%
% (c) The University of Glasgow 2006
% (c) The AQUA Project, Glasgow University, 1994-1998
%
Core-syntax unfoldings
Unfoldings (which can travel across module boundaries) are in Core
syntax (namely @CoreExpr@s).
The type @Unfolding@ sits ``above'' simply-Core-expressions
unfoldings, capturing ``higher-level'' things we know about a binding,
usually things that the simplifier found out (e.g., ``it's a
literal''). In the corner of a @CoreUnfolding@ unfolding, you will
find, unsurprisingly, a Core expression.
\begin{code}
module CoreUnfold (
Unfolding, UnfoldingGuidance,
noUnfolding, mkImplicitUnfolding,
mkUnfolding, mkCoreUnfolding,
mkTopUnfolding, mkSimpleUnfolding,
mkInlineUnfolding, mkInlinableUnfolding, mkWwInlineRule,
mkCompulsoryUnfolding, mkDFunUnfolding,
interestingArg, ArgSummary(..),
couldBeSmallEnoughToInline, inlineBoringOk,
certainlyWillInline, smallEnoughToInline,
callSiteInline, CallCtxt(..),
exprIsConApp_maybe, exprIsLiteral_maybe
) where
#include "HsVersions.h"
import DynFlags
import CoreSyn
import PprCore ()
import OccurAnal ( occurAnalyseExpr )
import CoreSubst hiding( substTy )
import CoreArity ( manifestArity, exprBotStrictness_maybe )
import CoreUtils
import Id
import DataCon
import Literal
import PrimOp
import IdInfo
import BasicTypes ( Arity )
import Type
import PrelNames
import TysPrim ( realWorldStatePrimTy )
import Bag
import Util
import FastTypes
import FastString
import Outputable
import ForeignCall
import qualified Data.ByteString as BS
import Data.Maybe
\end{code}
%************************************************************************
%* *
\subsection{Making unfoldings}
%* *
%************************************************************************
\begin{code}
mkTopUnfolding :: DynFlags -> Bool -> CoreExpr -> Unfolding
mkTopUnfolding dflags = mkUnfolding dflags InlineRhs True
mkImplicitUnfolding :: DynFlags -> CoreExpr -> Unfolding
mkImplicitUnfolding dflags expr
= mkTopUnfolding dflags False (simpleOptExpr expr)
mkSimpleUnfolding :: DynFlags -> CoreExpr -> Unfolding
mkSimpleUnfolding dflags = mkUnfolding dflags InlineRhs False False
mkDFunUnfolding :: [Var] -> DataCon -> [CoreExpr] -> Unfolding
mkDFunUnfolding bndrs con ops
= DFunUnfolding { df_bndrs = bndrs
, df_con = con
, df_args = map occurAnalyseExpr ops }
mkWwInlineRule :: CoreExpr -> Arity -> Unfolding
mkWwInlineRule expr arity
= mkCoreUnfolding InlineStable True
(simpleOptExpr expr) arity
(UnfWhen unSaturatedOk boringCxtNotOk)
mkCompulsoryUnfolding :: CoreExpr -> Unfolding
mkCompulsoryUnfolding expr
= mkCoreUnfolding InlineCompulsory True
(simpleOptExpr expr) 0
(UnfWhen unSaturatedOk boringCxtOk)
mkInlineUnfolding :: Maybe Arity -> CoreExpr -> Unfolding
mkInlineUnfolding mb_arity expr
= mkCoreUnfolding InlineStable
True
expr' arity
(UnfWhen unsat_ok boring_ok)
where
expr' = simpleOptExpr expr
(unsat_ok, arity) = case mb_arity of
Nothing -> (unSaturatedOk, manifestArity expr')
Just ar -> (needSaturated, ar)
boring_ok = inlineBoringOk expr'
mkInlinableUnfolding :: DynFlags -> CoreExpr -> Unfolding
mkInlinableUnfolding dflags expr
= mkUnfolding dflags InlineStable True is_bot expr'
where
expr' = simpleOptExpr expr
is_bot = isJust (exprBotStrictness_maybe expr')
\end{code}
Internal functions
\begin{code}
mkCoreUnfolding :: UnfoldingSource -> Bool -> CoreExpr
-> Arity -> UnfoldingGuidance -> Unfolding
mkCoreUnfolding src top_lvl expr arity guidance
= CoreUnfolding { uf_tmpl = occurAnalyseExpr expr,
uf_src = src,
uf_arity = arity,
uf_is_top = top_lvl,
uf_is_value = exprIsHNF expr,
uf_is_conlike = exprIsConLike expr,
uf_is_work_free = exprIsWorkFree expr,
uf_expandable = exprIsExpandable expr,
uf_guidance = guidance }
mkUnfolding :: DynFlags -> UnfoldingSource -> Bool -> Bool -> CoreExpr
-> Unfolding
mkUnfolding dflags src top_lvl is_bottoming expr
| top_lvl && is_bottoming
, not (exprIsTrivial expr)
= NoUnfolding
| otherwise
= CoreUnfolding { uf_tmpl = occurAnalyseExpr expr,
uf_src = src,
uf_arity = arity,
uf_is_top = top_lvl,
uf_is_value = exprIsHNF expr,
uf_is_conlike = exprIsConLike expr,
uf_expandable = exprIsExpandable expr,
uf_is_work_free = exprIsWorkFree expr,
uf_guidance = guidance }
where
(arity, guidance) = calcUnfoldingGuidance dflags expr
\end{code}
Note [Occurrence analysis of unfoldings]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We do occurrence-analysis of unfoldings once and for all, when the
unfolding is built, rather than each time we inline them.
