October 2014 - ghc-devs

emailing to Trac?
by Richard Eisenberg 29 Oct '14

29 Oct '14

Hi devs, There's a feature I've wanted for some time, and I don't see a good reason not to ask: Is it possible to email comments into Trac? My commute involves a ~40 minute train ride, which I generally can use productively despite no connection. One of the biggest annoyances, though, is that I can't respond to Trac comments while on the train. It would be great if there were an email address to write to that would post a comment. Of course, it's easy enough for me to write and locally cache a response, but then I have to remember to flush the cache and it's all a bit annoying. Does Trac have this feature? Is it perhaps already enabled but not advertised? Thanks! Richard

2 1

Is USE_REPORT_PRELUDE still useful?
by David Feuer 29 Oct '14

29 Oct '14

A lot of code in GHC.List and perhaps elsewhere compiles differently depending on whether USE_REPORT_PRELUDE is defined. Not all code differing from the Prelude implementation. Furthermore, I don't know to what extent, if any, such code actually works these days. Some of it certainly was not usable for *years* because GHC.List did not import GHC.Num. Should we 1. Convert all those code blocks to comments? 2. Go through everything, check it to make sure it's written as in the Prelude or has an alternative block, and then actually set up all the infrastructure so that works? 3. Leave it alone? My general inclination is to go to 1. I don't *really* like option 3 for four reasons: a. It leaves untouched code to rot b. It forces us to run CPP on files that otherwise have no need for it. c. It interrupts the flow of the code with stuff that *looks* like real code (and is highlighted as such) but is actually inactive. d. It's not hard to accidentally move code into or out of the #ifdef blocks.

4 4

Call Arity, oneShot or both
by Joachim Breitner 28 Oct '14

28 Oct '14

Hi, some months ago I tried to make foldl a good consumer in the common case. The starting point is always to write foldl k a xs = foldr (\v f a -> f (v `k` a)) id xs a and then somehow make GHC produce good code with this. I came up with two solutions: A more sophisticated compiler analysis (Call Arity), or an explicit annotation in the form of foldlB k a xs = foldr (\v f -> oneShot (\a -> f (v `k` a))) id xs a where oneShot :: (a -> b) -> (a -> b) is a built-in function, semantically the identity, but telling the compiler that it can assume that the (\a -> ...) is called at most once. Back then, we decided to use Call Arity, on the grounds that it might improve other code as well, despite not having a lot of evindence that this may happen. Then recently David Feuer built on my work by making more functions fuse, including functions like scanl that are not built on foldl, but benefit from the same analysis. This supports the usefulness of Call Arity. But he also found cases where Call Arity is not powerful enough and GHC would produce bad code. So I wanted to properly compare Call Arity with the oneShot approach. Based on todays master (0855b249), I disabled Call Arity and changed the definitions of foldl, foldl', scanl and scanl' to use oneShot, and ran nofib. The result are mixed. With the current code, the oneShot machinery does not always work as expected: Program Size Allocs Runtime Elapsed TotalMem Min -0.1% -1.5% -2.8% -2.8% -5.8% Max +0.4% +4.7% +5.8% +5.6% +5.4% Geometric Mean -0.0% +0.1% +0.3% +0.3% +0.1% The biggest loser is calendar, which uses scanl. I am not fully sure what went wrong here: Either the one-shot annotation on the lambda’s variable got lost somewhere in the pipeline, or despite it being there, the normal arity analysis did not use it. But there is also a winner, fft2, with -1.5% allocations. Here Call Arity was not good enough, but oneShot did the jobs. There is also the option of combining both. Then we do not get the regression, but still the improvement for fft2: Min -0.1% -1.5% -3.9% -3.8% -5.8% Max +0.2% +0.1% +6.4% +6.3% +13.1% Geometric Mean -0.0% -0.0% +0.0% +0.0% +0.1% The oneShot code is on the branch wip/oneShot. The changes are clearly not ready to be merged. In particular, there is the question of how to best keep the oneShot annotation in the unfoldings: The isOneShotLambda flag is currently not stored in the interface. I work around this by making sure that the oneShot function is never inlined in unfoldings, but maybe it would be better to serialize the isOneShotLambda flag in interfaces, which might have other good effects? If we want as much performance as possible, we should simply include both approaches. But there might be other things to consider... so not sure what the best thing to do is. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata(a)joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

