- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] multi-core programming in Haskell
by Galchin, Vasili 24 Aug '08

24 Aug '08

Thank you Murray. My post was not so clear .... I was referring to "automatic" parellelization vs "manual" parallelization. By "automatic" I mean the programmer doesn't have to indicate where to parallelize ... instead the compiler decides how to parallize! Vasili On Sat, Aug 23, 2008 at 12:58 AM, Murray Gross <mgross21(a)verizon.net> wrote: > > > Vasili: > > Each "par" "sparks" a new thread, which is then queued for execution. At > appropriate points, the threads are distributed to available (free) > processors (cores). The result is that parallelization scales automatically > with the number of available processors. Take a look at the GPH site for > papers that will provide more information on how parallel (and distributed) > Haskell does things. > > Best, > > Murray Gross, > Brooklyn College > > > > > > On Fri, 22 Aug 2008, Galchin, Vasili wrote: > > Hello, >> >> With pure side of the Haskell house, there is hope that the generated >> code could automagically scale as more cores are added yes? It seems that >> it >> is on the stateful monadic side of the house in an appplication that it is >> the programmer responsibility to design the software so that it scales >> across increasing cores? (I am assuming that things like "par" construct >> are >> monadic). On Monday, I am starting a several month project with a company. >> Alledgely some of the code will be written in Python. I would like engage >> the manager in a discussion about multi-core "enabling" the code now when >> we >> design and implement not later as an afterthought. Seems like a gnarly >> subject given current "state-of-the-art" software tools. Ideas?! >> >> Regards, Vasili >> >>

3 2

Why doesn't this work?
by Andrew Coppin 24 Aug '08

24 Aug '08

colour_grid :: (Particle -> IO ()) -> Grid ph -> IO () colour_grid fn g = sequence_ $ runST $ do ps <- grid_coords g mapM (\pix -> do particle <- read_grid g pix return $ fn particle ) ps When I attempt to run this, GHCi just gives me a very cryptic type checker error. I can't figure out what's wrong here. As far as I can tell, this should run...

5 5

(no subject)
by john lask 24 Aug '08

24 Aug '08

Hi has anyone had any success in running a recent version of hs-plugins on a windows platform. I refer to this post to the haskell list in December last year. The current version of hs-plugins dosn't appear to work either. Same problem ... http://www.haskell.org/pipermail/haskell/2007-December/020043.html jvl _________________________________________________________________ Meet singles near you. Try ninemsn dating now! http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fdating%2Eninemsn%2Ecom%2Eau…

1 0

Type family fun
by Chris Eidhof 24 Aug '08

24 Aug '08

Hey all, I was playing around with type families, and I have a strange problem. Suppose we have an alternative to an Either datatype: > data (:|:) a b = Inl a | Inr b and a class Ix: > class Ix i where > type IxMap i :: * -> * > empty :: IxMap i [Int] Now I want to give an instance for (a :|: b): > instance (Ix l, Ix r) => Ix (l :|: r) where > type IxMap (l :|: r) = BiApp (IxMap l) (IxMap r) > empty = BiApp empty empty BiApp is defined as following: > data BiApp a b c = BiApp (a c) (b c) However, it looks like the recursive calls to empty can't be unified, I get the following error message: Couldn't match expected type `IxMap l' against inferred type `IxMap i' Expected type: IxMap (l :|: r) [Int] Inferred type: BiApp (IxMap i) (IxMap i1) [Int] In the expression: BiApp empty empty In the definition of `empty': empty = BiApp empty empty In the inferred type, there should be IxMap l instead of IxMap i, does anybody know what I'm doing wrong? Thanks, -chris

4 3

ANN: benchpress 0.2.1
by Johan Tibell 24 Aug '08

24 Aug '08

I'm pleased to announce the first public release of benchpress. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/benchpress benchpress is a micro-benchmark library that produces statistics such as min, mean, standard deviation, median, and max execution time. It also computes execution time percentiles. Release bundle: http://hackage.haskell.org/packages/archive/benchpress/0.2.1/benchpress-0.2… Docs: http://hackage.haskell.org/packages/archive/benchpress/0.2.1/doc/html/Test-… Code: git clone git://github.com/tibbe/benchpress.git An example benchmark: > import qualified Data.ByteString as B > import System.IO > import Test.BenchPress > > inpath, outpath :: String > inpath = "/tmp/infile" > outpath = "/tmp/outfile" > > blockSize :: Int > blockSize = 4 * 1024 > > copyUsingByteString :: Handle -> Handle -> IO () > copyUsingByteString inf outf = go > where > go = do > bs <- B.hGet inf blockSize > let numRead = B.length bs > if numRead > 0 > then B.hPut outf bs >> go > else return () > > main :: IO () > main = bench 100 $ do > inf <- openBinaryFile inpath ReadMode > outf <- openBinaryFile outpath WriteMode > copyUsingByteString inf outf > hClose outf > hClose inf And the output, best viewed using a fixed-width font: $ ./example Times (ms) min mean +/-sd median max 232.774 273.611 53.317 266.648 722.332 Percentiles (ms) 50% 266.644 66% 266.826 75% 269.616 80% 295.040 90% 295.360 95% 305.855 98% 350.742 99% 450.855 100% 722.332 Cheers, Johan