But given this decision it's vital that we do
*always* do it. Consider this unfolding
\x -> letrec { f = ...g...; g* = f } in body
where g* is (for some strange reason) the loop breaker. If we don't
occ-anal it when reading it in, we won't mark g as a loop breaker, and
we may inline g entirely in body, dropping its binding, and leaving
the occurrence in f out of scope. This happened in Trac #8892, where
the unfolding in question was a DFun unfolding.
But more generally, the simplifier is designed on the
basis that it is looking at occurrence-analysed expressions, so better
ensure that they acutally are.
Note [Calculate unfolding guidance on the non-occ-anal'd expression]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Notice that we give the non-occur-analysed expression to
calcUnfoldingGuidance. In some ways it'd be better to occur-analyse
first; for example, 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* and *must not* occ-analyse before computing
the size because
a) The size computation bales out after a while, whereas occurrence
analysis does not.
b) Residency increases sharply if you occ-anal first. I'm not
100% sure why, but it's a large effect. Compiling Cabal went
from residency of 534M to over 800M with this one change.
This can occasionally mean that the guidance is very pessimistic;
it gets fixed up next round. And it should be rare, because large
let-bound things that are dead are usually caught by preInlineUnconditionally
%************************************************************************
%* *
\subsection{The UnfoldingGuidance type}
%* *
%************************************************************************
\begin{code}
inlineBoringOk :: CoreExpr -> Bool
inlineBoringOk e
= go 0 e
where
go :: Int -> CoreExpr -> Bool
go credit (Lam x e) | isId x = go (credit+1) e
| otherwise = go credit e
go credit (App f (Type {})) = go credit f
go credit (App f a) | credit > 0
, exprIsTrivial a = go (credit1) f
go credit (Tick _ e) = go credit e
go credit (Cast e _) = go credit e
go _ (Var {}) = boringCxtOk
go _ _ = boringCxtNotOk
calcUnfoldingGuidance
:: DynFlags
-> CoreExpr
-> (Arity, UnfoldingGuidance)
calcUnfoldingGuidance dflags expr
= case collectBinders expr of { (bndrs, body) ->
let
bOMB_OUT_SIZE = ufCreationThreshold dflags
val_bndrs = filter isId bndrs
n_val_bndrs = length val_bndrs
guidance
= case sizeExpr dflags (iUnbox bOMB_OUT_SIZE) val_bndrs body of
TooBig -> UnfNever
SizeIs size cased_bndrs scrut_discount
| uncondInline expr n_val_bndrs (iBox size)
-> UnfWhen unSaturatedOk boringCxtOk
| otherwise
-> UnfIfGoodArgs { ug_args = map (discount cased_bndrs) val_bndrs
, ug_size = iBox size
, ug_res = iBox scrut_discount }
discount :: Bag (Id,Int) -> Id -> Int
discount cbs bndr = foldlBag combine 0 cbs
where
combine acc (bndr', disc)
| bndr == bndr' = acc `plus_disc` disc
| otherwise = acc
plus_disc :: Int -> Int -> Int
plus_disc | isFunTy (idType bndr) = max
| otherwise = (+)
in
(n_val_bndrs, guidance) }
\end{code}
Note [Computing the size of an expression]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The basic idea of sizeExpr is obvious enough: count nodes. But getting the
heuristics right has taken a long time. Here's the basic strategy:
* Variables, literals: 0
(Exception for string literals, see litSize.)
* Function applications (f e1 .. en): 1 + #value args
* Constructor applications: 1, regardless of #args
* Let(rec): 1 + size of components
* Note, cast: 0
Examples
Size Term
--------------
0 42#
0 x
0 True
2 f x
1 Just x
4 f (g x)
Notice that 'x' counts 0, while (f x) counts 2. That's deliberate: there's
a function call to account for. Notice also that constructor applications
are very cheap, because exposing them to a caller is so valuable.
[25/5/11] All sizes are now multiplied by 10, except for primops
(which have sizes like 1 or 4. This makes primops look fantastically
cheap, and seems to be almost unversally beneficial. Done partly as a
result of #4978.
Note [Do not inline top-level bottoming functions]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The FloatOut pass has gone to some trouble to float out calls to 'error'
and similar friends. See Note [Bottoming floats] in SetLevels.
