Machine Learning-enabled Scalable Performance Prediction of Scientific Codes

Chennupati, Gopinath; Santhi, Nandakishore; Romero, Phillip R.; Eidenbenz, Stephan Johannes

doi:10.1145/3450264

Title: Machine Learning-enabled Scalable Performance Prediction of Scientific Codes

Journal Article · Fri Apr 23 00:00:00 EDT 2021 · ACM Transactions on Modeling and Computer Simulation

DOI:https://doi.org/10.1145/3450264· OSTI ID:1841921

^[1];

^[1]

Los Alamos National Lab. (LANL), Los Alamos, NM (United States)

Hardware architectures become increasingly complex as the compute capabilities grow to exascale. Here, we present the Analytical Memory Model with Pipelines (AMMP) of the Performance Prediction Toolkit (PPT). PPT-AMMP takes high-level source code and hardware architecture parameters as input and predicts runtime of that code on the target hardware platform, which is defined in the input parameters. PPT-AMMP transforms the code to an (architecture-independent) intermediate representation, then (i) analyzes the basic block structure of the code, (ii) processes architecture-independent virtual memory access patterns that it uses to build memory reuse distance distribution models for each basic block, and (iii) runs detailed basic-block level simulations to determine hardware pipeline usage. PPT-AMMP uses machine learning and regression techniques to build the prediction models based on small instances of the input code, then integrates into a higher-order discrete-event simulation model of PPT running on Simian PDES engine. We validate PPT-AMMP on four standard computational physics benchmarks and present a use case of hardware parameter sensitivity analysis to identify bottleneck hardware resources on different code inputs. We further extend PPT-AMMP to predict the performance of a scientific application code, namely, the radiation transport mini-app SNAP. To this end, we analyze multi-variate regression models that accurately predict the reuse profiles and the basic block counts. We validate predicted SNAP runtimes against actual measured times.

View Accepted Manuscript (DOE)

Cite

Export

Save

Research Organization:: Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)

Sponsoring Organization:: USDOE National Nuclear Security Administration (NNSA)

Grant/Contract Number:: 89233218CNA000001

OSTI ID:: 1841921

Alternate ID(s):: OSTI ID: 1896442

Report Number(s):: LA-UR-20-23203; LA-UR-21-21328

Journal Information:: ACM Transactions on Modeling and Computer Simulation, Vol. 31, Issue 2; ISSN 1049-3301

Publisher:: Association for Computing MachineryCopyright Statement

Country of Publication:: United States

Language:: English

References (37)

