Improving reactivity and communication overlap in MPI using a generic I/O manager

Authors:
François Trahay;Alexandre Denis;Olivier Aumage;Raymond Namyst
Affiliations:
INRIA, LaBRI, Université Bordeaux 1, Talence, France;INRIA, LaBRI, Université Bordeaux 1, Talence, France;INRIA, LaBRI, Université Bordeaux 1, Talence, France;INRIA, LaBRI, Université Bordeaux 1, Talence, France
Venue:
PVM/MPI'07 Proceedings of the 14th European conference on Recent Advances in Parallel Virtual Machine and Message Passing Interface
Year:
2007

Citing 7
Cited 2

Scheduler activations: effective kernel support for the user-level management of parallelism

ACM Transactions on Computer Systems (TOCS)
Polling watchdog: combining polling and interrupts for efficient message handling

ISCA '96 Proceedings of the 23rd annual international symposium on Computer architecture
Integrating Kernel Activations in a Multithreaded Runtime System on Top of LINUX

IPDPS '00 Proceedings of the 15 IPDPS 2000 Workshops on Parallel and Distributed Processing
Integrating polling, interrupts, and thread management

FRONTIERS '96 Proceedings of the 6th Symposium on the Frontiers of Massively Parallel Computation
Quantifying the potential benefit of overlapping communication and computation in large-scale scientific applications

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
Measuring MPI send and receive overhead and application availability in high performance network interfaces

EuroPVM/MPI'06 Proceedings of the 13th European PVM/MPI User's Group conference on Recent advances in parallel virtual machine and message passing interface
Open MPI: a flexible high performance MPI

PPAM'05 Proceedings of the 6th international conference on Parallel Processing and Applied Mathematics

A speculative and adaptive MPI rendezvous protocol over RDMA-enabled interconnects

International Journal of Parallel Programming
A high performance superpipeline protocol for infiniband

Euro-Par'11 Proceedings of the 17th international conference on Parallel processing - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

MPI applications may waste thousands of CPU cycles if they do not efficiently overlap communications and computation. In this paper, we present a generic and portable I/O manager that is able to make communication progress asynchronously using tasklets. It chooses automatically the most appropriate communication method, depending on the context: multi-threaded application or not, SMP machine or not. We have implemented and evaluated our I/O manager with Mad-MPI, our own MPI implementation, and compared it to other existing MPI implementations regarding the ability to efficiently overlap communication and computation.