[Beowulf] MS HPC... Oh dear...

[James Cownie]

On 12 Jun 2006, at 15:49, Joe Landman wrote:

>
>
> Ashley Pittman wrote:
>> On Mon, 2006-06-12 at 00:02 -0400, Joe Landman wrote:
>>
>>> What Microsoft will do is to take away as much of this as they  
>>> can.  I
>>> haven't seen it yet, but I believe they will offer MPICH as a  
>>> DLL, so if
>>> PathScale wants to work along side some other device, you can select
>>> this at runtime, and just have it work.  This is a nice idea.
>>
>> Perhaps I've missed something here, what do windows DLLs provide  
>> that a
>> linux .so doesn't?
>
> Nothing.  Unfortunately most folks use statically linked binaries for
> MPI, so .so's are not a factor.  I could be wrong, and maybe there  
> is a
> way to get statically linked binaries to respect LD_PRELOAD or
> LD_LIBRARY_PATH, but I am not aware of it.
>
> More to the point, this dynamic binding allows you to write to the  
> API,
> present a consistent ABI, and handle the hardware details elsewhere  
> in a
> driver which can be linked in by the .so/.dll/.eieio method at  
> runtime.
>  Which is about the complexity that most end users/customers want.

As I understand it this is what Intel's MPI (for LInux) does, so that  
you can choose the underlying hardware transport without relinking  
your code.

http://www.intel.com/cd/software/products/asmo-na/eng/cluster/mpi/ 
index.htm

(FWIW I work for Intel, but not on MPI, and I certainly don't speak  
for them ;-)

--
-- Jim
--
James Cownie <jcownie at cantab.net>