Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

high-performance logic computing via web

0 views
Skip to first unread message

Alan Baljeu

unread,
Dec 8, 2009, 2:28:09 PM12/8/09
to
I'm exploring this cloud concept (parallel computing grid or whatever)
as a way to scale up computations. I don't have specific performance
requirements, but I want to know what's available for production use:
languages and platforms.

The problems I'm solving have arbitrary complexity, and can choose
more advanced problems if I have faster computational resources, and/
or I can spend less effort optimizing generated problem sets, and thus
solve more problems faster.

I currently have algorithms in Prolog and CHR, and am looking for
tools to take these computations and run them out there. These
languages are supposed to be amenable to implicit parallelization, and
I am open to recoding my algorithms to get this right.

So what are some good options?

Sicstus, Eclipse Prolog come to mind.

Erlang seems like a strong option. (It looks like Prolog. I'm not sure
whether it behaves like it, but that's okay).

Possibly an Operations Research platform could be useful, although
they don't tend to have strong abstractions, and I'd want a symbolic
language to manage it.

What else?

Ulrich Neumerkel

unread,
Dec 9, 2009, 8:17:54 AM12/9/09
to
Alan Baljeu <alanb...@yahoo.com> writes:
>I'm exploring this cloud concept (parallel computing grid or whatever)
>as a way to scale up computations.
...
>What else?

SWI on Condor.

A.L.

unread,
Dec 9, 2009, 10:09:25 AM12/9/09
to
On Tue, 8 Dec 2009 11:28:09 -0800 (PST), Alan Baljeu
<alanb...@yahoo.com> wrote:

>I'm exploring this cloud concept (parallel computing grid or whatever)
>as a way to scale up computations. I don't have specific performance
>requirements, but I want to know what's available for production use:
>languages and platforms.
>
>The problems I'm solving have arbitrary complexity, and can choose
>more advanced problems if I have faster computational resources, and/
>or I can spend less effort optimizing generated problem sets, and thus
>solve more problems faster.
>
>I currently have algorithms in Prolog and CHR, and am looking for
>tools to take these computations and run them out there. These
>languages are supposed to be amenable to implicit parallelization, and
>I am open to recoding my algorithms to get this right.
>
>So what are some good options?
>
>Sicstus, Eclipse Prolog come to mind.

Mozart-OZ

http://www.mozart-oz.org/

It is multiparadigm language that covers logic programming, constraint
programming and distributed programming. There is excellent book about
programming in Mozart

Concepts, Techniques, and Models of Computer Programming (2004)
Peter Van Roy and Seif Haridi

Mozart-OZ is free

You can also try SICSts and Linda - SICStus has pretty good
implementation of Linda

A.L.

Alan Baljeu

unread,
Dec 9, 2009, 6:11:36 PM12/9/09
to
On Dec 9, 8:17 am, ulr...@mips.complang.tuwien.ac.at (Ulrich
Neumerkel) wrote:
Thanks for the suggestions, Ulrich, but I don't understand this
tool.

Condor analyzes conditional transformations, and runs in SWI.
Where do I get these transforms from and what would I do with the
analysis?

0 new messages