EuroMPI 2015 is presented in cooperation with ACM and SIGHPC in Bordeaux France, 21st – 23rd September, 2015.
EuroMPI is the prime annual meeting for researchers, developers, and students in message-passing parallel computing with MPI and related paradigms.
Deadline of early registration is Sept 1st, 2015
The conference will feature 14 strong technical paper presentations, 3 invited talks, 3 tutorials and posters. The detailed conference information is being incrementally updated at the main conference web site.
TUTORIALS
- Performance analysis for High Performance Systems, François Trahay, Telecom SudParis
- Understanding and managing hardware affinities with Hardware Locality (hwloc), Brice Goglin, INRIA Bordeaux sud-Ouest
- Insightful Automatic Performance Modeling, Alexandru Calotoiu, TU Darmstadt
INVITED TALKS
- A new high performance network for HPC systems: Bull eXascale Interconnect (BXI), Jean-Pierre Panziera, Chief Technology Director for Extreme Computing at Atos
- Computational Fluid Dynamics and High performance computing, Gabriel Staffelbach, Senior Researcher at CERFACS
- Is your software ready for exascale? – How the next generation of performance tools can give you the answer, Prof. Felix Wolf, Department of Computer Science of TU Darmstadt
IMPORTANT DATES
- Early Registration Deadline: September 1st, 2015
- Tutorials: September 21st, 2015
- Conference: September 22nd-23rd, 2015
Hi Jeff,
could you made a blog post about how the MPI standard specifies constants, or, e.g., the actual type of “magic” like MPI_DATATYPE?
I’m in particular interested in what does the standard guarantees, and what problems this introduces for e.g. those trying to write wrappers to the MPI C interface for other languages like Rust, Python, Julia,… For example, one typically needs a wrapper for each implementation, which complicates things in typed languages like Rust or Haskell.
Does the committee see this as a problem? Is there a plan to make it easier to use current MPI implementations from other languages? (E.g. having a single ABI at least per platform).
I would love to hear your thoughts about this.
Best regards
Yes, that sounds like a great idea — I’ll write about this soon.
Hi Jeff, will there be an adoption of the functional (lambda-calculus) programming model for MPI 3.1?
Hi Anonymous.
No, there is no current proposal in front of the MPI Forum for a lambda-calculus-based type of model. Do you have some ideas that could turned into a proposal to be presented?
It would be wonderful (and useful) if “lambda-calculus-based” type model be used to for applications requiring sparse matrices in parallel.
Thanks for your comments, John! 🙂