October 2003 - Glasgow-haskell-users

Re: UArray question....
by heron_carvalho 27 Oct '03

27 Oct '03

Hi, thanks for your answer. In the distributed parallel extension that we are developing, programmers are free to use immutable or mutable arrays and to transmit them between processes (the language have mechanisms to make transparent send/recv operations). Thus, it is important to give efficient marshalling for any array type. With StorableArrays, GHC offers good support, but for immutable arrays types, we have to copy element by element (worse performance). But someone can explain how unboxed arrays are stored in memory ? Heron > AFAIK, no. That's the whole point of StorableArrays. Why can't you use > these instead of UArrays? > > - Hal > > On Mon, 27 Oct 2003, heron_carvalho wrote: > > > Dear colleagues, > > > > In GHC/FFI, is there some way (is it possible ?) t o > > access DIRECTLY an array of UArray type (immutable) i n a > > sequential contiguous memory buffer (in C side, for > > example) without to need to copy the array elements o ne > > by one ? GHC hackers and implementors are wellcome... :-) > > My interest is to transmit the array over a network u sing > > MPI, but avoiding marshalling overheads. > > > > It is easy to do this with StorableArray's in the IO > > monad, of course, but I need to use UArray. > > > > > > Heron de Carvalho > > > > > > > > > > _____________________________________________________ _____________________ > > Acabe com aquelas janelinhas que pulam na sua tela. > > AntiPop-up UOL - É grátis! > > http://antipopup.uol.com.br/ > > > > > > _______________________________________________ > > Glasgow-haskell-users mailing list > > Glasgow-haskell-users(a)haskell.org > > http://www.haskell.org/mailman/listinfo/glasgow- haskell-users > > > > -- > Hal Daume III | hdau me(a)isi.edu > "Arrest this man, he talks in maths." | www. isi.edu/~hdaume > > _______________________________________________ > Glasgow-haskell-users mailing list > Glasgow-haskell-users(a)haskell.org > http://www.haskell.org/mailman/listinfo/glasgow- haskell-users > __________________________________________________________________________ Acabe com aquelas janelinhas que pulam na sua tela. AntiPop-up UOL - É grátis! http://antipopup.uol.com.br/

1 0

Optimization options?
by Juanma Barranquero 27 Oct '03

27 Oct '03

User's Guide 6.2 says: "So, when we want very fast code, we use: -O -fvia-C." but I've seen programs which use the (apparently undocumented) -O2-for-C flag too. In fact, generally speaking anyone of: -O -fvia-C -O2 -fvia-C -O -fvia-C -O2-for-C -O2 -fvia-C -O2-for-C produces different output... So, which is the best choice? Juanma

2 2

UArray question....
by heron_carvalho 26 Oct '03

26 Oct '03

Dear colleagues, In GHC/FFI, is there some way (is it possible ?) to access DIRECTLY an array of UArray type (immutable) in a sequential contiguous memory buffer (in C side, for example) without to need to copy the array elements one by one ? GHC hackers and implementors are wellcome... :-) My interest is to transmit the array over a network using MPI, but avoiding marshalling overheads. It is easy to do this with StorableArray's in the IO monad, of course, but I need to use UArray. Heron de Carvalho __________________________________________________________________________ Acabe com aquelas janelinhas que pulam na sua tela. AntiPop-up UOL - É grátis! http://antipopup.uol.com.br/

2 1

Question about profiling in GHC...
by Heron 26 Oct '03

26 Oct '03

Dear GHC users, I am using GHC profilling tool and I think you can help me to answer a (apparently) simple question: When I compile my program without "-prof -auto-all" option (no profiling support), its execution time is about 140s (compiled with -O2). When compiled with profiling support, the time spent by the program is about 180s (I used my own timer to measure this). Of course, the additional 40s is caused by the profiling annotation code. However, the profiler says (in the ".prof" file produced at the end of execution) that the time spent is about 85s. I suppose that the time measured by the profiler is only for evaluation of the main function of my program (I have compiled all modules with "-prof-all"). But what kind of computation is performed in the rest 95s (180s - 85s) ????? Garbage collection ??? The use of -O2 option have some effect in the profiling ? I am asking this because when I use -O2 when compiling some concurrent Haskell programs that I have implemented in my work, the code assumes strange and not expected behaviour. Best Regards, Heron de Carvalho

