- Haskell-Cafe - Haskell.org

[ANN] fclabels 2.0
by Sebastiaan Visser 01 Oct '13

01 Oct '13

Hello everyone, I’m excited to announce a brand new major version of fclabels. This library provides first class labels that can act as bidirectional record fields. The labels are implemented as lenses. This new version contains some interesting new features like support for polymorphic updates and lens derivation for GADTs. Furthermore the API has been cleanup significantly and should be easier to work with. You can read everything about this release at: http://fvisser.nl/post/2013/okt/1/fclabels-2.0.html All feedback more than welcome! -Sebastiaan -- On Hackage: http://hackage.haskell.org/package/fclabels-2.0 Source on github: https://github.com/sebastiaanvisser/fclabels Full changelog: https://github.com/sebastiaanvisser/fclabels/blob/master/CHANGELOG

1 0

The RTSOPTS "-qm" flag's impact on runtime
by Iustin Pop 30 Sep '13

30 Sep '13

Hi all, I found an interesting case where the rtsopts -qm flag makes a significant difference in runtime (~50x). This is using GHC 7.6.3, llvm 3.4, program compiled with "-threaded -O2 -fllvm" and a couple of language extension. Source is at http://benchmarksgame.alioth.debian.org/u64q/benchmark.php?test=chameneosre…, on the language shootout benchmarks. Running the code without -N results (on my computer) in around 4 seconds of runtime: $ time ./orig 6000000 … real 0m3.919s user 0m3.903s sys 0m0.010s This is reasonably consistent. Running -N4 (this is an 8-core machine) results in the surprising: $ time ./orig 6000000 +RTS -N4 … real 1m15.154s user 1m38.790s sys 2m7.947s The cores are all used very erratically (continuously changing 5%-20%-40%) and the overall cpu usage is ~27-28%. Note the surprising 2m7s of sys usage, which means the kernel is involved a lot… Note that removing the explicit forkOn and running with -N4 results in somewhat worse performance: real 2m6.548s user 2m13.470s sys 2m3.043s So in that sense the forkOn itself is not at fault. What I have found is that -qm is here a life saver: $ time ./orig 6000000 +RTS -N4 -qm real 0m2.773s user 0m5.610s sys 0m0.123s Adding -qa doesn't make a big difference. To summarise more runs (in terms of cpu used, user+sys): with forkOn: - -N4: 228s - -N4 -qa: 110s - -N4 -qm: 6s - -N4 -qm -qa: 6s without forkOn: - -N4: 253s - -N4 -qa: 252s - -N4 -qm: 5s - -N4 -qm -qa: 5s (Note that "without forkOn" is a bit slower in term of wall-clock, as the "with forkOn" version distributes the work a bit better, even if it uses overall a tiny bit more CPU.) So the question is, what does -qm actually do that it affects this benchmark so much (~50x)? (The docs are not very clear on it) And furthermore, could there be an heuristic inside the runtime such that automatic thread migration is suspended if threads are "over-migrated" (which is what I suppose happens here)? thanks for any explanations, iustin

1 0

Proposal: Partitionable goes somewhere + containers instances
by Ryan Newton 30 Sep '13

