Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: [OMPI devel] Amateur Guidance
From: Timothy Hayes (hayesti_at_[hidden])
Date: 2008-11-02 12:18:46


Hello everyone,

I'm a regular OpenMPI user but I'm new to the strange world of development
and hence this mailing list. I'm currently working on a project that
involves OpenMPI and I was wondering if I might get some guidance and
pointers in the right direction.
In brief, my project is trying to make an efficient transport module in
OpenMPI for use in virtualised environments (specifically Xen). At the
moment, Xen has a very inefficient way of inter-domain communication, and
although a myriad of solutions have appeared, they are all usually quite
application specific. I'm basing my project on a paper I read about
efficient message passing in interdomain virtualised guests.
The problem I'm having is jumping into the OpenMPI code. I've read two
papers I found on the homepage: "Open MPI: Goals, Concept, and Design of a
Next Generation MPI Implementation" and "TEG: A High-Performance, Scalable,
Multi-Network Point-to-Point Communications Methodology" which gave me some
insight about the MCA, PML and PTL. However, I'm finding it quite difficult
to get a foothold into the codebase and I'm wondering if anyone might be
able to point me to a guide or some documentation that might help get me
started.

I have a vague idea that I'm going to need to make a specific PML that uses
my own PTL for inter-domain Xen guests. Should the PTL fail (e.g. if the
guest migrated to another physical machine), the PML would then switch to
TCP/IP instead. Migrating back and switching back to my own PTL sounds
trickier. I've read there's an out of band communication socket between all
guests for this sort of thing, but once again I'm a bit lost in the detail.

I'm very eager to do this project well and contribute to the OpenMPI
community, and if anyone has some advice or pointers I'd really appreciate
it.

Kind regards

Tim