- Haskell-Cafe - Haskell.org

Haskell Speed Myth
by Thomas Davie 26 Aug '08

26 Aug '08

Lo guys, I thought you'd like to know about this result. I've been playing with the debian language shootout programs under OS X, looking at how fast Haskell code is compared to C on OS X, rather than linux. Interestingly, Haskell comes out rather better on OS X than it did on Linux. Here's my results (times in seconds): C Haskell Relative speed Inverse Binary Trees 6.842 1.228 0.179479684302835 5.57166123778502 Fannkuch 5.683 15.73 2.76790427591061 0.361284170375079 Mandelbrot 1.183 2.287 1.93322062552832 0.517271534761697 nbody 10.275 16.219 1.57849148418491 0.633516246377705 nsieve 0.167 0.253 1.51497005988024 0.660079051383399 nsieve-bits 0.471 0.713 1.51380042462845 0.660589060308555 partial sums 1.047 1.313 1.25405921680993 0.797410510281797 pidigits 1.238 1.4 1.13085621970921 0.884285714285714 recursive 1.554 3.594 2.31274131274131 0.432387312186978 spectral-norm 27.939 19.165 0.685958695729983 1.45781372293243 threadring 91.284 1.389 0.0152162481924543 65.719222462203 ------------------------------------------------------------------------- Averages 1.35333620432893 0.738914688605306 Some notes: Hardware: 2Ghz Core2Duo, enough ram to not worry about paging Some programs are not included, this is because the C code produced compile errors. The Haskell code appears to be portable in *all* cases. The average slowdown for running Haskell is only 1.3 times on OS X! That's pretty damn good. I'm sure some people will say "yeh, but you have to optimise your code pretty heavily to get that kind of result". Interestingly, the programs that have the biggest speed advantage over C here are also the most "naïvely" written ones. The thing that seems to make C slower is the implementation of malloc in OS X. This has a couple of implications, first, if apple replaced the malloc library, it would probably push the result back to the 1.7 times slower we see under Linux. Second, Fannkuch can probably be optimised differently for Haskell to look even better -- at the moment, the haskell code actually uses malloc! Finally, that threading example... WOW! 65 times faster, and the code is *really* simple. The C on the other hand is a massive mess. Bob

7 13

Valid Haskell characters
by Maurício 26 Aug '08

26 Aug '08

Hi, In Haskell reference, I see the following definitions: uniWhite -> any Unicode character defined as whitespace; uniSmall -> any Unicode lowercase letter; uniLarge -> any uppercase or titlecase Unicode letter; uniSymbol -> any Unicode symbol or punctuation. Where do I get lists for those characters? My first attempt was to check: http://unicode.org/Public/UNIDATA/UnicodeData.txt and consider large anything marked as CAPITAL and small anything marked as SMALL. I didn't know what to guess about the symbols. Am I using the right reference? How can I recognize (or get a list of) valid uppercase and lowercase unicode letters, as well as symbols and punctuation? Thanks for your help, Maurício

3 4

Re: [Haskell-cafe] Valid Haskell characters
by Richard A. O'Keefe 26 Aug '08

26 Aug '08

On 26 Aug 2008, at 3:42 pm, Deborah Goldsmith wrote: > All characters with general category Lu have the property Uppercase, > but the converse is not true. It depends on what the OP wants to do with the information. For example, Unicode Standard Annex 31, http://www.unicode.org/reports/tr31/tr31-9.html is defined in terms of the General Character classification, *not* in terms of the binary properties "Upper", "Alpha", &c. When the Haskell report says uniSmall -> any Unicode lowercase letter uniLarge -> any uppercase or titlecase Unicode letter it is really unclear what definition is meant: are we talking about characters in general category Lu or Lt, or are we talking about characters with the Uppercase property? Since it's _identifiers_, I'd expect UAX#31 to apply, so it should be general category. The specification of the Char module is similarly ambiguous. Since this is *not* about identifiers, I suppose this time the Other_Uppercase characters might well be included. It would be nice to have this spelled out clearly somewhere not too far from the Report on haskell.org.

1 0

Re: [Haskell-cafe] Valid Haskell characters
by Richard A. O'Keefe 26 Aug '08

26 Aug '08

On 26 Aug 2008, at 1:31 pm, Deborah Goldsmith wrote: > You can't determine Unicode character properties by analyzing the > names of the characters. However, the OP *does* have a copy of the UnicodeData...txt file, and you *can* determine the relevant Unicode character properties from that. For example, consider the entry for space: 0020;SPACE;Zs;0;WS;;;;;N;;;;; ^^ The Zs bit says it's a white space character (Zs: separator/space, Zl: separator/line, Zp: separator/paragraph). Or look at capital A: 0041;LATIN CAPITAL LETTER A;Lu;0;L;;;;;N;;;;0061;^ ^^ The Lu bit says it's a L(etter) that is u(pper case). Upper case: Lu, lower case: Ll, title case: Lt, modifier letter: Lm, other letter: Lo, digit: Nd, ... If memory serves me correctly, this is explained in the UnicodeData.html file, under a heading something like Normative Categories.