30 Sep '13

<subject change> On Sun, Sep 29, 2013 at 3:31 AM, Mike Izbicki <mike(a)izbicki.me> wrote: > I've got a Partitionable class that I've been using for this purpose: > > https://github.com/mikeizbicki/ConstraintKinds/blob/master/src/Control/Cons… > Mike -- Neat, that's a cool library! Edward -- ideally, where in the standard libraries should the Partitionable comonoid go? Btw, I'm not sure what the ideal return type for comappend is, given that it needs to be able to "bottom out". Mike, our partition function's list return type seems more reasonable. Or maybe something simple would be this: *class Partitionable t where* * partition :: t -> Maybe (t,t)* That is, at some point its not worth splitting and returns Nothing, and you'd better be able to deal with the 't' directly. So what I really want is for the *containers package to please get some kind of Partitionable instances! * Johan & others, I would be happy to provide a patch if the class can be agreed on. This is important because currently the balanced tree structure of Data.Set/Map is an *amazing and beneficial property* that is *not* exposed at all through the API. For example, it would be great to have a parallel traverse_ for Maps and Sets in the Par monad. The particular impetus is that our new and enhanced Par monad makes extensive use of Maps and Sets, both the pure, balanced ones, and lockfree/inplace ones based on concurrent skip lists: http://www.cs.indiana.edu/~rrnewton/haddock/lvish/ Alternatively, it would be ok if there were a "Data.Map.Internal" module that exposed the Bin/Tip, but I assume people would rather have a clean Partitionable instance... Best, -Ryan On Sun, Sep 29, 2013 at 3:31 AM, Mike Izbicki <mike(a)izbicki.me> wrote: > I've got a Partitionable class that I've been using for this purpose: > > > https://github.com/mikeizbicki/ConstraintKinds/blob/master/src/Control/Cons… > > The function called "parallel" in the HLearn library will automatically > parallelize any homomorphism from a Partionable to a Monoid. I > specifically use that to parallelize machine learning algorithms. > > I have two thoughts for better abstractions: > > 1) This Partitionable class is essentially a comonoid. By reversing the > arrows of mappend, we get: > > comappend :: a -> (a,a) > > By itself, this works well if the number of processors you have is a power > of two, but it needs some more fanciness to get things balanced properly > for other numbers of processors. I bet there's another algebraic structure > that would capture these other cases, but I'm not sure what it is. > > 2) I'm working with parallelizing tree structures right now (kd-trees, > cover trees, oct-trees, etc.). The real problem is not splitting the > number of data points equally (this is easy), but splitting the amount of > work equally. Some points take longer to process than others, and this > cannot be determined in advance. Therefore, an equal split of the data > points can result in one processor getting 25% of the work load, and the > second processor getting 75%. Some sort of lazy Partitionable class that > was aware of processor loads and didn't split data points until they were > needed would be ideal for this scenario. > > On Sat, Sep 28, 2013 at 6:46 PM, adam vogt <vogt.adam(a)gmail.com> wrote: > >> On Sat, Sep 28, 2013 at 1:09 PM, Ryan Newton <rrnewton(a)gmail.com> wrote: >> > Hi all, >> > >> > We all know and love Data.Foldable and are familiar with left folds and >> > right folds. But what you want in a parallel program is a balanced fold >> > over a tree. Fortunately, many of our datatypes (Sets, Maps) actually >> ARE >> > balanced trees. Hmm, but how do we expose that? >> >> Hi Ryan, >> >> At least for Data.Map, the Foldable instance seems to have a >> reasonably balanced fold called fold (or foldMap): >> >> > fold t = go t >> > where go (Bin _ _ v l r) = go l `mappend` (v `mappend` go r) >> >> This doesn't seem to be guaranteed though. For example ghc's derived >> instance writes the foldr only, so fold would be right-associated for >> a: >> >> > data T a = B (T a) (T a) | L a deriving (Foldable) >> >> Regards, >> Adam >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe >> > >

6 9

What class for splittable data / balanced-fold?
by Ryan Newton 30 Sep '13

30 Sep '13

Hi all, We all know and love Data.Foldable and are familiar with left folds and right folds. But what you want in a parallel program is a balanced fold over a tree. Fortunately, many of our datatypes (Sets, Maps) actually ARE balanced trees. Hmm, but how do we expose that? It seems like it would be nice to have a* standard class t*hat allows you to split a datatype into roughly even halves, until you get down to the leaves. This goes along with Guy Steele's argument that we should use "append lists" as primitive rather than "cons-lists", and it's why we added append-lists within the monad-par library<http://hackage.haskell.org/package/monad-par-extras-0.3.3/docs/Control-Mona…> . Does this class exist already? A random google search brought up this module by the name Data.Splittable<http://hackage.haskell.org/package/unfoldable-0.2.0/docs/Data-Splittable.ht…>, but it's not quite the right thing. Thanks, -Ryan

4 6

Vacancy Assistant Professor Software Technology Utrecht University (1, 0 fte)
by Johan Jeuring 30 Sep '13