Do not re-inline them! But we *do* still inline if they are very small
(the uncondInline stuff).
Note [INLINE for small functions]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider {-# INLINE f #-}
f x = Just x
g y = f y
Then f's RHS is no larger than its LHS, so we should inline it into
even the most boring context. In general, f the function is
sufficiently small that its body is as small as the call itself, the
inline unconditionally, regardless of how boring the context is.
Things to note:
(1) We inline *unconditionally* if inlined thing is smaller (using sizeExpr)
than the thing it's replacing. Notice that
(f x) --> (g 3) -- YES, unconditionally
(f x) --> x : [] -- YES, *even though* there are two
-- arguments to the cons
x --> g 3 -- NO
x --> Just v -- NO
It's very important not to unconditionally replace a variable by
a non-atomic term.
(2) We do this even if the thing isn't saturated, else we end up with the
silly situation that
f x y = x
...map (f 3)...
doesn't inline. Even in a boring context, inlining without being
saturated will give a lambda instead of a PAP, and will be more
efficient at runtime.
(3) However, when the function's arity > 0, we do insist that it
has at least one value argument at the call site. (This check is
made in the UnfWhen case of callSiteInline.) Otherwise we find this:
f = /\a \x:a. x
d = /\b. MkD (f b)
If we inline f here we get
d = /\b. MkD (\x:b. x)
and then prepareRhs floats out the argument, abstracting the type
variables, so we end up with the original again!
(4) We must be much more cautious about arity-zero things. Consider
let x = y +# z in ...
In *size* terms primops look very small, because the generate a
single instruction, but we do not want to unconditionally replace
every occurrence of x with (y +# z). So we only do the
unconditional-inline thing for *trivial* expressions.
NB: you might think that PostInlineUnconditionally would do this
but it doesn't fire for top-level things; see SimplUtils
Note [Top level and postInlineUnconditionally]
\begin{code}
uncondInline :: CoreExpr -> Arity -> Int -> Bool
uncondInline rhs arity size
| arity > 0 = size <= 10 * (arity + 1)
| otherwise = exprIsTrivial rhs
\end{code}
\begin{code}
sizeExpr :: DynFlags
-> FastInt
-> [Id]
-> CoreExpr
-> ExprSize
sizeExpr dflags bOMB_OUT_SIZE top_args expr
= size_up expr
where
size_up (Cast e _) = size_up e
size_up (Tick _ e) = size_up e
size_up (Type _) = sizeZero
size_up (Coercion _) = sizeZero
size_up (Lit lit) = sizeN (litSize lit)
size_up (Var f) | isRealWorldId f = sizeZero
| otherwise = size_up_call f [] 0
size_up (App fun arg)
| isTyCoArg arg = size_up fun
| otherwise = size_up arg `addSizeNSD`
size_up_app fun [arg] (if isRealWorldExpr arg then 1 else 0)
size_up (Lam b e)
| isId b && not (isRealWorldId b) = lamScrutDiscount dflags (size_up e `addSizeN` 10)
| otherwise = size_up e
size_up (Let (NonRec binder rhs) body)
= size_up rhs `addSizeNSD`
size_up body `addSizeN`
(if isUnLiftedType (idType binder) then 0 else 10)
size_up (Let (Rec pairs) body)
= foldr (addSizeNSD . size_up . snd)
(size_up body `addSizeN` (10 * length pairs))
pairs
size_up (Case (Var v) _ _ alts)
| v `elem` top_args
= alts_size (foldr addAltSize sizeZero alt_sizes)
(foldr maxSize sizeZero alt_sizes)
where
alt_sizes = map size_up_alt alts
alts_size (SizeIs tot tot_disc tot_scrut)
(SizeIs max _ _)
= SizeIs tot (unitBag (v, iBox (_ILIT(20) +# tot -# max)) `unionBags` tot_disc) tot_scrut
alts_size tot_size _ = tot_size
size_up (Case e _ _ alts) = size_up e `addSizeNSD`
foldr (addAltSize . size_up_alt) case_size alts
where
case_size
| is_inline_scrut e, not (lengthExceeds alts 1) = sizeN (10)
| otherwise = sizeZero
is_inline_scrut (Var v) = isUnLiftedType (idType v)
is_inline_scrut scrut
| (Var f, _) <- collectArgs scrut
= case idDetails f of
FCallId fc -> not (isSafeForeignCall fc)
PrimOpId op -> not (primOpOutOfLine op)
_other -> False
| otherwise
= False
size_up_app (App fun arg) args voids
| isTyCoArg arg = size_up_app fun args voids
| isRealWorldExpr arg = size_up_app fun (arg:args) (voids + 1)
| otherwise = size_up arg `addSizeNSD`
size_up_app fun (arg:args) voids
size_up_app (Var fun) args voids = size_up_call fun args voids
size_up_app other args voids = size_up other `addSizeN` (length args voids)
size_up_call :: Id -> [CoreExpr] -> Int -> ExprSize
size_up_call fun val_args voids
= case idDetails fun of
FCallId _ -> sizeN (10 * (1 + length val_args))
DataConWorkId dc -> conSize dc (length val_args)
