February 2014 - Haskell-Cafe

Re: [Haskell-cafe] Rewrite NetBSD kernel driver using Ajhc Haskell compiler
by Kiwamu Okabe 19 Feb '14

19 Feb '14

Hi Johnny, On Wed, Feb 19, 2014 at 11:57 PM, Johnny Billquist <bqt(a)update.uu.se> wrote: > Maybe someone with more insight could explain to an idiot like me how > Haskell garbage collection is handled when running in the kernel? First, s_alloc() is Ajhc Haskell compiler's allocator, and gc_perform_gc() is GC runner. https://github.com/ajhc/ajhc/blob/6c053b940812689fe22d2dbe2b7ad835dab8a3c7/… https://github.com/ajhc/ajhc/blob/6c053b940812689fe22d2dbe2b7ad835dab8a3c7/… The s_alloc() try to find free area. But if not, it runs GC. https://github.com/ajhc/ajhc/blob/6c053b940812689fe22d2dbe2b7ad835dab8a3c7/… By the way, where from s_alloc() get a memory chunk? Now using kern_malloc(9). https://github.com/metasepi/netbsd-arafura-s1/blob/arafura-s1/metasepi/sys/… Regards, -- Kiwamu Okabe

2 3

Functor instance derived from Monad
by Frantisek Farka 19 Feb '14

19 Feb '14

Hi cafe, when reading the Functor-Applicative-Monad proposal [1] on Haskell wiki I remembered that I have seen instance similar to instance (Monad m) => Functor m where fmap f a = a >>= \x -> return $ f x with FlexibleInstances and UndecidableInstances language extensions. But I cannot remember where it was nor can I find any useful reference. Is there some library in Hackage with this and similar instances? Can you point out some leading marks to me? Thanks FF [1] http://www.haskell.org/haskellwiki/Functor-Applicative-Monad_Proposal

2 3

Re: [Haskell-cafe] How to benckmark a Word64 -> Bool?
by Charles-Pierre Astolfi 19 Feb '14

19 Feb '14

Thanks Erik! I was kind of hoping for such a benchmark. On my system, the C version is consistently twice as fast as the Haskell version. (I tried -fllvm, but it doesn't change perf) Then I added bang patterns for the argument of only012 and in the tuple (p,q) and it now has the same runtime as C. Just for fun, I tried to void writing explicit recursion and used only combinators and this Haskell code is still as competitive as C, so the real bottleneck really was the laziness in the tuple/argument. http://lpaste.net/100153 In fact, this Haskell version takes 8 sec vs 10 sec for the C version on my laptop (input: 1000000000). I tried using Data.List.Stream, I couldn't get any performance gain, probably because the code is very close to optimal. So remember kids: - Bang pattern the heck out of your code (not a silver bullet, but definitely worth trying). - Don't rely on too much IO for benchmark. - Don't be afraid of combinators even if you want performance. - Ask on haskell-cafe, chaps here are very helpful :) -- Cp On Wed, Feb 19, 2014 at 8:39 PM, Erik Rantapaa <erantapaa(a)gmail.com> wrote: > I think a lot is being obscured by the IO subsystem. > > If you print just a count of the only012 numbers I get very different > results. > > Here are the versions I compared: > > http://lpaste.net/100142 > > and on my system the C version is about 10x faster. > > > On Wednesday, February 19, 2014 12:49:36 PM UTC-6, Charles-Pierre Astolfi > wrote: > >> Switching to quotRem gave no measurable improvements. >> After switching to ByteString, the code now runs in 9 seconds, which >> outperforms my C version. But honestly, I have no idea why. >> >> New code: >> http://lpaste.net/100136 >> >> $ ghc --make -O3 303only012.hs && time ./303only012 50000000 > /dev/null >> ./303only012 50000000 > /dev/null 9.72s user 0.21s system 90% cpu 10.961 >> total >> >> @Alois, I'm not sure how criterion can help compare my code with the C >> version, since in the C version I cannot measure the exec time of only012 >> only. What did you have in mind? >> >> Thanks everyone! >> >> >> -- >> Cp >> >> >> On Wed, Feb 19, 2014 at 7:24 PM, Levent Erkok <erk...(a)gmail.com> wrote: >> >>> Also, prefer quotRem over divMod as the former is faster. See here: >>> http://stackoverflow.com/questions/339719/when-is- >>> the-difference-between-quotrem-and-divmod-useful >>> >>> >>> On Wed, Feb 19, 2014 at 10:19 AM, Gregory Collins < >>> gr...(a)gregorycollins.net> wrote: >>> >>>> >>>> On Wed, Feb 19, 2014 at 9:36 AM, Charles-Pierre Astolfi <c...(a)crans.org >>>> > wrote: >>>> >>>>> So there's a difference, but I'm not sure if it's related to my >>>>> algorithm or related to IO/RTS. >>>> >>>> >>>> Text.Printf is slower than the C version (and uses String). Use >>>> Data.ByteString.putStr instead. >>>> >>>> >>>> >>>> -- >>>> Gregory Collins <gr...(a)gregorycollins.net> >>>> >>>> _______________________________________________ >>>> Haskell-Cafe mailing list >>>> Haskel...(a)haskell.org >>>> http://www.haskell.org/mailman/listinfo/haskell-cafe >>>> >>>> >>> >>

