NexTech 2021 Congress
October 03, 2021 to October 07, 2021 - Barcelona, Spain

  • UBICOMM 2021, The Fifteenth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies
  • ADVCOMP 2021, The Fifteenth International Conference on Advanced Engineering Computing and Applications in Sciences
  • SEMAPRO 2021, The Fifteenth International Conference on Advances in Semantic Processing
  • AMBIENT 2021, The Eleventh International Conference on Ambient Computing, Applications, Services and Technologies
  • EMERGING 2021, The Thirteenth International Conference on Emerging Networks and Systems Intelligence
  • DATA ANALYTICS 2021, The Tenth International Conference on Data Analytics
  • GLOBAL HEALTH 2021, The Tenth International Conference on Global Health Challenges
  • CYBER 2021, The Sixth International Conference on Cyber-Technologies and Cyber-Systems

SoftNet 2021 Congress
October 03, 2021 to October 07, 2021 - Barcelona, Spain

  • ICSEA 2021, The Sixteenth International Conference on Software Engineering Advances
  • ICSNC 2021, The Sixteenth International Conference on Systems and Networks Communications
  • CENTRIC 2021, The Fourteenth International Conference on Advances in Human-oriented and Personalized Mechanisms, Technologies, and Services
  • VALID 2021, The Thirteenth International Conference on Advances in System Testing and Validation Lifecycle
  • SIMUL 2021, The Thirteenth International Conference on Advances in System Simulation
  • SOTICS 2021, The Eleventh International Conference on Social Media Technologies, Communication, and Informatics
  • INNOV 2021, The Tenth International Conference on Communications, Computation, Networks and Technologies
  • HEALTHINFO 2021, The Sixth International Conference on Informatics and Assistive Technologies for Health-Care, Medical Support and Wellbeing

NetWare 2021 Congress
November 14, 2021 to November 18, 2021 - Athens, Greece

  • SENSORCOMM 2021, The Fifteenth International Conference on Sensor Technologies and Applications
  • SENSORDEVICES 2021, The Twelfth International Conference on Sensor Device Technologies and Applications
  • SECURWARE 2021, The Fifteenth International Conference on Emerging Security Information, Systems and Technologies
  • AFIN 2021, The Thirteenth International Conference on Advances in Future Internet
  • CENICS 2021, The Fourteenth International Conference on Advances in Circuits, Electronics and Micro-electronics
  • ICQNM 2021, The Fifteenth International Conference on Quantum, Nano/Bio, and Micro Technologies
  • FASSI 2021, The Seventh International Conference on Fundamentals and Advances in Software Systems Integration
  • GREEN 2021, The Sixth International Conference on Green Communications, Computing and Technologies

TrendNews 2021 Congress
November 14, 2021 to November 18, 2021 - Athens, Greece

  • CORETA 2021, Advances on Core Technologies and Applications
  • DIGITAL 2021, Advances on Societal Digital Transformation

 


ThinkMind // International Journal On Advances in Systems and Measurements, volume 9, numbers 1 and 2, 2016 // View article sysmea_v9_n12_2016_5


Impact of the Entering Time on the Performance of MPI Collective Operations

Authors:
Christoph Niethammer
Dmitry Khabi
Huan Zhou
Vladimir Marjanovic
José Gracia

Keywords: collectives; late-arrivals; benchmarking; MPI collective operations

Abstract:
Collective operations strongly affect the performance of many MPI applications, as they involve large numbers, or frequently all, of the processes communicating with each other. One critical issue for the performance of collective operations is load imbalance, which causes processes to enter collective operations at different times. The influence of such late-arrivals is not well understood at the moment. Earlier work showed that even small system noise can have a tremendous effect on the collective performance. Thus, although algorithms are optimized for large process counts, they do not seem to tolerate noise or consider delay of involved processes and even a small perturbation from a single process can already have a negative effect on the overall collective execution. In this work, we show a first detailed study about the effect of late arrivals onto the collective performance in MPI. For the evaluation a new, specialized benchmark was designed and a new metric, which we call delay overlap benefit, was used. Our results show that there is already some potential tolerance to late arrivals for the most common collective operations - namely barrier, broadcast, allreduce and alltoall - but there is also a lot of room for future optimizations.

Pages: 48 to 57

Copyright: Copyright (c) to authors, 2016. Used with permission.

Publication date: June 30, 2016

Published in: journal

ISSN: 1942-261x

SERVICES CONTACT
2010 - 2017 © ThinkMind. All rights reserved.
Read Terms of Service and Privacy Policy.