PrimOpId op -> primOpSize op (length val_args)
ClassOpId _ -> classOpSize dflags top_args val_args
_ -> funSize dflags top_args fun (length val_args) voids
size_up_alt (_con, _bndrs, rhs) = size_up rhs `addSizeN` 10
addSizeN TooBig _ = TooBig
addSizeN (SizeIs n xs d) m = mkSizeIs bOMB_OUT_SIZE (n +# iUnbox m) xs d
addAltSize TooBig _ = TooBig
addAltSize _ TooBig = TooBig
addAltSize (SizeIs n1 xs d1) (SizeIs n2 ys d2)
= mkSizeIs bOMB_OUT_SIZE (n1 +# n2)
(xs `unionBags` ys)
(d1 +# d2)
addSizeNSD TooBig _ = TooBig
addSizeNSD _ TooBig = TooBig
addSizeNSD (SizeIs n1 xs _) (SizeIs n2 ys d2)
= mkSizeIs bOMB_OUT_SIZE (n1 +# n2)
(xs `unionBags` ys)
d2
isRealWorldId id = idType id `eqType` realWorldStatePrimTy
isRealWorldExpr (Var id) = isRealWorldId id
isRealWorldExpr _ = False
\end{code}
\begin{code}
litSize :: Literal -> Int
litSize (LitInteger {}) = 100
litSize (MachStr str) = 10 + 10 * ((BS.length str + 3) `div` 4)
litSize _other = 0
classOpSize :: DynFlags -> [Id] -> [CoreExpr] -> ExprSize
classOpSize _ _ []
= sizeZero
classOpSize dflags top_args (arg1 : other_args)
= SizeIs (iUnbox size) arg_discount (_ILIT(0))
where
size = 20 + (10 * length other_args)
arg_discount = case arg1 of
Var dict | dict `elem` top_args
-> unitBag (dict, ufDictDiscount dflags)
_other -> emptyBag
funSize :: DynFlags -> [Id] -> Id -> Int -> Int -> ExprSize
funSize dflags top_args fun n_val_args voids
| fun `hasKey` buildIdKey = buildSize
| fun `hasKey` augmentIdKey = augmentSize
| otherwise = SizeIs (iUnbox size) arg_discount (iUnbox res_discount)
where
some_val_args = n_val_args > 0
size | some_val_args = 10 * (1 + n_val_args voids)
| otherwise = 0
arg_discount | some_val_args && fun `elem` top_args
= unitBag (fun, ufFunAppDiscount dflags)
| otherwise = emptyBag
res_discount | idArity fun > n_val_args = ufFunAppDiscount dflags
| otherwise = 0
conSize :: DataCon -> Int -> ExprSize
conSize dc n_val_args
| n_val_args == 0 = SizeIs (_ILIT(0)) emptyBag (_ILIT(10))
| isUnboxedTupleCon dc = SizeIs (_ILIT(0)) emptyBag (iUnbox (10 * (1 + n_val_args)))
| otherwise = SizeIs (_ILIT(10)) emptyBag (iUnbox (10 * (1 + n_val_args)))
\end{code}
Note [Constructor size and result discount]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Treat a constructors application as size 10, regardless of how many
arguments it has; 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
(Just x) has size 0, which is the same as a lone variable; and hence
'v' will always be replaced by (Just x), where v is bound to Just x.
The "result discount" is applied if the result of the call is
scrutinised (say by a case). For a constructor application that will
mean the constructor application will disappear, so we don't need to
charge it to the function. So the discount should at least match the
cost of the constructor application, namely 10. But to give a bit
of extra incentive we give a discount of 10*(1 + n_val_args).
Simon M tried a MUCH bigger discount: (10 * (10 + n_val_args)),
and said it was an "unambiguous win", but its terribly dangerous
because a fuction with many many case branches, each finishing with
a constructor, can have an arbitrarily large discount. This led to
terrible code bloat: see Trac #6099.
Note [Unboxed tuple size and result discount]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
However, unboxed tuples count as size zero. I found occasions where we had
f x y z = case op# x y z of { s -> (# s, () #) }
and f wasn't getting inlined.
I tried giving unboxed tuples a *result discount* of zero (see the
commented-out line). Why? When returned as a result they do not
allocate, so maybe we don't want to charge so much for them If you
have a non-zero discount here, we find that workers often get inlined
back into wrappers, because it look like
f x = case $wf x of (# a,b #) -> (a,b)
and we are keener because of the case. However while this change
shrank binary sizes by 0.5% it also made spectral/boyer allocate 5%
more. All other changes were very small. So it's not a big deal but I
didn't adopt the idea.
Note [Function and non-function discounts]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We want a discount if the function is applied. A good example is
monadic combinators with continuation arguments, where inlining is
quite important.
But we don't want a big discount when a function is called many times
(see the detailed comments with Trac #6048) because if the function is
big it won't be inlined at its many call sites and no benefit results.