30 Sep '13

Assistant Professor Software Technology Utrecht University (1,0 fte) Job description The division Software Systems of the Department of Information and Computing Sciences is looking for an Assistant Professor for the bachelor programmes Computing Science and Information Science and the master programmes of the department, and for research in the area of software technology. You are expected to develop an independent line of research within the field of software technology in cooperation with the other members of the division Software Systems. The tasks include: performing scientific research in the field of software systems, in particular software technology; supervising PhD students and acquiring research funding; developing and teaching courses within the bachelor and master programmes of the department; supervising internships and theses; organizational activities within the division, department or faculty. Qualifications We are looking for candidates with a PhD with expertise and experience in scientific education and research in computer science. Expertise in the fields of software technology, programming languages and typing systems, advanced programming methods, or compilers is required. Experience in academic education within a university setting is desired. You have published on the aforementioned areas in national and international conferences and / or journals. Well-developed teaching skills and command of English in speaking and writing are a requirement. Candidates who prefer part-time employment are also invited to apply by specifying the desired part-time ratio. Offer We offer a position in a dynamic environment. The position is for at most five years. The total size of the position is 100%, but part-time is possible. The salary depends on education and experience and ranges between € 2,919 (scale 10) and € ,5070 (scale 12) gross per month for a full-time appointment. Additionally, excellent secondary benefits are provided, such as 8% holiday allowance and 8.3% end of year bonus. We also offer a pension scheme, partially paid parental leave and flexible working conditions. For more information see the terms of employment. The department provides the candidate the necessary support for the arrangements of the education and research line. About the organisation Utrecht University has great ambitions for its teaching quality and study success rates. This also applies to its clear research profiles which are centred around four themes: Sustainability, Life Sciences, Youth & Identity, and Institutions. Utrecht University plays a prominent role in our society and contributes to finding the answers to topical and future societal issues. The Faculty of Science at Utrecht University comprises six departments: Biology, Pharmacy, Information and Computer Sciences, Chemistry, Mathematics and Physics and Astronomy. It has 3500 students and nearly 2000 employees, and is internationally known for its quality in research. The academic programs of the Faculty reflect developments in society and science of today. The Department of Information and Computer Science is nationally and internationally renowned for its fundamental research in computer science and information science. Its research is positioned around game technology, one of the four research themes of the Faculty of Science. The research of the department is grouped into four divisions: Software Systems, Artificial Intelligence, Virtual Worlds and Interaction Technology. The Department offers bachelor programmes in computer science and information science, and four English-language research master-programmes including Computing Science. High enrollment figures and good student ratings make the education very successful. Part of the research of the division Software Systems focuses on how programming languages, methods, and tools can be adapted to support program construction. More specifically, the research covers the following areas: programming languages and compiler construction-tools program analysis interaction analysis, error diagnosis advanced programming methods: generic programming, dependently-typed programming program testing and verification The division aims to design better programming languages and methods; to implement tools and languages that help programmers work more effectively; and to develop products and systems demonstrating the feasibility (and/or validity) of the ideas. An important application area studied in the department and the division is (serious) games. The division mainly contributes to the education in the bachelor programs information science and computer science, and to the master program Computing Science. Additional information For more information please contact Prof.dr. Johan Jeuring +31.30.253.4115 or +31.6.400.100.53, email:J.T.Jeuring@uu.nl. Apply Your application must contain a letter of motivation, your CV with publication list, teaching and research statement, and contact details of at least two references. You can respond via the application button below. The application deadline is 21/10/2013 Assistant Professor Software Technology (1,0 fte)

1 0

New Functional Programming Job Opportunities
by Functional Jobs 30 Sep '13

30 Sep '13

Here are some functional programming job opportunities that were posted recently: Senior Scala Developer for Green Building Software at Sefaira http://functionaljobs.com/jobs/8649-senior-scala-developer-for-green-buildi… Cheers, Sean Murphy FunctionalJobs.com