3 8

Re: Call Arity, oneShot or both
by David Feuer 28 Oct '14

28 Oct '14

Simon Peyton Jones wrote: > > But since it is plausible that there are cases out there where it might > help, even if just a little, we could go forward ?unless the > implementation becomes ugly. > Based on our experience with Call Arity, it's much more likely that it will help a lot in a few cases than that it will help a little in a lot of cases. David

1 0

Re: [commit: ghc] master: Implementation of hsig (module signatures), per #9252 (aa47995)
by Joachim Breitner 28 Oct '14

28 Oct '14

Hi Edwardd, Am Freitag, den 24.10.2014, 23:47 +0000 schrieb git(a)git.haskell.org: > >--------------------------------------------------------------- > > commit aa4799534225e3fc6bbde0d5e5eeab8868cc3111 > Author: Edward Z. Yang <ezyang(a)cs.stanford.edu> > Date: Thu Aug 7 18:32:12 2014 +0100 > > Implementation of hsig (module signatures), per #9252 this breaks a few test cases: Actual stderr output differs from expected: --- ./ghci/scripts/T5979.stderr 2014-10-24 23:49:33.395524791 +0000 +++ ./ghci/scripts/T5979.run.stderr 2014-10-25 00:24:08.934279006 +0000 @@ -2,6 +2,6 @@ <no location info>: Could not find module â€˜Control.Monad.Trans.Stateâ€™ Perhaps you meant - Control.Monad.Trans.State (from transformers-0.4.1.0@trans_GjLVjHaAO8fEGf8lChbngr) - Control.Monad.Trans.Class (from transformers-0.4.1.0@trans_GjLVjHaAO8fEGf8lChbngr) - Control.Monad.Trans.Cont (from transformers-0.4.1.0@trans_GjLVjHaAO8fEGf8lChbngr) + Control.Monad.Trans.State (from transformers-0.4.1.0@trans_5jw4w9yTgmZ89ByuixDAKP) + Control.Monad.Trans.Class (from transformers-0.4.1.0@trans_5jw4w9yTgmZ89ByuixDAKP) + Control.Monad.Trans.Cont (from transformers-0.4.1.0@trans_5jw4w9yTgmZ89ByuixDAKP) *** unexpected failure for T5979(ghci) Actual stdout output differs from expected: --- ./safeHaskell/check/pkg01/safePkg01.stdout 2014-10-24 23:49:33.705509654 +0000 +++ ./safeHaskell/check/pkg01/safePkg01.run.stdout 2014-10-25 00:19:17.451490530 +0000 @@ -29,17 +29,17 @@ require own pkg trusted: True M_SafePkg6 -package dependencies: array-0.5.0.1@array_5q713e1nmXtAgNRa542ahu +package dependencies: array-0.5.0.1@array_GX4NwjS8xZkC2ZPtjgwhnz trusted: trustworthy require own pkg trusted: False M_SafePkg7 -package dependencies: array-0.5.0.1@array_5q713e1nmXtAgNRa542ahu +package dependencies: array-0.5.0.1@array_GX4NwjS8xZkC2ZPtjgwhnz trusted: safe require own pkg trusted: False M_SafePkg8 -package dependencies: array-0.5.0.1@array_5q713e1nmXtAgNRa542ahu +package dependencies: array-0.5.0.1@array_GX4NwjS8xZkC2ZPtjgwhnz trusted: trustworthy require own pkg trusted: False *** unexpected failure for safePkg01(normal) https://s3.amazonaws.com/archive.travis-ci.org/jobs/38981598/log.txt It seems you need to adjust the testsuite to remove these hashes before comparing the output. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata(a)joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

2 1

RE: [commit: ghc] wip/oneShot: Avoid inlining oneShot in unfoldings (16066a1)
by Simon Peyton Jones 27 Oct '14