Indeed, we can get exponentially big inlinings this way; that is what
Trac #6048 is about.
On the other hand, for data-valued arguments, if there are lots of
case expressions in the body, each one will get smaller if we apply
the function to a constructor application, so we *want* a big discount
if the argument is scrutinised by many case expressions.
Conclusion:
- For functions, take the max of the discounts
- For data values, take the sum of the discounts
Note [Literal integer size]
~~~~~~~~~~~~~~~~~~~~~~~~~~~
Literal integers *can* be big (mkInteger [...coefficients...]), but
need not be (S# n). We just use an aribitrary big-ish constant here
so that, in particular, we don't inline top-level defns like
n = S# 5
There's no point in doing so -- any optimisations will see the S#
through n's unfolding. Nor will a big size inhibit unfoldings functions
that mention a literal Integer, because the float-out pass will float
all those constants to top level.
\begin{code}
primOpSize :: PrimOp -> Int -> ExprSize
primOpSize op n_val_args
= if primOpOutOfLine op
then sizeN (op_size + n_val_args)
else sizeN op_size
where
op_size = primOpCodeSize op
buildSize :: ExprSize
buildSize = SizeIs (_ILIT(0)) emptyBag (_ILIT(40))
augmentSize :: ExprSize
augmentSize = SizeIs (_ILIT(0)) emptyBag (_ILIT(40))
lamScrutDiscount :: DynFlags -> ExprSize -> ExprSize
lamScrutDiscount dflags (SizeIs n vs _) = SizeIs n vs (iUnbox (ufFunAppDiscount dflags))
lamScrutDiscount _ TooBig = TooBig
\end{code}
Note [addAltSize result discounts]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When adding the size of alternatives, we *add* the result discounts
too, rather than take the *maximum*. For a multi-branch case, this
gives a discount for each branch that returns a constructor, making us
keener to inline. I did try using 'max' instead, but it makes nofib
'rewrite' and 'puzzle' allocate significantly more, and didn't make
binary sizes shrink significantly either.
Note [Discounts and thresholds]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Constants for discounts and thesholds are defined in main/DynFlags,
all of form ufXxxx. They are:
ufCreationThreshold
At a definition site, if the unfolding is bigger than this, we
may discard it altogether
ufUseThreshold
At a call site, if the unfolding, less discounts, is smaller than
this, then it's small enough inline
ufKeenessFactor
Factor by which the discounts are multiplied before
subtracting from size
ufDictDiscount
The discount for each occurrence of a dictionary argument
as an argument of a class method. Should be pretty small
else big functions may get inlined
ufFunAppDiscount
Discount for a function argument that is applied. Quite
large, because if we inline we avoid the higher-order call.
ufDearOp
The size of a foreign call or not-dupable PrimOp
Note [Function applications]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In a function application (f a b)
- If 'f' is an argument to the function being analysed,
and there's at least one value arg, record a FunAppDiscount for f
- If the application if a PAP (arity > 2 in this example)
record a *result* discount (because inlining
with "extra" args in the call may mean that we now
get a saturated application)
Code for manipulating sizes
\begin{code}
data ExprSize = TooBig
| SizeIs FastInt
!(Bag (Id,Int))
FastInt
instance Outputable ExprSize where
ppr TooBig = ptext (sLit "TooBig")
ppr (SizeIs a _ c) = brackets (int (iBox a) <+> int (iBox c))
mkSizeIs :: FastInt -> FastInt -> Bag (Id, Int) -> FastInt -> ExprSize
mkSizeIs max n xs d | (n -# d) ># max = TooBig
| otherwise = SizeIs n xs d
maxSize :: ExprSize -> ExprSize -> ExprSize
maxSize TooBig _ = TooBig
maxSize _ TooBig = TooBig
maxSize s1@(SizeIs n1 _ _) s2@(SizeIs n2 _ _) | n1 ># n2 = s1
| otherwise = s2
sizeZero :: ExprSize
sizeN :: Int -> ExprSize
sizeZero = SizeIs (_ILIT(0)) emptyBag (_ILIT(0))
sizeN n = SizeIs (iUnbox n) emptyBag (_ILIT(0))
\end{code}
%************************************************************************
%* *
\subsection[considerUnfolding]{Given all the info, do (not) do the unfolding}
%* *
%************************************************************************
We use 'couldBeSmallEnoughToInline' 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 :: DynFlags -> Int -> CoreExpr -> Bool
couldBeSmallEnoughToInline dflags threshold rhs
= case sizeExpr dflags (iUnbox threshold) [] body of
TooBig -> False
_ -> True
where
(_, body) = collectBinders rhs
smallEnoughToInline :: DynFlags -> Unfolding -> Bool
smallEnoughToInline dflags (CoreUnfolding {uf_guidance = UnfIfGoodArgs {ug_size = size}})
= size <= ufUseThreshold dflags
smallEnoughToInline _ _
= False
certainlyWillInline :: DynFlags -> Unfolding -> Bool
certainlyWillInline dflags (CoreUnfolding { uf_arity = n_vals, uf_guidance = guidance })
= case guidance of
UnfNever -> False
UnfWhen {} -> True
UnfIfGoodArgs { ug_size = size}
-> n_vals > 0
&& size (10 * (n_vals +1)) <= ufUseThreshold dflags
certainlyWillInline _ _
= False
\end{code}
Note [certainlyWillInline: be caseful of thunks]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Don't claim that thunks will certainly inline, because that risks work
duplication. Even if the work duplication is not great (eg is_cheap
holds), it can make a big difference in an inner loop In Trac #5623 we
found that the WorkWrap phase thought that
y = case x of F# v -> F# (v +# v)
was certainlyWillInline, so the addition got duplicated.