1 0

FFI: how to handle external dll crashes
by Miro Karpis 30 Sep '13

30 Sep '13

Please, can you help me with following: I have an external dll that I'm importing in my haskell program. In some particular cases the dll crashes. Simplified: first I need to send to dll with MethodA some parameters and then call MethodB to do some calculations on those parameters. If I didn't give enough parameters then MethodB will crash the whole dll and my Haskell application. Is there a way to handle this? Unfortunately there are no exceptions thrown from the dll. In ghci I'm getting following message: ERROR in InitNumericalSystem::initializeSystem. JuncLabel. I have tried to use "catchAny but that didn't help. c_run is my external dll method which takes 4 input parameters: catchAny :: IO a -> (SomeException -> IO a) -> IO a catchAny = Control.Exception.catch main :: IO () main = do let timeTot = []::[CDouble] timeNow = []::[CDouble] runType = 2::CInt timeTotPtr <- newArray timeTot timeNowPtr <- newArray timeNow result <- (catchAny $ c_run timeTotPtr runType timeNowPtr 0) $ \e -> do putStrLn $ "Got an exception: " ++ show e putStrLn "Returning dummy value of -1" return (-1) free timeTotPtr free timeNowPtr print result I have tried also with withAsync, and no luck tryAny :: IO a -> IO (Either SomeException a) tryAny action = withAsync action waitCatch catchAny :: IO a -> (SomeException -> IO a) -> IO a catchAny action onE = tryAny action >>= either onE return try2 :: IO () try2 = do let timeTot = []::[CDouble] timeNow = []::[CDouble] runType = 2::CInt timeTotPtr <- newArray timeTot timeNowPtr <- newArray timeNow putStrLn $ "c_run going to call c_run.." result <- catchAny (c_run timeTotPtr runType timeNowPtr 0) (const $ return (-1)) free timeTotPtr free timeNowPtr putStrLn $ "Result: " ++ show result Is there a way how I can handle this? cheers, m.

3 5

Richard Bird and the fast nub function
by Henning Thielemann 29 Sep '13

29 Sep '13

In Richard Bird's "Functional Pearls in Algorithm Design" there is chapter 10 "Removing duplicates" which is about a fast and sorting variant of 'nub'. After reading the introduction of the chapter I answered mentally "Set.toAscList . Set.fromList - next chapter please". However after the introduction eight pages follow that develop an efficient algorithm for the problem. I suspected there might be the additional difficulty of not using the standard Set type, however on page 70 the common Set API is "imported". In the final remarks of the chapter Bird writes: "A nagging doubt remains that there might be a much simpler solution to such a simply stated problem. But so far I have not been able to find one." Now I am lost. Can someone tell me, how the task differs from what "Set.toAscList . Set.fromList" does?

3 2

[ANN] New OpenCV Bindings
by Arjun Comar 29 Sep '13

29 Sep '13

Hi all, I've been hard at work on a new set of OpenCV bindings that will hopefully solve a lot of the shortcomings with previous attempts. An automatic header parser has been used to generate a full set of Haskell bindings for the C++ API, and I'm now working to create a pleasant Haskell API. The plan is to expose major functionality through pipes for two reasons: 1) OpenCV is not very referentially transparent, and so you're stuck in IO for anything non-trivial. Lazy IO is potentially problematic, and so immediate incorporation of a proper streaming library is probably best. 2) Computer vision algorithms are frequently expressed in terms of pipelines of functionality, and a pipes approach can provide a very natural API for expressing these algorithms. At this stage, I could very much use input from anyone who's interested in these bindings. I've got the basics of a library forming, but there's a ton to do and help in the form of pull requests are very very welcome. I could also use feedback on what's been done so far (especially criticisms) as well as feature requests. The current plan is to develop the pipes API for the core functionality. The plan is to provide the major functionality from core, highgui, imgproc, features2d, contrib, ml, flann, video, and objdetect as quickly as I can before trying to cover any of the more specialized parts of the API. The functions and types from these modules (baring a few major and important exceptions) are automatically translated into C wrappers and raw Haskell bindings. If there's sufficient interest in these raw bindings separately from the API I'm developing, I'll release them as a separate package on Hackage. I'm calling my project revelation for the moment because I'm bad at naming things and it was the best pun I could come up with. Github: www.github.com/arjuncomar/revelation Regards, Arjun

3 10

Proposal: new function for lifting
by Thiago Negri 29 Sep '13

29 Sep '13

Everybody is claiming that using lift is a bad thing. So, I come to remedy this problem. Stop lifting, start using shinny operators like this one: (^$) :: Monad m => m a -> (a -> b -> c) -> m b -> m c (^$) = flip liftM2 Then you can do wonderful stuff and you will never read the four-letter word in your code again: \> Just 42 ^$(+)$ Nothing Nothing \> Just 10 ^$(+)$ Just 20 Just 30 \> let add = (+) \> Just 30 ^$ add $ Just 12 Just 42

4 3