all go args - go (Con con args) = True -- con args are always atomic + go (Var v) = case isPrimOpId_maybe v of + Just op -> okToUnfoldPrimOp op + Nothing -> True + go (Lit lit) = not (isLitLitLit lit) go (App fun arg) = go fun && go arg go (Lam _ body) = go body go (Let binds body) = and (map go (body :rhssOfBind binds)) - go (Case scrut bndr alts) = and (map go (scrut:rhssOfAlts alts)) + go (Case scrut bndr alts) = and (map go (scrut:rhssOfAlts alts)) && + not (any isLitLitLit [ lit | (LitAlt lit, _, _) <- alts ]) go (Note _ body) = go body go (Type _) = True -- ok to unfold a PrimOp as long as it's not a _casm_ - okToUnfoldPrimOp (CCallOp _ is_casm _ _) = not is_casm - okToUnfoldPrimOp _ = True + okToUnfoldPrimOp (CCallOp ccall) = not (ccallIsCasm ccall) + okToUnfoldPrimOp _ = True \end{code} + + +%************************************************************************ +%* * +\subsection{callSiteInline} +%* * +%************************************************************************ + +This is the key function. It decides whether to inline a variable at a call site + +callSiteInline is used at call sites, so it is a bit more generous. +It's a very important function that embodies lots of heuristics. +A non-WHNF can be inlined if it doesn't occur inside a lambda, +and occurs exactly once or + occurs once in each branch of a case and is small + +If the thing is in WHNF, there's no danger of duplicating work, +so we can inline if it occurs once, or is small + +NOTE: we don't want to inline top-level functions that always diverge. +It just makes the code bigger. Tt turns out that the convenient way to prevent +them inlining is to give them a NOINLINE pragma, which we do in +StrictAnal.addStrictnessInfoToTopId + +\begin{code} +callSiteInline :: DynFlags + -> Bool -- True <=> the Id is black listed + -> Bool -- 'inline' note at call site + -> OccInfo + -> Id -- The Id + -> [Bool] -- One for each value arg; True if it is interesting + -> Bool -- True <=> continuation is interesting + -> Maybe CoreExpr -- Unfolding, if any + + +callSiteInline dflags black_listed inline_call occ id arg_infos interesting_cont + = case idUnfolding id of { + NoUnfolding -> Nothing ; + OtherCon cs -> Nothing ; + CompulsoryUnfolding unf_template | black_listed -> Nothing + | otherwise -> Just unf_template ; + -- Constructors have compulsory unfoldings, but + -- may have rules, in which case they are + -- black listed till later + CoreUnfolding unf_template is_top is_value is_cheap guidance -> + + let + result | yes_or_no = Just unf_template + | otherwise = Nothing + + n_val_args = length arg_infos + + yes_or_no + | black_listed = False + | otherwise = case occ of + IAmDead -> pprTrace "callSiteInline: dead" (ppr id) False + IAmALoopBreaker -> False + OneOcc in_lam one_br -> (not in_lam || is_cheap) && consider_safe in_lam True one_br + NoOccInfo -> is_cheap && consider_safe True False False + + consider_safe in_lam once once_in_one_branch + -- consider_safe decides whether it's a good idea to inline something, + -- given that there's no work-duplication issue (the caller checks that). + -- once_in_one_branch = True means there's a unique textual occurrence + | inline_call = True + + | once_in_one_branch + -- Be very keen to inline something if this is its unique occurrence: + -- + -- a) Inlining gives a good chance of eliminating the original + -- binding (and hence the allocation) for the thing. + -- (Provided it's not a top level binding, in which case the + -- allocation costs nothing.) + -- + -- b) Inlining a function that is called only once exposes the + -- body function to the call site. + -- + -- The only time we hold back is when substituting inside a lambda; + -- then if the context is totally uninteresting (not applied, not scrutinised) + -- there is no point in substituting because it might just increase allocation, + -- by allocating the function itself many times + -- + -- Note: there used to be a '&& not top_level' in the guard above, + -- but that stopped us inlining top-level functions used only once, + -- which is stupid + = not in_lam || not (null arg_infos) || interesting_cont + + | otherwise + = case guidance of + UnfoldNever -> False ; + UnfoldIfGoodArgs n_vals_wanted arg_discounts size res_discount + + | enough_args && size <= (n_vals_wanted + 1) + -- No size increase + -- Size of call is n_vals_wanted (+1 for the function) + -> True + + | otherwise + -> some_benefit && small_enough + + where + some_benefit = or arg_infos || really_interesting_cont || + (not is_top && (once || (n_vals_wanted > 0 && enough_args))) + -- If it occurs more than once, there must be something interesting + -- about some argument, or the result context, to make it worth inlining + -- + -- If a function has a nested defn we also record some-benefit, + -- on the grounds that we are often able to eliminate the binding, + -- and hence the allocation, for the function altogether; this is good + -- for join points. But this only makes sense for *functions*; + -- inlining a constructor doesn't help allocation unless