GCL, GMP linkage

Michael Koehne kraehe at copyleft.de
Thu Dec 18 20:39:19 CET 2003


Moin Richard Fateman,

> >We have a gcl developer (Michael Koehne) working on something similar.
> Michael: I suggest you look at the specs for Allegro's  "dumplisp" or 
> equivalent, as well as the Lispworks spec.

  my elf::save-system is realy simple. It raises a 'dirty' flag and
  calls si::save-system. The dirty flag will cause a dlopen of every
  library the first time one of my dynamic functions is called. The
  current status is 'it works' - but the syntax of defining a dynamic
  function is not compatible to defining a static function. So one
  idea would be to provide GCL, CLisp's FFI or UFFI look like wrapper
  macros.

  see the elf-loader at http://www.copyleft.de/lisp/

> It is very hard to judge what might be of interest. Some people get very 
> excited about finding the bazillionth digit of pi. [...]
> This is OK,  we all live in niches of some sort....

  my last job was bootstraping the computering department of a fresh
  university (www.clamv.iu-bremen.de). I therefore had to talk to a
  lot of mathematicans. Most of them use MatLab, Mathematic, Mapple
  or SPSS and only a few even know about Octave or Fiasco. None of
  them spoke LISP but they tried hard to code MPI in C++ or Fortran.

  I tried to convince some to GCL/MPI or other free software and
  Computing PI is a good example. Using MPI to compute lisp only
  makes sense, if you have numbers larger than IEEE. Else a simple
  stochatic process over pytagoras would give its best results
  with a few dots and only adds noise to precission afterwards.
  So Octave/MPI did'nt gave any benefit over MatLab. But most
  mathematics stayed shy of LISP, even if GCL/MPI was able to
  compute better numbers.

Bye Michael
-- 
  mailto:kraehe at copyleft.de             UNA:+.? 'CED+2+:::Linux:2.4.22'UNZ+1'
  http://www.xml-edifact.org/           CETERUM CENSEO WINDOWS ESSE DELENDAM


More information about the gmp-discuss mailing list