Cluster Multiprocessing

Cluster Multiprocessing

[David McClain]

I am about to embark on a project that seeks to reduce the computation time
of a large problem from 100+ hours to less than 20 minutes. This is
initially though to require a cluster of multiprocessors using either a
Beowulf architecture, or a large, shared memory, message passing,
architecture. I am interested in the suitability (or otherwise) of OCaml to
such computations and the experiences of anyone else in this group on such
architectures. Garbage collection across nodes in a shared memory
architecture seems like a particularly daunting problem.

Any responses would be appreciated.

Thanks,

- D.McClain, Sr. Scientist, Raytheon Systems Co., Tucson, AZ

Re: Cluster Multiprocessing

[reig]

David McClain wrote:
> 
> I am about to embark on a project that seeks to reduce the computation time
> of a large problem from 100+ hours to less than 20 minutes. This is
> initially though to require a cluster of multiprocessors using either a
> Beowulf architecture, or a large, shared memory, message passing,
> architecture. I am interested in the suitability (or otherwise) of OCaml to
> such computations and the experiences of anyone else in this group on such
> architectures. Garbage collection across nodes in a shared memory
> architecture seems like a particularly daunting problem.
> 
> Any responses would be appreciated.
> 
> Thanks,
> 
> - D.McClain, Sr. Scientist, Raytheon Systems Co., Tucson, AZ
> 
> 
> 

Hi,
 
This may not be exactly what you're after, but I'd thought I'd mention
it in case you want to look at it.
 
The compilation manager (CM) for SML/NJ has a parallel and distributed
make option. It seems to use coarse parallelism (multiprocessors, rsh
+ NFS)
 
The manual describes it in section 12
 
http://www.kurims.kyoto-u.ac.jp/~blume/SMLNJ-DEV/manual/index.html
 
Fermin Reig

Windows CE

[Christophe Raffalli]

Is someone working on a prot of Ocaml For Windows CE (at least the
bytecode runtime) ?

-- 
Christophe Raffalli
Université de Savoie
Batiment Le Chablais, bureau 21
73376 Le Bourget-du-Lac Cedex

tél: (33) 4 79 75 81 03
fax: (33) 4 79 75 87 42
mail: Christophe.Raffalli@univ-savoie.fr
www: http://www.lama.univ-savoie.fr/~RAFFALLI

Re: Windows CE

[Francois Rouaix]

> Is someone working on a prot of Ocaml For Windows CE (at least the
> bytecode runtime) ?

I've tried for about 15 minutes (take the Microsoft Embedded Visual C++ 
environment -- you can order for S&H cost only from MSFT !--, move the byterun 
source files into a project, compile).
Although WinCE 3.0 is supposed to be a Win32 environment, it didn't compile 
immediately. I fooled around a bit, but rapidly abandonned (too hard for me, 
as a first attempt at Windows programming ;-)
Note: for the iPaq 36xx, the Linux ports really seem to go forward, with at 
least two companies promoting their own version. I'm waiting for my serial 
sync cable, and I'll let people know when/if I can run OCaml on it...

--f

Re: Cluster Multiprocessing

[Vitaly Lugovsky]

On Thu, 16 Nov 2000, David McClain wrote:

> I am about to embark on a project that seeks to reduce the computation time
> of a large problem from 100+ hours to less than 20 minutes. This is
> initially though to require a cluster of multiprocessors using either a
> Beowulf architecture, or a large, shared memory, message passing,
> architecture. I am interested in the suitability (or otherwise) of OCaml to
> such computations and the experiences of anyone else in this group on such
> architectures. Garbage collection across nodes in a shared memory
> architecture seems like a particularly daunting problem.
> 
> Any responses would be appreciated.

  Shared memory architectures is just the same as ordinary multithereaded
environment. And, there are PVM3 and MPI bindings for OCaml, if you want
to use message-passing clustering.

--

   V.S.Lugovsky aka Mauhuur (http://ontil.ihep.su/~vsl) (UIN=45482254)

Re: Cluster Multiprocessing

[Ward Wheeler]

Hello,
         I've been using CAML and PVM for about 5 years now, most recently 
on a cluster of 256 CPUs in 128 boxes.  The code is used for the 
construction of phylogenetic trees from molecular and other data.  All of 
our code is at ftp.amnh.org /pub/molecular/poy. The setup seems to work 
very well, but as mentioned in a previous note, this is message passing and 
has no shared memory among machines.
         Ward Wheeler

>On Thu, 16 Nov 2000, David McClain wrote:
>
> > I am about to embark on a project that seeks to reduce the computation time
> > of a large problem from 100+ hours to less than 20 minutes. This is
> > initially though to require a cluster of multiprocessors using either a
> > Beowulf architecture, or a large, shared memory, message passing,
> > architecture. I am interested in the suitability (or otherwise) of OCaml to
> > such computations and the experiences of anyone else in this group on such
> > architectures. Garbage collection across nodes in a shared memory
> > architecture seems like a particularly daunting problem.
> >
> > Any responses would be appreciated.
>
>   Shared memory architectures is just the same as ordinary multithereaded
>environment. And, there are PVM3 and MPI bindings for OCaml, if you want
>to use message-passing clustering.
>
>--
>
>    V.S.Lugovsky aka Mauhuur (http://ontil.ihep.su/~vsl) (UIN=45482254)


*********************************
Ward Wheeler
Division of Invertebrate Zoology
American Museum of Natural History
Central Park West @ 79th St.
New York, NY 10024-5192
1-212-769-5754 (Voice)
1-212-769-5783 (FAX)
*********************************

Re: Cluster Multiprocessing

[William Chesters]

David McClain writes:
 > I am about to embark on a project that seeks to reduce the computation time
 > of a large problem from 100+ hours to less than 20 minutes. This is
 > initially though to require a cluster of multiprocessors using either a
 > Beowulf architecture, or a large, shared memory, message passing,
 > architecture. I am interested in the suitability (or otherwise) of OCaml to
 > such computations and the experiences of anyone else in this group on such
 > architectures. Garbage collection across nodes in a shared memory
 > architecture seems like a particularly daunting problem.
 > 
 > Any responses would be appreciated.

Have you seen Starfish?  I haven't ever looked at it in detail but
it's at least in the same general area of ocaml + cluster MP:

http://dsl.cs.technion.ac.il/Starfish/

Re: Cluster Multiprocessing

[jean-marc alliot]

David McClain wrote:

> I am about to embark on a project that seeks to reduce the computation time
> of a large problem from 100+ hours to less than 20 minutes. This is
> initially though to require a cluster of multiprocessors using either a
> Beowulf architecture, or a large, shared memory, message passing,
> architecture. I am interested in the suitability (or otherwise) of OCaml to
> such computations and the experiences of anyone else in this group on such
> architectures. Garbage collection across nodes in a shared memory
> architecture seems like a particularly daunting problem.
>

We have such an experience regarding Air Traffic arithmetic simulation with
conflict resolutions. We use PVM (messages passing), but our problem has a
quite large "granularity". I have a personal experience in games programming
(parallel alpha-beta) and message passing is there less efficient.
My advice on such problems would be the following : if you can "divide" your
problem in quite large independant chunks, the message passing solution is the
one you are looking for. You can build extremely powerful clusters with very
little money, and the messages passing mechanism is extremely easy to use.
On the opposite, such mechanisms will become slow if your problem needs
variable sharing etc...

JMA
(Note : PVM bindings for OCAML are at
ftp://ftp.recherche.enac.fr/pub/SRC/ocamlpvm.tgz)