%************************************************************************
%* *
\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
-> Id
-> Bool
-> Bool
-> [ArgSummary]
-> CallCtxt
-> Maybe CoreExpr
instance Outputable ArgSummary where
ppr TrivArg = ptext (sLit "TrivArg")
ppr NonTrivArg = ptext (sLit "NonTrivArg")
ppr ValueArg = ptext (sLit "ValueArg")
data CallCtxt
= BoringCtxt
| RhsCtxt
| DiscArgCtxt
| RuleArgCtxt
| ValAppCtxt
| CaseCtxt
instance Outputable CallCtxt where
ppr CaseCtxt = ptext (sLit "CaseCtxt")
ppr ValAppCtxt = ptext (sLit "ValAppCtxt")
ppr BoringCtxt = ptext (sLit "BoringCtxt")
ppr RhsCtxt = ptext (sLit "RhsCtxt")
ppr DiscArgCtxt = ptext (sLit "DiscArgCtxt")
ppr RuleArgCtxt = ptext (sLit "RuleArgCtxt")
callSiteInline dflags id active_unfolding lone_variable arg_infos cont_info
= case idUnfolding id of
CoreUnfolding { uf_tmpl = unf_template, uf_is_top = is_top
, uf_is_work_free = is_wf, uf_arity = uf_arity
, uf_guidance = guidance, uf_expandable = is_exp }
| active_unfolding -> tryUnfolding dflags id lone_variable
arg_infos cont_info unf_template is_top
is_wf is_exp uf_arity guidance
| dopt Opt_D_dump_inlinings dflags && dopt Opt_D_verbose_core2core dflags
-> pprTrace "Inactive unfolding:" (ppr id) Nothing
| otherwise -> Nothing
NoUnfolding -> Nothing
OtherCon {} -> Nothing
DFunUnfolding {} -> Nothing
tryUnfolding :: DynFlags -> Id -> Bool -> [ArgSummary] -> CallCtxt
-> CoreExpr -> Bool -> Bool -> Bool -> Arity -> UnfoldingGuidance
-> Maybe CoreExpr
tryUnfolding dflags id lone_variable
arg_infos cont_info unf_template is_top
is_wf is_exp uf_arity guidance
| dopt Opt_D_dump_inlinings dflags && dopt Opt_D_verbose_core2core dflags
= pprTrace ("Considering inlining: " ++ showSDocDump dflags (ppr id))
(vcat [text "arg infos" <+> ppr arg_infos,
text "uf arity" <+> ppr uf_arity,
text "interesting continuation" <+> ppr cont_info,
text "some_benefit" <+> ppr some_benefit,
text "is exp:" <+> ppr is_exp,
text "is work-free:" <+> ppr is_wf,
text "guidance" <+> ppr guidance,
extra_doc,
text "ANSWER =" <+> if yes_or_no then text "YES" else text "NO"])
result
| otherwise = result
where
n_val_args = length arg_infos
saturated = n_val_args >= uf_arity
cont_info' | n_val_args > uf_arity = ValAppCtxt
| otherwise = cont_info
result | yes_or_no = Just unf_template
| otherwise = Nothing
interesting_args = any nonTriv arg_infos
some_benefit
| not saturated = interesting_args
| otherwise = interesting_args
|| interesting_call
interesting_call
= case cont_info' of
CaseCtxt -> not (lone_variable && is_wf)
ValAppCtxt -> True
RuleArgCtxt -> uf_arity > 0
DiscArgCtxt -> uf_arity > 0
RhsCtxt -> uf_arity > 0
_ -> not is_top && uf_arity > 0
(yes_or_no, extra_doc)
= case guidance of
UnfNever -> (False, empty)
UnfWhen unsat_ok boring_ok
-> (enough_args && (boring_ok || some_benefit), empty )
where
enough_args = saturated || (unsat_ok && n_val_args > 0)
UnfIfGoodArgs { ug_args = arg_discounts, ug_res = res_discount, ug_size = size }
-> ( is_wf && some_benefit && small_enough
, (text "discounted size =" <+> int discounted_size) )
where
discounted_size = size discount
small_enough = discounted_size <= ufUseThreshold dflags
discount = computeDiscount dflags uf_arity arg_discounts
res_discount arg_infos cont_info'
\end{code}
Note [Unfold into lazy contexts], Note [RHS of lets]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When the call is the argument of a function with a RULE, or the RHS of a let,
we are a little bit keener to inline. For example
f y = (y,y,y)
g y = let x = f y in ...(case x of (a,b,c) -> ...) ...
