Re: langage for mapping algorithms on SIMD/MIMD computer?

Prahladavaradan Sampath <prahlad@newsgate.dircon.co.uk>
3 Feb 1996 12:23:30 -0500

          From comp.compilers

Related articles
langage for mapping algorithms on SIMD/MIMD computer? contandr@vlsi.polymtl.ca (nicolas contandriopoulos) (1996-01-31)
Re: langage for mapping algorithms on SIMD/MIMD computer? cdg@nullstone.com (1996-02-01)
langage for mapping algorithms on SIMD/MIMD computer? dave@occl-cam.demon.co.uk (Dave Lloyd) (1996-02-01)
Re: langage for mapping algorithms on SIMD/MIMD computer? prahlad@newsgate.dircon.co.uk (Prahladavaradan Sampath) (1996-02-03)
| List of all articles for this month |
From: Prahladavaradan Sampath <prahlad@newsgate.dircon.co.uk>
Newsgroups: comp.compilers
Date: 3 Feb 1996 12:23:30 -0500
Organization: Direct Connection
References: 96-01-149
Keywords: parallel, C

nicolas contandriopoulos <contandr@vlsi.polymtl.ca> wrote:
> We are looking for a programming language (base on C) which can
>include some partitionning directives (SIMD/MIMD). We are tending
>towards CC, CC++ or C*, but with these languages, we fear that we will
>be unable carry out a manual mapping (therefore a sufficiently fine
>grained mapping).


hi,
      I thought that I would just add, that there is some work going on
at Imperial college using "Coordination forms" to program parallel
computers. This is still research software but, the idea seems to be
quite powerful. Coordination forms are basically a higher-order
functional language which can be used to co-ordinate the execution of
pieces of computation (which can be coded in your favourite language,
I think only C and Fortran are supported at the moment). The
functional language is itself quite restricted and provides a number
of built-in functions to distribute computation and data and only
allows function application and composition.


      Anyway, this probably will be of no practical use to you
currently. Also look up the site:
            http://www-ala.doc.ic.ac.uk/~yg/
It has a number of pointers to other tools and literature for parallel
programming.


regards


Prahlad
--


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.