27 Oct '14

| Avoid inlining oneShot in unfoldings I'm sure there's a reason for this but, again, please say what is, lest I accidentally reverse it in 3 years time. Simon | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf Of | git(a)git.haskell.org | Sent: 25 October 2014 13:59 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] wip/oneShot: Avoid inlining oneShot in unfoldings | (16066a1) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : wip/oneShot | Link : | http://ghc.haskell.org/trac/ghc/changeset/16066a165dd7245be58d0fbee265e13 | fb21e2aed/ghc | | >--------------------------------------------------------------- | | commit 16066a165dd7245be58d0fbee265e13fb21e2aed | Author: Joachim Breitner <mail(a)joachim-breitner.de> | Date: Sat Oct 25 14:46:27 2014 +0200 | | Avoid inlining oneShot in unfoldings | | | >--------------------------------------------------------------- | | 16066a165dd7245be58d0fbee265e13fb21e2aed | compiler/basicTypes/MkId.lhs | 4 ++-- | 1 file changed, 2 insertions(+), 2 deletions(-) | | diff --git a/compiler/basicTypes/MkId.lhs b/compiler/basicTypes/MkId.lhs | index 05dcdd5..31cd426 100644 | --- a/compiler/basicTypes/MkId.lhs | +++ b/compiler/basicTypes/MkId.lhs | @@ -1124,8 +1124,8 @@ lazyId = pcMiscPrelId lazyIdName ty info | oneShotId :: Id | oneShotId = pcMiscPrelId oneShotName ty info | where | - info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma | - `setUnfoldingInfo` mkCompulsoryUnfolding rhs | + info = noCafIdInfo -- `setInlinePragInfo` alwaysInlinePragma | + `setUnfoldingInfo` mkWwInlineRule rhs 1 | ty = mkForAllTys [alphaTyVar, betaTyVar] (mkFunTy fun_ty fun_ty) | fun_ty = mkFunTy alphaTy betaTy | [body, x] = mkTemplateLocals [fun_ty, alphaTy] | | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | http://www.haskell.org/mailman/listinfo/ghc-commits

2 2

GHC on mingw-w64 / gcc 4.8.3 on Windows
by Gintautas Miliauskas 27 Oct '14

27 Oct '14

The patch to migrate to mingw-w64 / gcc 4.8.3 is finally starting to look reasonable (Phab:D339 <https://phabricator.haskell.org/D339>). Whew! The part about downloading the tarball was relatively straightforward, but the mingw migration has been really tricky, in part due to the gcc version bump, but mostly, I think, due to use of mingw-w64 on i686 as well (which is a supported configuration). In particular, "#define _MSVCRT_ 1" took ages to figure out after digging through heaps of headers and weird errors. I am not sure I would have signed on if I had an idea of how much effort this needed... Jeez. On the bright side, now that we have standartised on mingw-w64, further gcc version upgrades should be relatively straightforward. Could people with some experience in Windows matters take a look at the patch? In particular, rts/Linker.c has been a problem; my fixes there seem to work but I have no idea if the approach is right (the whole file does seem like a grabbag of ad-hocness though). The patch also could use some more testing (both for x86 and for x86-64). There are still some validate.sh failures (although I am actually seeing fewer than with the legacy setup), some of which could probably be fixed easily. Some additional eyes and fingers on keyboards would be very welcome there. If you want to experiment with the patch, make sure to also patch in the related changes listed in the last comment on the Phab page. -- Gintautas Miliauskas

4 5

Re: Call Arity, oneShot, or both
by David Feuer 27 Oct '14

27 Oct '14

Joachim Breitner כתב That would be great! But do we have evidence of this user-written code > that benefits? So far I have only seen relevant improvement due to > list-fusion a left-foldish function. > I was under the impression that the transformation was much more general than that, improving various recursive forms. Was I wrong? But aside from that, I would be astonished if the library authors were the only ones writing left-accumulating folds! David

2 1

RE: [commit: ghc] wip/oneShot: Use oneShot in the definition of foldl etc. (6f101d2)
by Simon Peyton Jones 26 Oct '14