3 2

can the default instances in GHC be recompiled with overlap?
by Jason Dusek 25 Aug '08

25 Aug '08

Per the GHC manual section on instance overlap: If an instance declaration is compiled without -XOverlappingInstances, then that instance can never be overlapped. This could perhaps be inconvenient. It is inconvenient for certain things. If I want to declare a special `instance Show [MyType] where...` I am out of luck. However, the solution need not involve a change to the way instances are handled -- if all the `GHC.*` instances were compiled with overlapping instances, I'd be able to declare my instance. Is there a substantial barrier, either theoretical or practical, to recompiling GHC this way? -- _jsn |...GHC manual section...| http://www.haskell.org/ghc/docs/latest/html/users_guide/type-class-extensio…

3 4

Prototype of a syntax reference
by Maurício 25 Aug '08

25 Aug '08

Hi, I'm trying to make a syntax reference in the form of source code. It's not ready, but if someone shares the same ideas I could use some help. (Should I put this on a wiki? Where?) My “design principles” are: • It should contain just pieces of uncommented code. The idea is not to use it as a fast tutorial, but as a reminder of how to write all kinds of things, and solve all questions of type “is this allowed?”; • It should be small, and fit on a printed A4 page; • It should show the complete syntax of everything, not the most common. For instance, ‘case of’ should show the use of guards; • It should be an example of valid code, not good one. The idea is to show what can be done, not what should :) • It should use expanded code, not identation rules. Replacement of identation for code (like {} etc.) have simple rules, and the proper way to ident can be deduced from expanded code. Sometimes this is important if we are trying to solve syntax errors; • It should use Unicode. It's nice to know which kind of Unicode characters can actually be used; • It should use no predefined types or classes except Bool. I’ll write a Prelude code reference if this first attempt go right. So I came up with the weird code below. It should contain a lot more, and be more dense (more syntax examples per lines of code), but it's already usefull for a begginer like me. What do you think? Best, Maurício module Ça (Exèmplo(Åwirù,Øwet'Л,Øwet'И),aҝҧү,Clááá) where { class Clááá ú where { (⊗) :: ú -> ú -> ú; τAmber :: ú -> ú; τAmber a = a ⊗ a; }; data D١ = U1 | U2 | U3; data D٢ = Vú1 | Vú2 | JustVú; data D٤ = Opção | OutraOpção; data Exèmplo = Åwirù | Øwet'Л D٤ | Øwet'И D٢ Exèmplo; osmesmos :: D٤ -> D٤ -> Bool; osmesmos Opção Opção = True; osmesmos OutraOpção OutraOpção = True; osmesmos _ _ = False; aҝҧү = (Øwet'И Vú2 aҝҧү) ⊗ Åwirù; instance Clááá Exèmplo where { Åwirù ⊗ _ = Åwirù; a@(Øwet'Л b) ⊗ x = case x of { Åwirù -> a; y@(Øwet'Л o) | (osmesmos o Opção) -> a | otherwise -> Øwet'И Vú1 y; y@(Øwet'И _ _) -> y ⊗ a }; a@(Øwet'И Vú1 _) ⊗ _ = a; a@(Øwet'И Vú2 x) ⊗ _ = x; a@(Øwet'И JustVú _) ⊗ y = y; } }

3 3

Re: [Haskell-cafe] Haskell Propeganda
by Niels Aan de Brugh 25 Aug '08

25 Aug '08

Christopher Lane Hinson wrote: >> Here's an error the Haskell run-time system might throw: >> *** Exception: Prelude.head: empty list >> (or whatever) >> >> So now what? Action plan = []. > > rgrep " head " . > > Lists 17 items for three projects I've been working on summing to > 10000 lines of haskell. Use listToMaybe or pattern matching and > -Wall, and raise an explicit (error "with explanation"), which is all > good practice in the C world as well. Please, don't make undocumented > use of partial functions in serious code. You're absolutely right, and the compiler can warn the programmer for such mistakes. However, I've found that analysing the run-time behaviour/errors (if they do occur) in a Haskell program to be harder than in a language like C(++) which has many very mature engineering tools available. No doubt this is a matter of time and money, not which language is better suited for "real work". Regards, Niels

2 1

ANN: The Monad.Reader - Issue 11
by Wouter Swierstra 25 Aug '08

25 Aug '08

I am pleased to announce that a new issue of The Monad.Reader is now available: http://www.haskell.org/haskellwiki/The_Monad.Reader Issue 11 consists of the following three articles: * David F. Place How to Refold a Map * Kenneth Knowles First-Order Logic a la Carte * Douglas M. Auclair MonadPlus: What a Super Monad! The Monad.Reader is a quarterly magazine about functional programming. It is less formal than a journal, but a bit more enduring than a wiki page or blog post. If you'd like to write something for the next issue of The Monad.Reader, please get in touch. I haven't fixed the deadline for the next issue just yet. Expect a deadline early 2009. Wouter

