Re: [Boost-users] Hybrid parallelism, no more + mpi+serialization, many questions

Subject: Re: [Boost-users] Hybrid parallelism, no more + mpi+serialization, many questions
From: Hicham Mouline (hicham_at_[hidden])
Date: 2010-11-16 15:52:19

> -----Original Message-----
> From: boost-users-bounces_at_[hidden] [mailto:boost-users-
> bounces_at_[hidden]] On Behalf Of David Abrahams
> Sent: 16 November 2010 20:19
> To: boost-users_at_[hidden]
> Subject: Re: [Boost-users] Hybrid parallelism, no more +
> mpi+serialization, many questions
> 
> At Tue, 16 Nov 2010 12:41:08 -0700,
> James C. Sutherland wrote:
> >
> > > Where do I deal with the endianness issue given that I may have
> > > Intel/Sparc/PowerPC CPUs?
> > >
> >
> > Not sure how boost::serialization handles that one...   There are
> probably
> > compiler flags that you can set to change endian-ness if needed
> though.
> 
> IIUC MPI, and thus Boost.MPI, handles it for you transparently.
>
I'm a bit unclear.
MPI uses serialization to serialize user-defined types (you write the
serialize template function).
I don't know if MPI lets you choose if you want a binary archive or a
text/xml archive.
If you can choose the binary archive, wouldn't the issue then be with
serialization and not MPI?

What about primitive types like a double? This will be a quick test.

regards,