We'd inline 'f' if the call was in a case context, and it kind-of-is,
only we can't see it. Also
x = f v
could be expensive whereas
x = case v of (a,b) -> a
is patently cheap and may allow more eta expansion.
So we treat the RHS of a let as not-totally-boring.
Note [Unsaturated applications]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When a call is not saturated, we *still* inline if one of the
arguments has interesting structure. That's sometimes very important.
A good example is the Ord instance for Bool in Base:
Rec {
$fOrdBool =GHC.Classes.D:Ord
@ Bool
...
$cmin_ajX
$cmin_ajX [Occ=LoopBreaker] :: Bool -> Bool -> Bool
$cmin_ajX = GHC.Classes.$dmmin @ Bool $fOrdBool
}
But the defn of GHC.Classes.$dmmin is:
$dmmin :: forall a. GHC.Classes.Ord a => a -> a -> a
{- Arity: 3, HasNoCafRefs, Strictness: SLL,
Unfolding: (\ @ a $dOrd :: GHC.Classes.Ord a x :: a y :: a ->
case @ a GHC.Classes.<= @ a $dOrd x y of wild {
GHC.Types.False -> y GHC.Types.True -> x }) -}
We *really* want to inline $dmmin, even though it has arity 3, in
order to unravel the recursion.
Note [Things to watch]
~~~~~~~~~~~~~~~~~~~~~~
* { y = I# 3; x = y `cast` co; ...case (x `cast` co) of ... }
Assume x is exported, so not inlined unconditionally.
Then we want x to inline unconditionally; no reason for it
not to, and doing so avoids an indirection.
* { x = I# 3; ....f x.... }
Make sure that x does not inline unconditionally!
Lest we get extra allocation.
Note [Inlining an InlineRule]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
An InlineRules is used for
(a) programmer INLINE pragmas
(b) inlinings from worker/wrapper
For (a) the RHS may be large, and our contract is that we *only* inline
when the function is applied to all the arguments on the LHS of the
source-code defn. (The uf_arity in the rule.)
However for worker/wrapper it may be worth inlining even if the
arity is not satisfied (as we do in the CoreUnfolding case) so we don't
require saturation.
Note [Nested functions]
~~~~~~~~~~~~~~~~~~~~~~~
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.
Note [Cast then apply]
~~~~~~~~~~~~~~~~~~~~~~
Consider
myIndex = __inline_me ( (/\a. ) |> co )
co :: (forall a. a -> a) ~ (forall a. T a)
... /\a.\x. case ((myIndex a) |> sym co) x of { ... } ...
We need to inline myIndex to unravel this; but the actual call (myIndex a) has
no value arguments. The ValAppCtxt gives it enough incentive to inline.
Note [Inlining in ArgCtxt]
~~~~~~~~~~~~~~~~~~~~~~~~~~
The condition (arity > 0) here is very important, because otherwise
we end up inlining top-level stuff into useless places; eg
x = I# 3#
f = \y. g x
This can make a very big difference: it adds 16% to nofib 'integer' allocs,
and 20% to 'power'.
At one stage I replaced this condition by 'True' (leading to the above
slow-down). The motivation was test eyeball/inline1.hs; but that seems
to work ok now.
NOTE: arguably, we should inline in ArgCtxt only if the result of the
call is at least CONLIKE. At least for the cases where we use ArgCtxt
for the RHS of a 'let', we only profit from the inlining if we get a
CONLIKE thing (modulo lets).
Note [Lone variables] See also Note [Interaction of exprIsWorkFree and lone variables]
~~~~~~~~~~~~~~~~~~~~~ which appears below
The "lone-variable" case is important. I spent ages messing about
with unsatisfactory varaints, but this is nice. The idea is that if a
variable appears all alone
as an arg of lazy fn, or rhs BoringCtxt
as scrutinee of a case CaseCtxt
as arg of a fn ArgCtxt
AND
it is bound to a cheap expression
then we should not inline it (unless there is some other reason,
e.g. is is the sole occurrence). That is what is happening at
the use of 'lone_variable' in 'interesting_call'.
Why? At least in the case-scrutinee situation, turning
let x = (a,b) in case x of y -> ...
into
let x = (a,b) in case (a,b) of y -> ...
and thence to
let x = (a,b) in let y = (a,b) in ...
is bad if the binding for x will remain.
Another example: I discovered that strings
were getting inlined straight back into applications of 'error'
because the latter is strict.
s = "foo"
f = \x -> ...(error s)...
Fundamentally such contexts should not encourage inlining because the
context can ``see'' the unfolding of the variable (e.g. case or a
RULE) so there's no gain. If the thing is bound to a value.
