MPI collective communications on the blue gene/p supercomputer: algorithms and optimizations

Authors:
Ahmad Faraj;Sameer Kumar;Brian Smith;Amith Mamidala;John Gunnels;Philip Heidelberger
Affiliations:
International Business Machines, Rochester, NY, USA;International Business Machines, Yorktown Heights, NY, USA;International Business Machines, Rochester, NY, USA;International Business Machines, Yorktown Heights, NY, USA;International Business Machines, Yorktown Heights, NY, USA;International Business Machines, Yorktown Heights, NY, USA
Venue:
Proceedings of the 23rd international conference on Supercomputing
Year:
2009

Citing 3
Cited 5

Optimization of MPI collective communication on BlueGene/L systems

Proceedings of the 19th annual international conference on Supercomputing
A study of process arrival patterns for MPI collective operations

Proceedings of the 21st annual international conference on Supercomputing
Overview of the IBM Blue Gene/P project

IBM Journal of Research and Development

Optimal bucket algorithms for large MPI collectives on torus interconnects

Proceedings of the 24th ACM International Conference on Supercomputing
Collective algorithms for sub-communicators

Proceedings of the 17th ACM SIGPLAN symposium on Principles and Practice of Parallel Programming
Automatic mapping of parallel applications on multicore architectures using the Servet benchmark suite

Computers and Electrical Engineering
Collective algorithms for sub-communicators

Proceedings of the 26th ACM international conference on Supercomputing
Collectives on two-tier direct networks

EuroMPI'12 Proceedings of the 19th European conference on Recent Advances in the Message Passing Interface

Quantified Score

Hi-index	0.00

Visualization

Abstract

The IBM Blue Gene/P (BG/P) system is a massively parallel supercomputer succeeding BG/L, and it is based on orders of magnitude in system size and significant power consumption efficiency. BG/P comes with many enhancements to the machine design and new architectural features at the hardware and software levels. In this work, we demonstrate techniques to leverage the architectural features of BG/P to deliver high performance MPI collective communication primitives.