skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: HPC Node Performance and Energy Modeling with the Co-Location of Applications

Abstract

Multicore processors have become an integral part of modern large-scale and high-performance parallel and distributed computing systems. This advance became necessary as the demand for increased system performance has exceeded the limits that single core processors can provide. Unfortunately, applications co-located on multicore processors can suffer from decreased performance and increased dynamic energy use as a result of access to shared resources, such as memory. Consequently, it is increasingly important to characterize the performance of applications that execute on these architectures. This work investigates some of the disadvantages of co-location, and presents a methodology for building models capable of utilizing varying amounts of information about a target application and its co-located applications to make predictions about the target application’s execution time and the system’s energy use under arbitrary co-locations of a wide range of application types. The proposed methodology is validated on three different server class Intel Xeon multicore processors using eleven applications from two scientific benchmark suites. The model’s utility for scheduling is also demonstrated in a simulated large-scale high-performance computing environment through the creation of a co-location aware scheduling heuristic. This heuristic demonstrates that scheduling using information generated with the proposed modeling methodology is capable of making significantmore » improvements over a scheduling heuristic that is naive to co-location interference.« less

Authors:
 [1];  [1]; ORCiD logo [2];  [1];  [1];  [3];  [3];  [1]
  1. Colorado State University
  2. BATTELLE (PACIFIC NW LAB)
  3. Georgia Institute Of Technology
Publication Date:
Research Org.:
Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1557684
Report Number(s):
PNNL-SA-118182
DOE Contract Number:  
AC05-76RL01830
Resource Type:
Journal Article
Journal Name:
Journal of Supercomputing
Additional Journal Information:
Journal Volume: 72; Journal Issue: 12
Country of Publication:
United States
Language:
English
Subject:
performance modeling, energy modeling, resource management, memory interference, application co-location,

Citation Formats

Dauwe, Daniel, Jonardi, Eric, Friese, Ryan D., Pasricha, Sudeep, Maciejewski, Anthony A., Bader, David A., Bader, David A., and Siegel, Howard J. HPC Node Performance and Energy Modeling with the Co-Location of Applications. United States: N. p., 2016. Web. doi:10.1007/s11227-016-1783-y.
Dauwe, Daniel, Jonardi, Eric, Friese, Ryan D., Pasricha, Sudeep, Maciejewski, Anthony A., Bader, David A., Bader, David A., & Siegel, Howard J. HPC Node Performance and Energy Modeling with the Co-Location of Applications. United States. doi:10.1007/s11227-016-1783-y.
Dauwe, Daniel, Jonardi, Eric, Friese, Ryan D., Pasricha, Sudeep, Maciejewski, Anthony A., Bader, David A., Bader, David A., and Siegel, Howard J. Thu . "HPC Node Performance and Energy Modeling with the Co-Location of Applications". United States. doi:10.1007/s11227-016-1783-y.
@article{osti_1557684,
title = {HPC Node Performance and Energy Modeling with the Co-Location of Applications},
author = {Dauwe, Daniel and Jonardi, Eric and Friese, Ryan D. and Pasricha, Sudeep and Maciejewski, Anthony A. and Bader, David A. and Bader, David A. and Siegel, Howard J.},
abstractNote = {Multicore processors have become an integral part of modern large-scale and high-performance parallel and distributed computing systems. This advance became necessary as the demand for increased system performance has exceeded the limits that single core processors can provide. Unfortunately, applications co-located on multicore processors can suffer from decreased performance and increased dynamic energy use as a result of access to shared resources, such as memory. Consequently, it is increasingly important to characterize the performance of applications that execute on these architectures. This work investigates some of the disadvantages of co-location, and presents a methodology for building models capable of utilizing varying amounts of information about a target application and its co-located applications to make predictions about the target application’s execution time and the system’s energy use under arbitrary co-locations of a wide range of application types. The proposed methodology is validated on three different server class Intel Xeon multicore processors using eleven applications from two scientific benchmark suites. The model’s utility for scheduling is also demonstrated in a simulated large-scale high-performance computing environment through the creation of a co-location aware scheduling heuristic. This heuristic demonstrates that scheduling using information generated with the proposed modeling methodology is capable of making significant improvements over a scheduling heuristic that is naive to co-location interference.},
doi = {10.1007/s11227-016-1783-y},
journal = {Journal of Supercomputing},
number = 12,
volume = 72,
place = {United States},
year = {2016},
month = {12}
}

Works referenced in this record:

Power-aware dynamic placement of HPC applications
conference, January 2008

  • Verma, Akshat; Ahuja, Puneet; Neogi, Anindya
  • Proceedings of the 22nd annual international conference on Supercomputing - ICS '08
  • DOI: 10.1145/1375527.1375555

Power-Aware Consolidation of Scientific Workflows in Virtualized Environments
conference, November 2010

  • Zhu, Qian; Zhu, Jiedan; Agrawal, Gagan
  • 2010 SC - International Conference for High Performance Computing, Networking, Storage and Analysis, 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis
  • DOI: 10.1109/SC.2010.43

The impact of memory subsystem resource sharing on datacenter applications
conference, January 2011

  • Tang, Lingjia; Mars, Jason; Vachharajani, Neil
  • Proceeding of the 38th annual international symposium on Computer architecture - ISCA '11
  • DOI: 10.1145/2000064.2000099

