- Haskell-Cafe - Haskell.org

Text.Html introduction
by Colin Paul Adams 19 Aug '09

19 Aug '09

http://www.haskell.org/ghc/docs/latest/html/libraries/xhtml/Text-XHtml.html says: "Based on the original Text.Html library by Andy Gill. See http://www.cse.ogi.edu/~andy/html/intro.htm for an introduction to that library. " But that link gives a not-found page. Anyone know where it is know? -- Colin Adams Preston Lancashire

4 3

simulation in the haskell way
by Eric Wong 19 Aug '09

19 Aug '09

Hello, Haskellers! I'm relatively new to haskell and due to my strong imperative background, it's really a pain to learn haskell. But I'm really indulged in it. :) Now I think I understand the basics of Haskell very well, such as the type system and monad. And for those data-flow kind of applications, I can easily structure the problem in a functional way and sketch out an intuitive picture of the computation. But for simulation kind-of problems, in which I think OO really fits the best, what's the haskell way to structure such problems? I once thought maybe I can use the State monad to simulate objects. But it's really hard for me to implement, because I think State monad is used to simulate a global shared state, is it right? Then what's the best way to simulate private states or just instead how to solve simulation problems such as a physical engine using the haskell way? Best regards. Eric

9 17

Re: [Haskell-beginners] Start file with associated prog (windows only?)
by Max Desyatov 18 Aug '09

18 Aug '09

Bernhard Lehnert <b.lehnert(a)gmx.de> writes: >> Windows has a command "start" which you can use e.g. via System.Process.proc. > > Thank you, Magnus - this is exactly what I was looking for. (By the way, > someone should implement something like this for GNOME and KDE, > too ;-) ) You can do that using MIME, look at /etc/mailcap, it already contains all what you need to run appropriate process for exact MIME-type. As long as it's supported in your distribution.

2 1

Installing documentation with cabal
by Maciej Piechotka 18 Aug '09

18 Aug '09

Is it possible to automatically (or non-automatically) install documentation when installing package with cabal-install? Regards

2 2

Library function for map+append
by Dusan Kolar 18 Aug '09

18 Aug '09

Hello all, During a small project I'm trying to develop a small application. It becomes quite often that I need a function mapapp: mapapp _ [] ap = ap mapapp f (a:as) ap = f a : map f as ap I tried hoogle to find such a function with no success. Is there any function/functions built-in "standard" libraries that could easily satisfy the functionality with the same or even better (?) efficiency? Of course, (map f list) ++ append would do the same as mapapp f list append but with less efficiency. Or am I wrong? Thanks Dusan

9 14

Looking up functions inQ monad (Template Haskell).
by Zefirov Sergey 18 Aug '09

18 Aug '09

Why isn't possible to lookup and call user-defined functions while performing Template Haskell actions? Just like the way Template Haskell looks up and calls 'f' in $(f ...). 'f' gets looked up and called. And user of Q monad cannot do that. A colleague of mine, a Lisp user, says that almost everything is in place. It's just not enabled. Is there any problem we do not understand?

2 1

ANN: OpenCLRaw 1.0.1000
by Jeff Heard 18 Aug '09

18 Aug '09

All, I've released a raw binding to the OpenCL platform on Hackage. The main differences between it and the C bindings are that constants have been replaced by newtypes for type safety reasons, void-essential functions that return a token errorcode return a Maybe ErrorCode, and functions that return a handle/ptr and return an error-code through an out-parameter return instead an Either ErrorCode a. Modules are grouped roughly by the section in which they appear in the standard. I'm working on an OpenCL binding that is a little more "cooked" as well as a high-level OpenCL binding. OpenCL is a platform for single-host heterogenous, data-parallel computing that follows roughly the OpenGL and OpenAL conventions for how the library is architected. It will be available by default in the next version of Apple's OS-X, and there are drivers for AMD Opteron/Athlons and nVidia CUDA-based graphics cards. The basic procedure behind running an OpenCL program is: 1. get a list of platforms 2. choose a device on the platform that fits your needs 3. create a context for the platform and the device 4. compile a program and "kernels" written in OpenCL/C 5. create memory buffer objects. 6. execute kernels on the memory buffer objects. 7. rinse hands, repeat. Obviously this is not very functional, but we can't get there without a raw binding. Data Parallel computing is somewhere Haskell is excelling; hopefully this will generate some interest in creating a DPH binding to OpenCL's C99-based language, OpenCL/C.

2 2

Re: [Haskell-cafe] Can't install Haskell Platform on 64-bit Linux
by Colin Paul Adams 18 Aug '09

18 Aug '09

>>>>> "Magnus" == Magnus Therning <magnus(a)therning.org> writes: Magnus> Ouch, there only seems to be a version of 6.10.3 in Fedora Magnus> 11, if I read this correctly: Magnus> https://admin.fedoraproject.org/pkgdb/packages/name/ghc Magnus> (I'm not a Fedora user so I might be completely confused Magnus> here :-) Magnus> It looks like the pre-built ghc can't find the installed Magnus> libgmp. I don't think it's a problem with libgmp. I think that's just a cautionary message, based on experience that this is a common error. -- Colin Adams Preston Lancashire

1 0

Grouping and SIMD in parallel Haskell (using Nested Data Parallel Haskell ideas in legacy code)
by Zefirov Sergey 18 Aug '09

