March 2005 - Glasgow-haskell-users

RE: Kind of StableName
by Simon Marlow 29 Mar '05

29 Mar '05

On 28 March 2005 11:25, Nicolas Oury wrote: > I try to use StableNames for hash consing on complex types - I have to > mix in a same table StableNames of different types - and I wonder why > StableName is of kind * -> * and not *. I think we gave StableName the extra type parameter to help with type safety. But I can't think of a convincing example right now... anyway, it would be quite easy to build an alternative StableName implementation without the type parameter. Take a look at the source code for System.Mem.StableName, it's quite short. Cheers, Simon

1 0

RE: Biographical Profiling
by Simon Marlow 29 Mar '05

29 Mar '05

On 25 March 2005 15:59, Andreas Marth wrote: > The ghc user guide says that there are 4 states a heap object may be > in. Unfortunately in my program there are 5 states (inherent_use, > use, drag, lag, void) with inherent_use being the biggest part (many > MBs) and the one not mentioned in the users guide. > My question is what does it mean? What can I do to get rid of it? IIRC, inherently used closures are primitive objects that we don't record use-stamps on. This includes things like the stack (part of a primitive thread object), arrays, MVars, and IORefs. Cheers, Simon

1 0

RE: [Haskell] Haskell 6.4 perfomance
by Simon Marlow 29 Mar '05

29 Mar '05

GHC 6.4 uses C-- as an intermediate language, it doesn't compile via C-- (yet). A full C-- backend has the potential to generate faster code that going via GCC, because C-- would have control over the stack and the calling convention. In my experience, extra optimisation flags for gcc make very little difference for GHC-compiled code, and may even make things worse. By all means do some benchmarks, though! Cheers, Simon On 25 March 2005 08:19, Alexandre wrote: > Keean, thank you. > > Does any tests/benchmarks available? > > Regards, > /Alexandre. > > > On Mar 25, 2005, at 00:47, Keean Schupke wrote: > >> Think this should really go to glasgow-haskell-users... >> >> If this is true - how do I get ghc to use C--, and is it really >> faster than using gcc as a backend with all the bells & whistles >> turned on (for a pentium-III) something like >> >> "-O3 -mcpu=pentium3 -march=pentium3 -pipe -fomit-frame-pointer >> -momit-leaf-frame-pointer -ftracer -fno-crossjumping -mfpmath=sse,387 >> -ffast-math -fsched-spec-load -fprefetch-loop-arrays >> -maccumulate-outgoing-args -fmove-all-movables >> -freduce-all-givs" >> >> Keean. >> >> Alexandre wrote: >> >>> As I heard, 6.4 version of the Haskell using C-- backend and make >>> lots of the resulting code perfomance (programs executed faster). >>> If so, does any test/comparison with other languages available? >>> >>> Thank you in advance, >>> Regards, >>> /Alexandre. >>> >> _______________________________________________ >> Haskell mailing list >> Haskell(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell >> > > > _______________________________________________ > Glasgow-haskell-users mailing list > Glasgow-haskell-users(a)haskell.org > http://www.haskell.org/mailman/listinfo/glasgow-haskell-users

1 0

moving from ghc-6.2 to 6.4
by Johannes Waldmann 29 Mar '05

29 Mar '05

I am trying to bring a larger heap of code (http://141.57.11.163/auto/ ) into 6.4 land (because of wonder stories about faster compilation, faster execution, Data.Map, and so on ...) Here are a few observations and questions that may be useful to others as well. * what is the situation with ghc-6.4 for sparc/solaris? I don't see a binary package in the download area. I started to build from source - can this be successful? (The rest of this report refers to i386/linux) * Cabal is very nice! - The only thing that was confusing me is that I have to list all modules in the *.cabal file: if I don't, it still happily builds and installs the package but it cannot be used, giving linker errors. Couldn't this be checked earlier? Or better, couldn't it infer the needed hidden modules? Anyway I can generate the module list by a shell script but that does not feel right. - How do I build and install a profiling version of a package, how does Cabal support this? * (when building HSQL-1.4) CTime is no longer an instance of Integral, so I replaced fromIntegral by fromIntegral . fromEnum, is this OK? * the name of mkAppTy changed to mkTyConApp (I hope it's only the name that changed) * I don't see "dramatic" improvements in execution times - are there some magic ghc options I missed? I used -O -fvia-C. Still, executables are maybe 2 .. 5 % smaller and faster than they were with 6.2 - and compilation without -O is really fast. Best regards, and thanks to the GHC team. -- -- Johannes Waldmann -- Tel/Fax (0341) 3076 6479/80 -- ---- http://www.imn.htwk-leipzig.de/~waldmann/ -------