Kokkos: Enabling manycore performance portability through polymorphic memory access patterns Carter Edwards, H.; Trott, Christian R.; Sunderland, Daniel Journal of Parallel and Distributed Computing, Vol. 74, Issue 12 https://doi.org/10.1016/j.jpdc.2014.07.003	journal	December 2014
Estimating cache misses and locality using stack distances CaΒcaval, Calin; Padua, David A. Proceedings of the 17th annual international conference on Supercomputing - ICS '03 https://doi.org/10.1145/782814.782836	conference	January 2003
Roofline: An Insightful Visual Performance Model for Floating-Point Programs and Multicore Architectures Williams, Samuel; Waterman, Andrew; Patterson, David https://doi.org/10.2172/1407078	report	September 2009
Accurate and efficient regression modeling for microarchitectural performance and power prediction Lee, Benjamin C.; Brooks, David M. ACM SIGARCH Computer Architecture News, Vol. 34, Issue 5 https://doi.org/10.1145/1168919.1168881	journal	October 2006
Predict the performance of GE with an ACO based machine learning algorithm Chennupati, Gopinath; Azad, R. Muhammad Atif; Ryan, Conor GECCO '14: Genetic and Evolutionary Computation Conference, Proceedings of the Companion Publication of the 2014 Annual Conference on Genetic and Evolutionary Computation https://doi.org/10.1145/2598394.2609860	conference	July 2014
Legion: Expressing locality and independence with logical regions Bauer, Michael; Treichler, Sean; Slaughter, Elliott 2012 SC - International Conference for High Performance Computing, Networking, Storage and Analysis, 2012 International Conference for High Performance Computing, Networking, Storage and Analysis https://doi.org/10.1109/SC.2012.71	conference	November 2012
COMPASS: A Framework for Automated Performance Modeling and Prediction Lee, Seyong; Meredith, Jeremy S.; Vetter, Jeffrey S. Proceedings of the 29th ACM on International Conference on Supercomputing - ICS '15 https://doi.org/10.1145/2751205.2751220	conference	January 2015
Hardware-independent application characterization Pakin, Scott; McCormick, Patrick 2013 IEEE International Symposium on Workload Characterization (IISWC) https://doi.org/10.1109/IISWC.2013.6704676	conference	September 2013
An Integrated Interconnection Network Model for Large-Scale Performance Prediction Ahmed, Kishwar; Obaida, Mohammad; Liu, Jason SIGSIM-PADS '16: SIGSIM Principles of Advanced Discrete Simulation, Proceedings of the 2016 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation https://doi.org/10.1145/2901378.2901396	conference	May 2016
Predicting whole-program locality through reuse distance analysis Ding, Chen; Zhong, Yutao ACM SIGPLAN Notices, Vol. 38, Issue 5 https://doi.org/10.1145/780822.781159	journal	May 2003
PPT-SASMM: Scalable Analytical Shared Memory Model: Predicting the Performance of Multicore Caches from a Single-Threaded Execution Trace Barai, Atanu; Chennupati, Gopinath; Santhi, Nandakishore MEMSYS 2020: The International Symposium on Memory Systems https://doi.org/10.1145/3422575.3422806	conference	March 2021
The gem5 simulator Binkert, Nathan; Sardashti, Somayeh; Sen, Rathijit ACM SIGARCH Computer Architecture News, Vol. 39, Issue 2 https://doi.org/10.1145/2024716.2024718	journal	August 2011
The Simian concept: Parallel Discrete Event Simulation with interpreted languages and just-in-time compilation Santhi, Nandakishore; Eidenbenz, Stephan; Liu, Jason 2015 Winter Simulation Conference (WSC) https://doi.org/10.1109/WSC.2015.7408405	conference	December 2015
New Performance Modeling Methods for Parallel Data Processing Applications Bhimani, Janki; Mi, Ningfang; Leeser, Miriam ACM Transactions on Modeling and Computer Simulation, Vol. 29, Issue 3 https://doi.org/10.1145/3309684	journal	July 2019
LogP: towards a realistic model of parallel computation Culler, David; Karp, Richard; Patterson, David ACM SIGPLAN Notices, Vol. 28, Issue 7 https://doi.org/10.1145/173284.155333	journal	July 1993
Miss rate prediction across all program inputs Zhong, Y.; Dropsho, S. G.; Ding, C. 12th International Conference on Parallel Architectures and Compilation Techniques. PACT 2003, Oceans 2002 Conference and Exhibition. Conference Proceedings (Cat. No.02CH37362) https://doi.org/10.1109/PACT.2003.1238004	conference	January 2003
LogGP: incorporating long messages into the LogP model---one step closer towards a realistic model for parallel computation Alexandrov, Albert; Ionescu, Mihai F.; Schauser, Klaus E. Proceedings of the seventh annual ACM symposium on Parallel algorithms and architectures - SPAA '95 https://doi.org/10.1145/215399.215427	conference	January 1995
Durango: Scalable Synthetic Workload Generation for Extreme-Scale Application Performance Modeling and Simulation Carothers, Christopher D.; Meredith, Jeremy S.; Blanco, Mark P. SIGSIM-PADS '17: SIGSIM Principles of Advanced Discrete Simulation, Proceedings of the 2017 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation https://doi.org/10.1145/3064911.3064923	conference	May 2017
Parallel Application Performance Prediction Using Analysis Based Models and HPC Simulations Obaida, Mohammad Abu; Liu, Jason; Chennupati, Gopinath SIGSIM-PADS '18: SIGSIM Principles of Advanced Discrete Simulation, Proceedings of the 2018 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation https://doi.org/10.1145/3200921.3200937	conference	May 2018