18 Aug '09

I haven't had enough time to polish a paper about the subject, so I decided to post my results here, in Haskell Café. When Simon Peyton-Jones was in Moscow about a month ago I made a bold statement that Parallel Haskell programs, expressed with the help of par and pseq, can be transformed into Nested Data Parallel Haskell. I wrote a simple model of what will be if we group arguments of par and pseq into queues based on parallel arrays from NDPH. We could then evaluate all thunks from that queues in parallel using SIMD (or just getting higher ILP). We also do not have to lock out main spark queue, we lock only queue we should put argument in. The latter turned out to be beneficial in itself, at least in my simple model. Let us look at famous parallel fib function: Fib n | N <= 1 = 1 | otherwise = a `par` b `pseq` a+b where a = fib (n-1) b = fib (n-2) When we evaluate fib we put a spark of a into spark queue and proceed to evaluate b and, then, a+b. When we put a into spark queue, we lock queue out, modify and release. There is a big probability that threads on different cores will compete for queue lock and some of them (most of them) will waste time waiting for lock release. If we group a's into different queue and put to main spark queue a spark to evaluate a complete group of a's at once, we will get less wasted time. This will work even for single CPU. Below is a run of (fib 15) on my model with cpuCount 1, 4 and 16 and a's or b's group length 0 (no grouping) and 16: cpuCount groupLength=0 groupLength=16 modelTicks modelTicks 1 34535 27189 4 12178 7472 16 7568 3157 speedup 2.19 times 3.11 times ticks1/ticks16 I think results speak for itself. I think the idea of `par` argument grouping could be viable. I should note that I made several digressions when I wrote model. One of digressions is that all evaluations are put into queue. In (a `par` b `pseq` a+b) a put into a_queue, b put into b_queue and (a+b) put into main queue. Each evaluated spark update it's "parent" - a spark that wait for it. Also, main loop of single CPU changed from simple (reading main spark queue + execute when get something) into a series of attepmts with fall back on failure: - first read main queue and execute spark if succeed, - else read current a_queue and execute all sparks there if succeed, - else read current b_queue and execute all sparks there if succeed, - else go to main loop. The new (transformed) code for our fib below: -- |Create a new queue based on parallel array. It holds a parallel array with current arguments and a function that performs -- computation in RTS monad (evaluation function). newQueueParArray :: (x -> RTS ()) -> RTSRef ([: x :],x -> RTS ()) a_queue = unsafePerformIO $ newQueueParArray (\x -> fib (x-1)) -- RTSRef (Int,Int -> Int) b_queue = unsafePerformIO $ newQueueParArray (\x -> fib (x-2)) -- RTSRef (Int,Int -> Int) fib n caller | n <= 1 = 1 | otherwise = unsafePerformIO $ do Ab <- addToMainQueue (defer (+) caller) A' <- addToParArrQueue a_queue x ab B' <- addToParArrQueue b_queue x ab addToMainQueue ab -- add a spark to check a' and b' evaluation status, compute a+b and update the caller. That transfomation cannot be done at the source level using usual type (class/families) hackery. It could be done, though, using core-to-core transformations. It is clear that several values of same type (Int for fib) and a function to perform operations over them leads to SIMD execution. I made some provisions to exploit SIMD and ILP in my model. It can load more than single task information and add several values per cycle. This also speeds execution up, but not so radically (about 3%). The source code for model is here: http://82.146.47.211/attachment/wiki/ndpph/ndp-ph.hs (it's a MskHUG.ru domain, we have temporary problems with DNS). It short of comments, but I tried to make understandable function names. Compile it with 'ghc -o ndp-ph --make -O2 ndp-ph' and run with a command line like the following: ndp-ph.exe cpuCount 1 usePrivateGroups 0 maxABTaskLength 64 taskFetchsPerCycle 1 thunksPerAddition 8 cyclesPerAddition 1 I decided to create a model of exexcution instead of modifying an existing implementation because I have not enough time. I wrote it over evenings and a weekend, so it is simple, it's rude and it does the job pretty fine.

2 2

Can't install Haskell Platform on 64-bit Linux
by Colin Paul Adams 18 Aug '09

18 Aug '09

I installed GHC 6.10.4 (as ./configure said it was required) from a .bz2 file on the GHC downloads page. I then tried ./configure for the platform again, and got: checking ghc actually works... no configure: error: Your installation of ghc does not appear to work. It cannot compile a simple program (see config.log for the details). If you installed ghc from a generic binary tarball then it is worth checking that you have the 'gmp' C library and header files installed. (On debian-based systems this package is called libgmp3-dev.) Looking in config.log, I see: configure:2287: checking ghc actually works configure:2296: /usr/local/bin/ghc -o conftest conftest.hs /tmp/ghc22651_0/ghc22651_0.s: Assembler messages: /tmp/ghc22651_0/ghc22651_0.s:43:0: Error: suffix or operands invalid for `push' /tmp/ghc22651_0/ghc22651_0.s:89:0: Error: suffix or operands invalid for `push' /tmp/ghc22651_0/ghc22651_0.s:135:0: Error: suffix or operands invalid for `push' configure:2299: $? = 1 configure: failed program was: main = putStr "Hello world!\n" -- this file generated by TRY-COMPILE-GHC end of failed program. configure:2308: result: no configure:2314: error: Your installation of ghc does not appear to work. It cannot compile a simple program (see config.log for the details). If you installed ghc from a generic binary tarball then it is worth checking that you have the 'gmp' C library and header files installed. (On debian-based systems this package is called libgmp3-dev.) Libgmp is installed. I don't know the significance of the assembler error messages. If I type from the command line: ghci I get the Prelude> prompt, so it looks like the installation is OK at first glance. What should I check next? -- Colin Adams Preston Lancashire

2 2