MPI and compiler technology: a love-hate relationship

Authors:
Anthony Danalis
Affiliations:
Inovative Computing Laboratory, University of Tennessee, Knoxville, TN
Venue:
EuroMPI'12 Proceedings of the 19th European conference on Recent Advances in the Message Passing Interface
Year:
2012

Citing 7
Cited 0

Co-array Fortran for parallel programming

ACM SIGPLAN Fortran Forum
Data-Flow Analysis for MPI Programs

ICPP '06 Proceedings of the 2006 International Conference on Parallel Processing
Quantifying the potential benefit of overlapping communication and computation in large-scale scientific applications

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
MPI-aware compiler optimizations for improving communication-computation overlap

Proceedings of the 23rd international conference on Supercomputing
Communication-Sensitive Static Dataflow for Parallel Message Passing Applications

Proceedings of the 7th annual IEEE/ACM International Symposium on Code Generation and Optimization
Transforming MPI source code based on communication patterns

Future Generation Computer Systems
Exploitation of Dynamic Communication Patterns through Static Analysis

ICPP '10 Proceedings of the 2010 39th International Conference on Parallel Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

MPI has been the de facto standard for parallel computing for over a decade now, and remains mostly unchallenged as the only viable options for harnessing the processing power of massively parallel supercomputers. Alternative parallel programming paradigms have existed for quite some time, mainly in the form of PGAS languages [4,7], but have yet to deliver the necessary performance, robustness and portability needed to drive developers away from MPI.