2 5

Lines of code metrics
by Dimitry Golubovsky 24 Aug '08

24 Aug '08

Hi, Greg Fitzgerald wrote: >Does anyone know of a good case study comparing a project written in C >versus one written in Haskell? I'm mostly looking for a comparison of lines >of code, but any other metric, such as time to market and code quality >metrics could also be Just curious, has anybody tried to apply Halstead's code metrics (see e. g. http://www.sei.cmu.edu/str/descriptions/halstead_body.html) but there is a 70s' book titled "Elements of Software Science" to Haskell and other functional languages vs. C and other imperative languages? I myself played with these calculations in late 80s trying to estimate code quality of Pascal programs on PDP-11, but that was a pain to count functions' operands properly as they might come from global variables. Application of these formulas to functional languages might be mich cleaner, so has anybody tried? Thanks. -- Dimitry Golubovsky Anywhere on the Web

3 2

parsec, hGetContents
by brian 23 Aug '08

23 Aug '08

I'm writing an NNTP client API. When I connect with connectTo, I get a Handle 'h'. I immediately hGetContents h to get a string 'c' that I hope contains all the data the server will ever send. I put h and c in a data Connection and pass it to the functions that send or receive. I send a command and receive a response. The response might look like "200 OK\r\n". Parsec parses c. No problem. But if I send another command and receive another response, parsec parses the same string because (I think) I haven't trimmed the part I parsed on the previous command. Maybe c looks like "200 OK\r\n200 second thing\r\n". So I was thinking I need to parse and return the unparsed part so that the next parse will work. Parsec has a getInput that returns the unparsed part, but it blocks if there is no more to parse. That is, after I parse "200 OK\r\n", if I ask parsec for the "rest", it blocks because there isn't any "rest" yet. How should I be doing this? Before people suggest I hGetLine and parse that, I don't like hGetLine for this because 1) it assumes '\n' signifies the end of a line, whereas in my case the protocol specifies "\r\n" and 2) if the "line" doesn't have an ending, I can't tell it from what hGetLine returns. That is, hGetLine on a handle having "abc\n" returns the same thing as on a handle having "abc". Also, if I can parse c directly, I can write parsers (especially those for multiline respones) much more naturally than I can if I have to get lines separately. Thanks very much for any help.

1 0

"Real World Haskell" hits a milestone
by Bryan O'Sullivan 23 Aug '08

23 Aug '08

Hi, all - I'm proud to report that John, Don and I have finished the draft manuscript of our book. It is now available online in its entirety. For more details, see here: http://www.realworldhaskell.org/blog/2008/08/22/our-writing-is-now-complete/ We expect the final book to be published around the beginning of November, and to weigh in at about 700 pages. Thank you to everyone in the Haskell community who has encouraged and assisted us so far!

3 3

multi-core programming in Haskell
by Galchin, Vasili 22 Aug '08

22 Aug '08

Hello, With pure side of the Haskell house, there is hope that the generated code could automagically scale as more cores are added yes? It seems that it is on the stateful monadic side of the house in an appplication that it is the programmer responsibility to design the software so that it scales across increasing cores? (I am assuming that things like "par" construct are monadic). On Monday, I am starting a several month project with a company. Alledgely some of the code will be written in Python. I would like engage the manager in a discussion about multi-core "enabling" the code now when we design and implement not later as an afterthought. Seems like a gnarly subject given current "state-of-the-art" software tools. Ideas?! Regards, Vasili

2 1

question about uploads of code contribution
by Galchin, Vasili 22 Aug '08

22 Aug '08

Hello, 1) I want to upload a version with minor changes. Should I send out an "announcement"? 2) does the hackage database have a "reader/writer" lock to protect readers, i.e. people downloading when I am uploading? Regards, Vasili

3 3