Modeling performance variation due to cache sharing
conference, February 2013

  • Sandberg, A.; Sembrant, A.; Hagersten, E.
  • 2013 IEEE 19th International Symposium on High Performance Computer Architecture (HPCA)
  • DOI: 10.1109/HPCA.2013.6522315

Algorithmic Time, Energy, and Power on Candidate HPC Compute Building Blocks
conference, May 2014

  • Choi, Jee; Dukhan, Marat; Liu, Xing
  • 2014 IEEE International Parallel & Distributed Processing Symposium (IPDPS), 2014 IEEE 28th International Parallel and Distributed Processing Symposium
  • DOI: 10.1109/IPDPS.2014.54

The application slowdown model: quantifying and controlling the impact of inter-application interference at shared caches and main memory
conference, January 2015

  • Subramanian, Lavanya; Seshadri, Vivek; Ghosh, Arnab
  • Proceedings of the 48th International Symposium on Microarchitecture - MICRO-48
  • DOI: 10.1145/2830772.2830803

Resource-conscious scheduling for energy efficiency on multicore processors
conference, January 2010

  • Merkel, Andreas; Stoess, Jan; Bellosa, Frank
  • Proceedings of the 5th European conference on Computer systems - EuroSys '10
  • DOI: 10.1145/1755913.1755930

CPU Accounting for Multicore Processors
journal, February 2012

  • Luque, C.; Moreto, M.; Cazorla, F. J.
  • IEEE Transactions on Computers, Vol. 61, Issue 2
  • DOI: 10.1109/TC.2011.152

Bubble-Up: increasing utilization in modern warehouse scale computers via sensible co-locations
conference, January 2011

  • Mars, Jason; Tang, Lingjia; Hundt, Robert
  • Proceedings of the 44th Annual IEEE/ACM International Symposium on Microarchitecture - MICRO-44 '11
  • DOI: 10.1145/2155620.2155650

Dynamically Controlled Resource Allocation in SMT Processors
conference, January 2004

  • Cazorla, F. J.; Ramirez, A.; Valero, M.
  • 37th International Symposium on Microarchitecture (MICRO-37'04)
  • DOI: 10.1109/MICRO.2004.17

Exploiting unbalanced thread scheduling for energy and performance on a CMP of SMT processors
conference, January 2006

  • De Vuyst, M.; Kumar, R.; Tullsen, D. M.
  • Proceedings 20th IEEE International Parallel & Distributed Processing Symposium
  • DOI: 10.1109/IPDPS.2006.1639374

Addressing Fairness in SMT Multicores with a Progress-Aware Scheduler
conference, May 2015

  • Feliu, Josue; Sahuquillo, Julio; Petit, Salvador
  • 2015 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
  • DOI: 10.1109/IPDPS.2015.48

Deadline and energy constrained dynamic resource allocation in a heterogeneous computing environment
journal, February 2012

  • Young, B. Dalton; Apodaca, Jonathan; Briceño, Luis Diego
  • The Journal of Supercomputing, Vol. 63, Issue 2
  • DOI: 10.1007/s11227-012-0740-7

Power and Thermal-Aware Workload Allocation in Heterogeneous Data Centers
journal, February 2015

  • Al-Qawasmeh, Abdulla M.; Pasricha, Sudeep; Maciejewski, Anthony A.
  • IEEE Transactions on Computers, Vol. 64, Issue 2
  • DOI: 10.1109/TC.2013.116

Makespan and Energy Robust Stochastic Static Resource Allocation of a Bag-of-Tasks to a Heterogeneous Computing System
journal, October 2015

  • Oxley, Mark A.; Pasricha, Sudeep; Maciejewski, Anthony A.
  • IEEE Transactions on Parallel and Distributed Systems, Vol. 26, Issue 10
  • DOI: 10.1109/TPDS.2014.2362921

Supporting priorities and improving utilization of the IBM SP scheduler using slack-based backfilling
conference, January 1999

  • Talby, D.; Feitelson, D. G.
  • Proceedings 13th International Parallel Processing Symposium and 10th Symposium on Parallel and Distributed Processing. IPPS/SPDP 1999
  • DOI: 10.1109/IPPS.1999.760525

Scheduling parallel jobs on multicore clusters using CPU oversubscription
journal, March 2014

  • Utrera, Gladys; Corbalan, Julita; Labarta, Jesús
  • The Journal of Supercomputing, Vol. 68, Issue 3
  • DOI: 10.1007/s11227-014-1142-9

Utility Functions and Resource Management in an Oversubscribed Heterogeneous Computing Environment
journal, August 2015

  • Khemka, Bhavesh; Friese, Ryan; Briceno, Luis D.
  • IEEE Transactions on Computers, Vol. 64, Issue 8
  • DOI: 10.1109/TC.2014.2360513

A Methodology for Co-Location Aware Application Performance Modeling in Multicore Computing
conference, May 2015

  • Dauwe, Daniel; Jonardi, Eric; Friese, Ryan
  • 2015 IEEE International Parallel and Distributed Processing Symposium Workshop (IPDPSW)
  • DOI: 10.1109/IPDPSW.2015.38