Cetus: A Source-to-Source Compiler Infrastructure for Multicores Dave, Chirag; Bae, Hansang; Min, Seung-Jai Computer, Vol. 42, Issue 12 https://doi.org/10.1109/MC.2009.385	journal	December 2009
PARDA: A Fast Parallel Reuse Distance Analysis Algorithm Niu, Qingpeng; Dinan, James; Lu, Qingda 2012 IEEE International Symposium on Parallel & Distributed Processing (IPDPS), 2012 IEEE 26th International Parallel and Distributed Processing Symposium https://doi.org/10.1109/IPDPS.2012.117	conference	May 2012
Order of Nonlinearity as a Complexity Measure for Models Generated by Symbolic Regression via Pareto Genetic Programming Vladislavleva, E. J.; Smits, G. F.; den Hertog, D. IEEE Transactions on Evolutionary Computation, Vol. 13, Issue 2 https://doi.org/10.1109/TEVC.2008.926486	journal	April 2009
Evolving multidimensional transformations for symbolic regression with M3GP Muñoz, Luis; Trujillo, Leonardo; Silva, Sara Memetic Computing, Vol. 11, Issue 2 https://doi.org/10.1007/s12293-018-0274-5	journal	September 2018
Rose: Compiler Support for Object-Oriented Frameworks Quinlan, Dan Parallel Processing Letters, Vol. 10, Issue 02n03 https://doi.org/10.1142/S0129626400000214	journal	June 2000
Fast, accurate, and scalable memory modeling of GPGPUs using reuse profiles Arafa, Yehia; Badawy, Abdel-Hameed; Chennupati, Gopinath ICS '20: 2020 International Conference on Supercomputing, Proceedings of the 34th ACM International Conference on Supercomputing https://doi.org/10.1145/3392717.3392761	conference	June 2020
The structural simulation toolkit Rodrigues, A. F.; CooperBalls, E.; Jacob, B. ACM SIGMETRICS Performance Evaluation Review, Vol. 38, Issue 4 https://doi.org/10.1145/1964218.1964225	journal	March 2011
Reuse-distance-based miss-rate prediction on a per instruction basis Fang, Changpeng; Carr, Steve; Önder, Soner Proceedings of the 2004 workshop on Memory system performance - MSP '04 https://doi.org/10.1145/1065895.1065906	conference	January 2004
Predicting whole-program locality through reuse distance analysis Ding, Chen; Zhong, Yutao Proceedings of the ACM SIGPLAN 2003 conference on Programming language design and implementation - PLDI '03 https://doi.org/10.1145/781131.781159	conference	January 2003
Evaluation techniques for storage hierarchies Mattson, R. L.; Gecsei, J.; Slutz, D. R. IBM Systems Journal, Vol. 9, Issue 2 https://doi.org/10.1147/sj.92.0078	journal	January 1970
Inferred Models for Dynamic and Sparse Hardware-Software Spaces Wu, Weidan; Lee, Benjamin C. 2012 45th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO) https://doi.org/10.1109/MICRO.2012.45	conference	December 2012
Accelerating multicore reuse distance analysis with sampling and parallelization Schuff, Derek L.; Kulkarni, Milind; Pai, Vijay S. Proceedings of the 19th international conference on Parallel architectures and compilation techniques - PACT '10 https://doi.org/10.1145/1854273.1854286	conference	January 2010
Imcsim: Parameterized Performance Prediction for Implicit Monte Carlo Codes Chennupathi, Gopinath; Eidenbenz, Stephan; Long, Alex 2018 Winter Simulation Conference (WSC) https://doi.org/10.1109/WSC.2018.8632406	conference	December 2018
Discrete event performance prediction of speculatively parallel temperature-accelerated dynamics Zamora, Richard J.; Voter, Arthur F.; Perez, Danny SIMULATION, Vol. 92, Issue 12 https://doi.org/10.1177/0037549716674806	journal	October 2016
Program locality analysis using reuse distance Zhong, Yutao; Shen, Xipeng; Ding, Chen ACM Transactions on Programming Languages and Systems, Vol. 31, Issue 6 https://doi.org/10.1145/1552309.1552310	journal	August 2009
Analytical Processor Performance and Power Modeling using Micro-Architecture Independent Characteristics Van den Steen, Sam; Eyerman, Stijn; De Pestel, Sander IEEE Transactions on Computers https://doi.org/10.1109/TC.2016.2547387	journal	January 2016
MARSS: a full system simulator for multicore x86 CPUs Patel, Avadh; Afram, Furat; Chen, Shunfei Proceedings of the 48th Design Automation Conference on - DAC '11 https://doi.org/10.1145/2024724.2024954	conference	January 2011
Palm: easing the burden of analytical performance modeling Tallent, Nathan R.; Hoisie, Adolfy Proceedings of the 28th ACM international conference on Supercomputing - ICS '14 https://doi.org/10.1145/2597652.2597683	conference	January 2014

Similar Records

$\mathrm{PPT}$-Multicore: performance prediction of Open$\mathrm{MP}$ applications using reuse profiles and analytical modeling

Journal Article · Mon Jun 28 00:00:00 EDT 2021 · Journal of Supercomputing · OSTI ID:1841921

Barai, Atanu; Arafa, Yehia; Badawy, Abdel-Hameed; +3 more

PPT-GPU: Scalable GPU Performance Modeling

Journal Article · Tue Jan 01 00:00:00 EST 2019 · IEEE Computer Architecture Letters · OSTI ID:1841921

Arafa, Yehia; Badawy, Abdel Hameed A.; Chennupati, Gopinath; +2 more

ASC Tri-lab Co-design Level 2 Milestone Report 2015

Technical Report · Wed Sep 23 00:00:00 EDT 2015 · OSTI ID:1841921

Hornung, Rich; Jones, Holger; Keasler, Jeff; +11 more

Related Subjects

97 MATHEMATICS AND COMPUTING
computer hardware
performance modeling
deep learning
symbolic regression
genetic programming
hardware software codesign
performance prediction
machine learning

Title: Machine Learning-enabled Scalable Performance Prediction of Scientific Codes

Citation Formats

References (37)

Similar Records

Related Subjects