1 0

Kind of StableName
by Nicolas Oury 28 Mar '05

28 Mar '05

Hello, I try to use StableNames for hash consing on complex types - I have to mix in a same table StableNames of different types - and I wonder why StableName is of kind * -> * and not *. Best regards, Nicolas Oury

1 0

static linking with ghc
by Patrick Scheibe 27 Mar '05

27 Mar '05

Hi, I have a programm which is linked against some shared object libraries. Now I want to run my prog on a different machine (but same type, x86 linux). So I took the compiled prog and the needed .so files, put both in one directory and tried it on the other machine. That worked fine. Is it possible to tell ghc to take all stuff out of the .so libs and compile my source to a "all-including" executable? I know that ghc just calls the linker which is doing this part. The --static flag for the linker just says: "take the static library if possible for linking...". But I don't have a static version of these libs and I don't want to recompile all libraries. Am I totally wrong? Is this possible in general? Are there some good papers except of the "man ld"?? Cheers Patrick

2 1

Biographical Profiling
by Andreas Marth 25 Mar '05

25 Mar '05

Hallo! The ghc user guide says that there are 4 states a heap object may be in. Unfortunately in my program there are 5 states (inherent_use, use, drag, lag, void) with inherent_use being the biggest part (many MBs) and the one not mentioned in the users guide. My question is what does it mean? What can I do to get rid of it? Thanks, Andreas

1 0

Problem starting GHC 6.4 on windows 98
by J L Russell 25 Mar '05

25 Mar '05

When trying to run GHC 6.4 (from the installer) on Windows 98, I get the following error message: "The GHC.EXE file is linked to missing export SHELL32.DLL:SHGetFolderPathA" >From what I've been able to determine, on Windows version prior to Windows 2000, this function was instead exported from shfolder.dll, which is retained in later versions of Windows for backward compatibility, and on newer version of Windows, simply calls the version in shell32.dll. So it looks to me like GHC 6.4 from the installer won't be able to run on Windows 98 unless the executable in the installer is updated to link this function from shfolder.dll instead. -James

3 2

Re: [Haskell] Haskell 6.4 perfomance
by Keean Schupke 25 Mar '05

25 Mar '05

Think this should really go to glasgow-haskell-users... If this is true - how do I get ghc to use C--, and is it really faster than using gcc as a backend with all the bells & whistles turned on (for a pentium-III) something like "-O3 -mcpu=pentium3 -march=pentium3 -pipe -fomit-frame-pointer -momit-leaf-frame-pointer -ftracer -fno-crossjumping -mfpmath=sse,387 -ffast-math -fsched-spec-load -fprefetch-loop-arrays -maccumulate-outgoing-args -fmove-all-movables -freduce-all-givs" Keean. Alexandre wrote: > As I heard, 6.4 version of the Haskell using C-- backend and make lots > of the resulting code perfomance (programs executed faster). > If so, does any test/comparison with other languages available? > > Thank you in advance, > Regards, > /Alexandre. >

2 1

RE: comments on the GHC API
by Simon Marlow 23 Mar '05

23 Mar '05

On 22 March 2005 17:49, Matthias Neubauer wrote: > ... we also need the state after renaming, but we then have to stop, I > suppose! Yes, the HaRE folks also want the renamed abstract syntax too. We haven't quite figured out how best to make it available: renaming is intertwined with typechecking a bit. > Because we want to do do our own typechecking, stopping after ghc's > typechecking would be to late: in case of a type error, ghc falls back > into the wrong context (at least that's what's happening right now > when doing a :load inside ghci ...). > >>> We, for our part, would need parts of the intermediate result >>> (that is, the type environment and class declarations) existing >>> either before parsing A, after parsing A, or, even better, after >>> renaming A. Other users may be interested in other intermediate >>> results. > > Is there a way to get access to all the identifiers that are > *imported* into the current module? Having that piece of information > at hand, we could then use *lookupThing* to cumulate the initial type > environment and the initial class environment for our analysis. > > Or do you see another/better way of doing that? The ModIface contains the usages, which includes everything that is *used* by the current module (not quite the same as everything that is imported, because this is a list of original names, not imported names). There is also the mi_globals list (in ModIface), which contains everything in scope at the top level of the module. From that I guess you could find everything that is imported by subtracting things that are defined in the local module. Cheers, Simon

1 0