1 0

Re: [Haskell] Top Level <-
by Duncan Coutts 25 Aug '08

25 Aug '08

This topic came up in #haskell this evening... On Sun, 2008-08-24 at 16:12 -0700, Ashley Yakeley wrote: > 1. Global mutable state. For instance, here's the count variable for > Data.Unique rewritten: > > uniqSource :: MVar Integer > uniqSource <- newMVarTL 0 > > Isn't that much nicer? > <http://haskell.org/haskellwiki/Top_level_mutable_state> It's nicer syntax but it's still not true. There's still no such thing as a global variable. There's always a scope. In this case what scope are we looking for? Process scope? Only one instance of uniqSource per process? If so, it's not right. We can link in multiple copies of a package and thereby end up with multiple instances (this really happens, it's not a far fetched scenario). So uniqSource is a package-name/version scope mutable variable. To get the right thing for the annoying C libs that have 'global' init and finalise functions you need something else. Not actually process scope, that's too wide. You need exactly the same scope as the system linker is using, so that we get one variable per instance of the C library linked into the process. Consider a process using several dlls where each one links to a C lib that needs 'global' initialisation, we must have exactly one 'global' var instance per dll in this case because that's the linker namespace scope. The scope you need depends on the application. A language extension like this is likely to only cover package-name/version scope, not scopes like thread, OS thread, rts instance, linker, process, user account, OS, subnet, interweb, planet etc. So in practise, what scopes do we often need, and does package-name/version scope correspond to many of those use cases? Duncan

2 1

ANN: Mueval 0.5.1, 0.6, 0.6.1, 0.6.2, 0.6.3, 0.6.4
by Gwern Branwen 25 Aug '08

25 Aug '08

Evening folks. So it's that time again, to do a Mueval announcement! This time I'd like to announce version 0.6.4, & 0.6.3. And 0.6.2. And 0.6.1. And 0.6. And 0.5.1. These releases have some nice stuff to them; perhaps the most interesting aspect is that I've carried through on my threats^Wpromises, and Lambdabot in #haskell now uses Mueval. GETTING All the stuff from the previous ANNs still apply here: you can get it at <http://hackage.haskell.org/cgi-bin/hackage-scripts/package/mueval>, do a 'darcs get --lazy http://code.haskell.org/mubot/', cabal-install mueval, and so on. DEPENDENCIES 0.6.4 does add an additional dependency, on haskell-src-exts, for use in some experimental parsing code, so you'll now need that in addition to Hint, utf8-string, and show. (Upgrading Hint is mandatory, utf8-string unnecessary, and show nice but certainly not necessary.) BUILDING Same as usual. CHANGES 0.5.1 * SmallCheck support 0.6 * A new option, --loadfile, which loads a given *.{l}hs file with a module and makes its function available for evaluation. This is a complex option; see the README. * Cleaner output, thanks to a needless show call sjanssen identified. 0.6.2 * Speed improvements (shorter timeouts, optimizations) * Mueval imports many more modules by default, so one can freely use standard functions from Control.Monad, for example; this also brings with it many test examples from #haskell. * experimental --noimports flag, to disable default imports (such as the above, and also of the Prelude). This was motivated by Lambdabot's Caleskell, but I'm not sure why it doesn't interact well with --loadfile (and so doesn't work with Lambdabot). 0.6.3 * Saizan contributed some significant improvements to error reporting and thread handling, so the infamous time-out errors are replaced by normal errors more reminiscent of GHCi. 0.6.4 * Fixed an issue with lazy printing of results * Exceptions are now caught and printed more prettily; again, thanks to Saizan * Forced a dependency on the latest Hint; this fixes an error where one could successfully evaluate: bash-3.2$ mueval -e '1 + 1 {- addition example -}' 2 but not: bash-3.2$ mueval -e '1 + 1 -- addition example' * Added a --rlimits option. I decided that the POSIX resource limits were causing too many people problems. So now non-time resource limits are disabled unless specifically requested with --rlimits. * mmorrow contributed some experimental parsing patches intended to catch subtle invocations of dangerous functions. If haskell-src-exts is too burdensome a dependency, Mueval/Context.hs and mueval.cabal can be easily edited to remove the dependency. KNOWN ISSUES * Qualified imports do not work. That is, one cannot do $ mueval -e 'M.map (+1) $ M.fromList [(1,2)]' or whatever and expect the proper Data.Map functions to be substituted in. This is a weakness in the GHC API, AFAIK. It is possible that this will be fixed in future versions of Hint or GHC, but I do not expect that to happen for months. * Many limits in System.Posix.Resource are simply completely broken for many users; that is, *no* value can be set using them. I have been unable to figure out the heck why this is the case (I can't reproduce it), and AFAIK no one has filed a bug with the GHC developers, so this probably won't be fixed anytime soon. -- gwern Secure wwics P415 SASP Comirex mailbomb ISADC Sayeret SBI Jatti

2 2