the result is + -- scrutinised. UNLESS the constructor occurs just once, albeit possibly + -- in multiple case branches. Then inlining it doesn't increase allocation, + -- but it does increase the chance that the constructor won't be allocated at all + -- in the branches that don't use it. + + enough_args = n_val_args >= n_vals_wanted + really_interesting_cont | n_val_args < n_vals_wanted = False -- Too few args + | n_val_args == n_vals_wanted = interesting_cont + | otherwise = True -- Extra args + -- really_interesting_cont tells if the result of the + -- call is in an interesting context. + + small_enough = (size - discount) <= opt_UF_UseThreshold + discount = computeDiscount n_vals_wanted arg_discounts res_discount + arg_infos really_interesting_cont + + in +#ifdef DEBUG + if dopt_D_dump_inlinings dflags then + pprTrace "Considering inlining" + (ppr id <+> vcat [text "black listed:" <+> ppr black_listed, + text "occ info:" <+> ppr occ, + text "arg infos" <+> ppr arg_infos, + text "interesting continuation" <+> ppr interesting_cont, + text "is value:" <+> ppr is_value, + text "is cheap:" <+> ppr is_cheap, + text "guidance" <+> ppr guidance, + text "ANSWER =" <+> if yes_or_no then text "YES" else text "NO", + if yes_or_no then + text "Unfolding =" <+> pprCoreExpr unf_template + else empty]) + result + else +#endif + result + } + +computeDiscount :: Int -> [Int] -> Int -> [Bool] -> Bool -> Int +computeDiscount n_vals_wanted arg_discounts res_discount arg_infos result_used + -- We multiple the raw discounts (args_discount and result_discount) + -- ty opt_UnfoldingKeenessFactor because the former have to do with + -- *size* whereas the discounts imply that there's some extra + -- *efficiency* to be gained (e.g. beta reductions, case reductions) + -- by inlining. + + -- we also discount 1 for each argument passed, because these will + -- reduce with the lambdas in the function (we count 1 for a lambda + -- in size_up). + = 1 + -- Discount of 1 because the result replaces the call + -- so we count 1 for the function itself + length (take n_vals_wanted arg_infos) + + -- Discount of 1 for each arg supplied, because the + -- result replaces the call + round (opt_UF_KeenessFactor * + fromInt (arg_discount + result_discount)) + where + arg_discount = sum (zipWith mk_arg_discount arg_discounts arg_infos) + + mk_arg_discount discount is_evald | is_evald = discount + | otherwise = 0 + + -- Don't give a result discount unless there are enough args + result_discount | result_used = res_discount -- Over-applied, or case scrut + | otherwise = 0 +\end{code} + + +%************************************************************************ +%* * +\subsection{Black-listing} +%* * +%************************************************************************ + +Inlining is controlled by the "Inline phase" number, which is set +by the per-simplification-pass '-finline-phase' flag. + +For optimisation we use phase 1,2 and nothing (i.e. no -finline-phase flag) +in that order. The meanings of these are determined by the @blackListed@ function +here. + +The final simplification doesn't have a phase number. + +Pragmas +~~~~~~~ + Pragma Black list if + +(least black listing, most inlining) + INLINE n foo phase is Just p *and* p Maybe Int -- Inline phase + -> Id -> Bool -- True <=> blacklisted + +-- The blackListed function sees whether a variable should *not* be +-- inlined because of the inline phase we are in. This is the sole +-- place that the inline phase number is looked at. + +blackListed rule_vars Nothing -- Last phase + = \v -> isNeverInlinePrag (idInlinePragma v) + +blackListed rule_vars (Just phase) + = \v -> normal_case rule_vars phase v + +normal_case rule_vars phase v + = case idInlinePragma v of + NoInlinePragInfo -> has_rules + + IMustNotBeINLINEd from_INLINE Nothing + | from_INLINE -> has_rules -- Black list until final phase + | otherwise -> True -- Always blacklisted + + IMustNotBeINLINEd from_INLINE (Just threshold) + | from_INLINE -> (phase < threshold && has_rules) + | otherwise -> (phase < threshold || has_rules) + where + has_rules = v `elemVarSet` rule_vars + || not (isEmptyCoreRules (idSpecialisation v)) +\end{code} + + +SLPJ 95/04: Why @runST@ must be inlined very late: +\begin{verbatim} +f x = + runST ( \ s -> let + (a, s') = newArray# 100 [] s + (_, s'') = fill_in_array_or_something a x s' + in + freezeArray# a s'' ) +\end{verbatim} +If we inline @runST@, we'll get: +\begin{verbatim} +f x = let + (a, s') = newArray# 100 [] realWorld#{-NB-} + (_, s'') = fill_in_array_or_something a x s' + in + freezeArray# a s'' +\end{verbatim} +And now the @newArray#@ binding can be floated to become a CAF, which +is totally and utterly wrong: +\begin{verbatim} +f = let + (a, s') = newArray# 100 [] realWorld#{-NB-} -- YIKES!!! + in + \ x -> + let (_, s'') = fill_in_array_or_something a x s' in + freezeArray# a s'' +\end{verbatim} +All calls to @f@ will share a {\em single} array! + +Yet we do want to inline runST sometime, so we can avoid +needless code. Solution: black list it until the last moment. +