However, watch out:
* Consider this:
foo = _inline_ (\n. [n])
bar = _inline_ (foo 20)
baz = \n. case bar of { (m:_) -> m + n }
Here we really want to inline 'bar' so that we can inline 'foo'
and the whole thing unravels as it should obviously do. This is
important: in the NDP project, 'bar' generates a closure data
structure rather than a list.
So the non-inlining of lone_variables should only apply if the
unfolding is regarded as cheap; because that is when exprIsConApp_maybe
looks through the unfolding. Hence the "&& is_wf" in the
InlineRule branch.
* Even a type application or coercion isn't a lone variable.
Consider
case $fMonadST @ RealWorld of { :DMonad a b c -> c }
We had better inline that sucker! The case won't see through it.
For now, I'm treating treating a variable applied to types
in a *lazy* context "lone". The motivating example was
f = /\a. \x. BIG
g = /\a. \y. h (f a)
There's no advantage in inlining f here, and perhaps
a significant disadvantage. Hence some_val_args in the Stop case
Note [Interaction of exprIsWorkFree and lone variables]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The lone-variable test says "don't inline if a case expression
scrutines a lone variable whose unfolding is cheap". It's very
important that, under these circumstances, exprIsConApp_maybe
can spot a constructor application. So, for example, we don't
consider
let x = e in (x,x)
to be cheap, and that's good because exprIsConApp_maybe doesn't
think that expression is a constructor application.
In the 'not (lone_variable && is_wf)' test, I used to test is_value
rather than is_wf, which was utterly wrong, because the above
expression responds True to exprIsHNF, which is what sets is_value.
This kind of thing can occur if you have
{-# INLINE foo #-}
foo = let x = e in (x,x)
which Roman did.
\begin{code}
computeDiscount :: DynFlags -> Arity -> [Int] -> Int -> [ArgSummary] -> CallCtxt
-> Int
computeDiscount dflags uf_arity arg_discounts res_discount arg_infos cont_info
= 10
+ 10 * length (take uf_arity arg_infos)
+ round (ufKeenessFactor dflags *
fromIntegral (arg_discount + res_discount'))
where
arg_discount = sum (zipWith mk_arg_discount arg_discounts arg_infos)
mk_arg_discount _ TrivArg = 0
mk_arg_discount _ NonTrivArg = 10
mk_arg_discount discount ValueArg = discount
res_discount' = case cont_info of
BoringCtxt -> 0
CaseCtxt -> res_discount
ValAppCtxt -> res_discount
_ -> 40 `min` res_discount
\end{code}
%************************************************************************
%* *
Interesting arguments
%* *
%************************************************************************
Note [Interesting arguments]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
An argument is interesting if it deserves a discount for unfoldings
with a discount in that argument position. The idea is to avoid
unfolding a function that is applied only to variables that have no
unfolding (i.e. they are probably lambda bound): f x y z There is
little point in inlining f here.
Generally, *values* (like (C a b) and (\x.e)) deserve discounts. But
we must look through lets, eg (let x = e in C a b), because the let will
float, exposing the value, if we inline. That makes it different to
exprIsHNF.
Before 2009 we said it was interesting if the argument had *any* structure
at all; i.e. (hasSomeUnfolding v). But does too much inlining; see Trac #3016.
But we don't regard (f x y) as interesting, unless f is unsaturated.
If it's saturated and f hasn't inlined, then it's probably not going
to now!
Note [Conlike is interesting]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider
f d = ...((*) d x y)...
... f (df d')...
where df is con-like. Then we'd really like to inline 'f' so that the
rule for (*) (df d) can fire. To do this
a) we give a discount for being an argument of a class-op (eg (*) d)
b) we say that a con-like argument (eg (df d)) is interesting
\begin{code}
data ArgSummary = TrivArg
| NonTrivArg
| ValueArg
interestingArg :: CoreExpr -> ArgSummary
interestingArg e = go e 0
where
go (Lit {}) _ = ValueArg
go (Var v) n
| isConLikeId v = ValueArg
| idArity v > n = ValueArg
| n > 0 = NonTrivArg
| conlike_unfolding = ValueArg
| otherwise = TrivArg
where
conlike_unfolding = isConLikeUnfolding (idUnfolding v)
go (Type _) _ = TrivArg
go (Coercion _) _ = TrivArg
go (App fn (Type _)) n = go fn n
go (App fn (Coercion _)) n = go fn n
go (App fn _) n = go fn (n+1)
go (Tick _ a) n = go a n
go (Cast e _) n = go e n
go (Lam v e) n
| isTyVar v = go e n
| n>0 = go e (n1)
| otherwise = ValueArg
go (Let _ e) n = case go e n of { ValueArg -> ValueArg; _ -> NonTrivArg }
go (Case {}) _ = NonTrivArg
nonTriv :: ArgSummary -> Bool
nonTriv TrivArg = False
nonTriv _ = True
\end{code}