3 3

How to benckmark a Word64 -> Bool?
by Charles-Pierre Astolfi 19 Feb '14

19 Feb '14

Hi, I have a Word64 -> Bool function called only012 that returns True iff the digits in base 10 are 0, 1 or 2. I wand to make sure that I'm competitive with C, so I wrote the following code: Haskell version: http://lpaste.net/100129 C version: http://lpaste.net/100128 $ ghc --make -O3 303only012.hs && time ./303only012 50000000 > /dev/null ./303only012 50000000 > /dev/null 21.44s user 0.41s system 91% cpu 23.754 total $ g++ -O3 303only012.cpp -o 303cpponly012 && time ./303cpponly012 50000000 > /dev/null ./303cpponly012 50000000 > /dev/null 13.87s user 0.03s system 95% cpu 14.577 total So there's a difference, but I'm not sure if it's related to my algorithm or related to IO/RTS. The core output of only012 looks quite tight: http://lpaste.net/100130 Am I doing the IO part right, that is fast IO that is not screwing my benckmark? I'm not so sure about that since I use a convoluted way to print the boolean result in Haskell. Cheers, -- Cp

4 5

Haskell-platform release information
by Obscaenvs 19 Feb '14

19 Feb '14

I'd like a pointer to some place containing information as to why the new HP is delayed (although I have my suspicions the reason being Xcode 5), and if there is a new schedule. Why isn't there a pointer on [1]? If one can't stay on schedule, which is sometimes understandable and I guess in this case highly so, then some info as to why this is is the least damage control one could do. But maybe there is such info already, and I am just unable to find it - except for the by now rather old discussion [2]. /fredrik [1] http://www.haskell.org/platform/ [2] https://www.mail-archive.com/haskell-platform@projects.haskell.org/msg02594… -- bugthunk.net @bugthunk

3 4

[ANN] data-default-generics
by João Cristóvão 19 Feb '14

19 Feb '14

Hi, This project is a fork of the well known data-default package, since the original author does not seem to be reachable, and has not accepted a rather useful patch in the past. It mainly adds: * Regrouped everything into a single module * Added several instances, namely Either, Text, ByteString, etc * Added Generics support patch by Jonathan Fischoff, with some additional support for recursive datatypes by me, with the precious help of José Pedro Magalhães This last point means that you can easily defined a default value for a record. See: https://github.com/jcristovao/data-default-generics http://hackage.haskell.org/package/data-default-generics-0.3/candidate If this is a duplicate effort, or a better alternative exists, please let me know. If not, I will soon publish it to hackage as 'official'. Thanks, João

1 0

Rewrite NetBSD kernel driver using Ajhc Haskell compiler
by Kiwamu Okabe 19 Feb '14