26 Oct '14

No Notes! Surely, surely it deserves one! Simon | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf Of | git(a)git.haskell.org | Sent: 25 October 2014 11:27 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] wip/oneShot: Use oneShot in the definition of | foldl etc. (6f101d2) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : wip/oneShot | Link : | http://ghc.haskell.org/trac/ghc/changeset/6f101d20805fb52de0423bc8beab373 | b94bd4a7d/ghc | | >--------------------------------------------------------------- | | commit 6f101d20805fb52de0423bc8beab373b94bd4a7d | Author: Joachim Breitner <mail(a)joachim-breitner.de> | Date: Sat Oct 25 12:27:06 2014 +0200 | | Use oneShot in the definition of foldl etc. | | | >--------------------------------------------------------------- | | 6f101d20805fb52de0423bc8beab373b94bd4a7d | libraries/base/Data/OldList.hs | 5 +++-- | libraries/base/GHC/List.lhs | 6 +++--- | 2 files changed, 6 insertions(+), 5 deletions(-) | | diff --git a/libraries/base/Data/OldList.hs | b/libraries/base/Data/OldList.hs | index 0e6709e..75fba35 100644 | --- a/libraries/base/Data/OldList.hs | +++ b/libraries/base/Data/OldList.hs | @@ -499,7 +499,8 @@ pairWithNil x = (x, []) | | mapAccumLF :: (acc -> x -> (acc, y)) -> x -> (acc -> (acc, [y])) -> acc | -> (acc, [y]) | {-# INLINE [0] mapAccumLF #-} | -mapAccumLF f = \x r s -> let (s', y) = f s x | +mapAccumLF f = \x r -> oneShot $ \s -> | + let (s', y) = f s x | (s'', ys) = r s' | in (s'', y:ys) | | @@ -1058,7 +1059,7 @@ unfoldr f b0 = build (\c n -> | | -- | A strict version of 'foldl'. | foldl' :: forall a b . (b -> a -> b) -> b -> [a] -> b | -foldl' k z0 xs = foldr (\(v::a) (fn::b->b) (z::b) -> z `seq` fn (k z v)) | (id :: b -> b) xs z0 | +foldl' k z0 xs = foldr (\(v::a) (fn::b->b) -> oneShot (\(z::b) -> z | `seq` fn (k z v))) (id :: b -> b) xs z0 | -- Implementing foldl' via foldr is only a good idea if the compiler can | optimize | -- the resulting code (eta-expand the recursive "go"), so this needs - | fcall-arity! | -- Also see #7994 | diff --git a/libraries/base/GHC/List.lhs b/libraries/base/GHC/List.lhs | index 2d01678..c7a0cb3 100644 | --- a/libraries/base/GHC/List.lhs | +++ b/libraries/base/GHC/List.lhs | @@ -186,7 +186,7 @@ filterFB c p x r | p x = x `c` r | | foldl :: forall a b. (b -> a -> b) -> b -> [a] -> b | {-# INLINE foldl #-} | -foldl k z0 xs = foldr (\(v::a) (fn::b->b) (z::b) -> fn (k z v)) (id :: b | -> b) xs z0 | +foldl k z0 xs = foldr (\(v::a) (fn::b->b) -> oneShot (\(z::b) -> fn (k z | v))) (id :: b -> b) xs z0 | -- Implementing foldl via foldr is only a good idea if the compiler can | optimize | -- the resulting code (eta-expand the recursive "go"), so this needs - | fcall-arity! | -- Also see #7994 | @@ -221,7 +221,7 @@ scanl = scanlGo | | {-# INLINE [0] scanlFB #-} | scanlFB :: (b -> a -> b) -> (b -> c -> c) -> a -> (b -> c) -> b -> c | -scanlFB f c = \b g x -> let b' = f x b in b' `c` g b' | +scanlFB f c = \b g -> oneShot (\x -> let b' = f x b in b' `c` g b') | | {-# INLINE [0] constScanl #-} | constScanl :: a -> b -> a | @@ -258,7 +258,7 @@ scanl' = scanlGo' | | {-# INLINE [0] scanlFB' #-} | scanlFB' :: (b -> a -> b) -> (b -> c -> c) -> a -> (b -> c) -> b -> c | -scanlFB' f c = \b g x -> let b' = f x b in b' `seq` b' `c` g b' | +scanlFB' f c = \b g -> oneShot (\x -> let b' = f x b in b' `seq` b' `c` | g b') | | {-# INLINE [0] flipSeqScanl' #-} | flipSeqScanl' :: a -> b -> a | | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | http://www.haskell.org/mailman/listinfo/ghc-commits

1 0

RE: [commit: ghc] wip/oneShot: Add GHC.Prim.oneShot (955d9f5)
by Simon Peyton Jones 26 Oct '14

