MPI at Exascale: Challenges for Data Structures and Algorithms

  • Authors:

  • William Gropp

  • Affiliations:

  • University of Illinois at Urbana-Champaign, USA

  • Venue:
  • Proceedings of the 16th European PVM/MPI Users' Group Meeting on Recent Advances in Parallel Virtual Machine and Message Passing Interface
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Petascale computing is here and MPI continues to succeed as a effective, scalable programming model, despite previous predictions that MPI could not work at the Petascale. As the high-performance computing community considers Exascale systems, will MPI continue to be suitable? Already, a number of studies have looked at the behavior of MPI as the number of tasks approaches one million, and have identified scaling problems in the current state-of-the-art. In addition, the MPI Forum has begun to consider the scalability of the definition of some of the routines in MPI, based on the sizes of their input arguments. Concerns have also been raised about other routines, based on the scalability of current implementations.