19 Feb '14

Hi all, NetBSD kernel's HD Audio sound driver has interrupt handler. The interrupt handler of the demo is re-written by Haskell language using Ajhc Haskell Compiler http://ajhc.metasepi.org/. Demo movie could be found at following link. http://www.youtube.com/watch?v=XEYcR5RG5cA At the demo, run following operations. First, set breakpoint at the interrupt of finding headphone, and see Haskell function names on backtrace. Second, set breakpoint s_alloc() function, that allocate area in Haskell heap. Make sure of calling the function while anytime running kernel. Nevertheless, playing wav file does not break up. The source code is found at https://github.com/metasepi/netbsd-arafura-s1. And a picture of the architecture is also at http://www.flickr.com/photos/masterq/12616929055/. Thank's, -- Kiwamu Okabe

3 2

how to write cabal package file for executable with unit tests?
by Richard Cobbe 18 Feb '14

18 Feb '14

What's the preferred way to write a cabal configuration file for a package that contains an executable and a test suite? The only thing I've found that seems to work is to put a thin wrapper Main.hs into the executable, put all of the actual code into the library, and then have the test-suite include the library in its build-depends list. This works, but I find it mildly annoying that I have to separate the code out in this fashion -- especially since the unit tests require me to expose modules in my library that aren't really part of the library's public interface. In more detail: This is a follow-up to a thread I posted back on Feb 2; you can see it on GMane at <http://thread.gmane.org/gmane.comp.lang.haskell.cafe/109374>. The original example (tarball attached to the first message in that thread) was a package structured as a library with unit tests. After acting on Erik's suggestion to put the library and test-suite source in different directories, I was able to compile the library without the type error I originally mentioned. However, this isn't really the package structure I want. The package is actually intended to produce a standalone executable, not a library. My original example used a library for reasons that I don't honestly remember -- I think I more or less stumbled upon that structure as something that more or less worked, given the difficulty of getting Cabal, test suites, Happy, and Alex to all work together well. So I tried replacing the "Library" section of the cabal file with an "Executable" section (sample tarball attached), but then I can't get the tests to compile. If I leave sample as a build-depends for the test-suite, then "cabal configure --enable-tests" fails with Preprocessing test suite 'sample-tests' for sample-0.4... <command line>: cannot satisfy -package-id sample-0.4-inplace (use -v for more information) If I take sample out of the build-depends list for the test-suite, then "cabal configure --enable-tests" works, but "cabal build" fails: test-src/RunTests.hs:8:8: Could not find module `Parser' It is a member of the hidden package `ghc-7.6.3'. Perhaps you need to add `ghc' to the build-depends in your .cabal file. Use -v to see a list of the files searched for. Either way, "cabal configure && cabal build" works fine; it's only when I try to build the test suite that things break down. Thanks, Richard

6 8

Thank you for the donations to Haskell.org!
by Jason Dagit 18 Feb '14

18 Feb '14

On behalf of Haskell.org, we would like to thank all the generous donors we had in 2013! Just in the last week of the year, we received about half of the total donations for the year. Nearly $900! That brings the total donations for the year to $1976! Thank you very much! We are very grateful for all the donations we receive, financial and otherwise. As examples, GlobalSign graciously provides us with a free SSL certificate for use on hackage and each year community members collectively provide countless hours of time and energy towards making the Haskell community vibrant, positive, and relevant. At the moment our primary expenses are hosting for haskell.org and hackage. The donations we have received will help ensure that we continue to have adequate hosting for the ever growing hackage! Thank you and let's continue our momentum into 2014! Sincerely, The Haskell.org Committee

1 0

Yi documentation
by Mateusz Kowalczyk 18 Feb '14

18 Feb '14

Hi all, In #yi (and other Haskell-related outlets), we ofter have people who want to try Yi but struggle one way or another. It seems that a big barrier is lack of documentation. If you had tried Yi or wanted to try it and you met with difficulties due to lacking documentation, I'd like to hear about it. Perhaps you tried to get going but just couldn't figure out how to do something in your config, or perhaps you tried hacking on the source but didn't know where to get started. If you're looking for any specific guides, I'd also like to hear about it. I (and perhaps some others) can write documentation but it helps if we have specific topics to talk about rather than taking guesses in the dark. If you want to try it out, we recommend that you read the README at [1]. If you have problems that aren't related to the documentation, you should use the issue tracker[2] to report them. Thanks! [1]: https://github.com/yi-editor/yi [2]: https://github.com/yi-editor/yi/issues?state=open -- Mateusz K.

5 12