1 0

"Job Opening" - Mac OS X 10.2 Binary Packages
by Wolfgang Thaller 24 Oct '03

24 Oct '03

Dear GHC for Mac OS X Users, As you probably know, Mac OS X 10.3 a.k.a. Panther is being officially released today. I'm going to upgrade my Powerbook right away. Panther adds some new functions in Darwin (for example, dlfcn.h and proper wchar support). GHC will automatically make use of these new features, but that means that versions of GHC built on Panther will not work on previous versions of Mac OS X. Which brings me to the point of this mail: When GHC 6.2 is released, I probably won't have access to a Mac running Mac OS 10.2 any more, which means that my binaries for GHC 6.2 will _require_ Mac OS X 10.3. So would someone please step forward and accept the role of GHC packager for Mac OS X 10.2? Cheers, Wolfgang

2 1

Re: "Job Opening" - Mac OS X 10.2 Binary Packages
by Wolfgang Thaller 24 Oct '03

24 Oct '03

On 25.10.2003, at 00:01, Diederik van Arkel wrote: > Not necessary, the 10.3 dev tools includes the headers for 10.1 and > 10.2 so you can compile for all three OS revisions from your 10.3 box. Well, but those are only easily accessible from Apple's Xcode IDE; I have no idea how to get the command-line GCC to use them, let alone how to get GHC's makefiles and configure script to pass the correct flags on to GCC. And as I really dislike digging around in Makefiles and configure scripts, I think it would be much simpler if somebody else did the 10.2 build. Cheers, Wolfgang

1 0

RE: library-guide.ps
by Simon Marlow 24 Oct '03

24 Oct '03

> Simon Marlow points at .../ghc/ .../users_guide.ps.gz > > It is shown near it the hierarchical library guide, > which is also needed and which does not provide .ps. Haddock doesn't produce printable output (yet), so the library documentation is only available in HTML I'm afraid. Cheers, Simon

1 0

SPECIALIZEing something from another module
by Hal Daume III 23 Oct '03

23 Oct '03

Is this possible, or need the specializations go in the defining module. Is this decision based on a fundamental problem with allowing specializations anywhere, or was it just a design choice (in which case I would encourage it's removal)... Thanks! - Hal -- Hal Daume III | hdaume(a)isi.edu "Arrest this man, he talks in maths." | www.isi.edu/~hdaume

1 0

library-guide.ps
by Serge D. Mechveliani 23 Oct '03

23 Oct '03

Thanks. Simon Marlow points at .../ghc/ .../users_guide.ps.gz It is shown near it the hierarchical library guide, which is also needed and which does not provide .ps. Regards, ----------------- Serge Mechveliani mechvel(a)botik.ru

1 0

RE: package treating
by Simon Marlow 23 Oct '03

23 Oct '03

> > > By the way, I somehow lost the way to the Postscript file of the > > > GHC User Guide, how does one get it? > > > > http://www.haskell.org/ghc/documentation.html > > > > With .ps, I sometimes print the Manual to the paper, to book. > Say, print the pages 1 -- 300. > I do not know how to do this with .html. > I can convert some small part to .ps when viewing .html. > But it is desirable to be able to convert it all to one .ps > file. > I am sorry for the ignorance, but does .html replace fully .ps ? The postscript is available from the link I gave you. Here's a direct link: http://www.haskell.org/ghc/docs/latest/users_guide.ps.gz > On last reliable GHC version > ---------------------------- > For any occasion: > it is good to always keep available on public the latest reliable > GHC version. > > For example, the DoCon program failed (so far) with ghc-6.0.1, > due to the gc bug in ghc. > And the DoCon Installing guide says: > "it was tested for ghc-5.02.2 ...". > At least, DoCon user should be able to download and install this > ghc-5.02.2 (in this particular case, I suspect, 5.04 fits). We keep all previous GHC distributions available onlin. Version 5.02.2 is still available here: http://www.haskell.org/ghc/download_ghc_502.html (this is linked from the download page). Cheers, Simon

1 0