26 Oct '14

Is all this one-shot stuff carefully documented somewhere? I think it's basically a good plan, but I see no Notes, alas. Simon | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf Of | git(a)git.haskell.org | Sent: 25 October 2014 11:27 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] wip/oneShot: Add GHC.Prim.oneShot (955d9f5) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : wip/oneShot | Link : | http://ghc.haskell.org/trac/ghc/changeset/955d9f53b6c934585a90423dfc95d86 | d8a129908/ghc | | >--------------------------------------------------------------- | | commit 955d9f53b6c934585a90423dfc95d86d8a129908 | Author: Joachim Breitner <mail(a)joachim-breitner.de> | Date: Sun Jan 26 11:36:23 2014 +0000 | | Add GHC.Prim.oneShot | | Conflicts: | compiler/basicTypes/MkId.lhs | | | >--------------------------------------------------------------- | | 955d9f53b6c934585a90423dfc95d86d8a129908 | compiler/basicTypes/MkId.lhs | 17 +++++++++++++++-- | compiler/prelude/PrelNames.lhs | 3 ++- | 2 files changed, 17 insertions(+), 3 deletions(-) | | diff --git a/compiler/basicTypes/MkId.lhs b/compiler/basicTypes/MkId.lhs | index bf1c199..05dcdd5 100644 | --- a/compiler/basicTypes/MkId.lhs | +++ b/compiler/basicTypes/MkId.lhs | @@ -135,7 +135,8 @@ ghcPrimIds | seqId, | magicDictId, | coerceId, | - proxyHashId | + proxyHashId, | + oneShotId | ] | \end{code} | | @@ -1016,7 +1017,7 @@ another gun with which to shoot yourself in the | foot. | \begin{code} | lazyIdName, unsafeCoerceName, nullAddrName, seqName, | realWorldName, voidPrimIdName, coercionTokenName, | - magicDictName, coerceName, proxyName, dollarName :: Name | + magicDictName, coerceName, proxyName, dollarName, oneShotName :: Name | unsafeCoerceName = mkWiredInIdName gHC_PRIM (fsLit "unsafeCoerce#") | unsafeCoerceIdKey unsafeCoerceId | nullAddrName = mkWiredInIdName gHC_PRIM (fsLit "nullAddr#") | nullAddrIdKey nullAddrId | seqName = mkWiredInIdName gHC_PRIM (fsLit "seq") | seqIdKey seqId | @@ -1028,6 +1029,7 @@ magicDictName = mkWiredInIdName gHC_PRIM | (fsLit "magicDict") magicDict | coerceName = mkWiredInIdName gHC_PRIM (fsLit "coerce") | coerceKey coerceId | proxyName = mkWiredInIdName gHC_PRIM (fsLit "proxy#") | proxyHashKey proxyHashId | dollarName = mkWiredInIdName gHC_BASE (fsLit "$") | dollarIdKey dollarId | +oneShotName = mkWiredInIdName gHC_PRIM (fsLit "oneShot") | oneShotKey oneShotId | \end{code} | | \begin{code} | @@ -1119,6 +1121,17 @@ lazyId = pcMiscPrelId lazyIdName ty info | info = noCafIdInfo | ty = mkForAllTys [alphaTyVar] (mkFunTy alphaTy alphaTy) | | +oneShotId :: Id | +oneShotId = pcMiscPrelId oneShotName ty info | + where | + info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma | + `setUnfoldingInfo` mkCompulsoryUnfolding rhs | + ty = mkForAllTys [alphaTyVar, betaTyVar] (mkFunTy fun_ty fun_ty) | + fun_ty = mkFunTy alphaTy betaTy | + [body, x] = mkTemplateLocals [fun_ty, alphaTy] | + x' = setOneShotLambda x | + rhs = mkLams [alphaTyVar, betaTyVar, body, x'] $ Var body `App` Var | x | + | | ------------------------------------------------------------------------ | -------- | magicDictId :: Id -- See Note [magicDictId magic] | diff --git a/compiler/prelude/PrelNames.lhs | b/compiler/prelude/PrelNames.lhs | index e053b11..e2ade33 100644 | --- a/compiler/prelude/PrelNames.lhs | +++ b/compiler/prelude/PrelNames.lhs | @@ -1682,10 +1682,11 @@ rootMainKey, runMainKey :: Unique | rootMainKey = mkPreludeMiscIdUnique 101 | runMainKey = mkPreludeMiscIdUnique 102 | | -thenIOIdKey, lazyIdKey, assertErrorIdKey :: Unique | +thenIOIdKey, lazyIdKey, assertErrorIdKey, oneShotKey :: Unique | thenIOIdKey = mkPreludeMiscIdUnique 103 | lazyIdKey = mkPreludeMiscIdUnique 104 | assertErrorIdKey = mkPreludeMiscIdUnique 105 | +oneShotKey = mkPreludeMiscIdUnique 106 | | breakpointIdKey, breakpointCondIdKey, breakpointAutoIdKey, | breakpointJumpIdKey, breakpointCondJumpIdKey, | | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | http://www.haskell.org/mailman/listinfo/ghc-commits

1 0