• Home
  • Search
  • Predicting HPC parallel program performance based on LLVM compiler
  • Cite Icon18
  • https://doi.org/10.1007/s10586-016-0707-1Copy DOI Icon

Predicting HPC parallel program performance based on LLVM compiler

Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Performance prediction of parallel program plays key roles in many areas, such as parallel system design, parallel program optimization, and parallel system procurement. Accurate and efficient performance prediction on large-scale parallel systems is a challenging problem. To solve this problem, we present an effective framework for performance prediction based on the LLVM compiler technique in this paper. We can predict the performance of a parallel program on a small amount of nodes of the target parallel system using this framework toned but not execute this parallel program on a corresponding full-scale parallel system. This framework predicts the performance of computation and communication components separately and combines the two predictions to achieve full program prediction. As for sequential computation, we first combine the static branch probability and loop trip count identification and propose a new instrumentation method to acquire the number of each instruction type. We then construct a test program to measure the average execution time of each instruction type. Finally, we utilize the pruning technique to convert a parallel program into a corresponding sequential program to predict the performance on only one node of the target parallel system. As for communication, we utilize the LogGP model to model point-to-point communication and the artificial neural network technique to model collective communication. We validate our approach by a set of experiments that predict the performance of NAS parallel benchmarks and CGPOP parallel application. Experimental results show that the proposed framework can accurately predict the execution time of parallel programs, and the average error rate of these programs is 10.86%.

Similar Papers
  • Conference Article
  • Citations25

A source code analyzer for performance prediction

  • Apr 26, 2004
  • M Kuhnemann +2
  • Conference Article
  • Citations6

Automated Performance Modeling Based on Runtime Feature Detection and Machine Learning

  • Dec 01, 2017
  • Jingwei Sun +3
  • Research Article
  • Citations54

Automated Performance Modeling of HPC Applications Using Machine Learning

  • May 01, 2020
  • IEEE Transactions on Computers
  • Jingwei Sun +4
  • Book Chapter
  • Citations14

Efficient solutions for mapping parallel programs

  • Jan 01, 1995
  • P Bouvry +2
  • Research Article
  • Citations47

Stochastic bounds on execution times of parallel programs

  • Jan 01, 1991
  • IEEE Transactions on Software Engineering
  • N Yazia-Pekergin +1
  • Research Article

Research of Parallel Program Performance Analysis and Optimization on Multicore Platform

  • Aug 16, 2013
  • Yingyan Teng +1
  • Research Article
  • Citations7

Predicting the execution times of parallel-independent programs using Pearson distributions

  • Aug 01, 2005
  • Parallel Computing
  • G.L Reijns +1
  • Research Article
  • Citations50

Compiler techniques for data partitioning of sequentially iterated parallel loops

  • Jun 01, 1990
  • ACM SIGARCH Computer Architecture News
  • David E Hudak +1
  • PDF
  • Research Article

Расширение модели ParJava для случая кластеров с многоядерными узлами

  • Jan 01, 2012
  • Proceedings of the Institute for System Programming of RAS
  • M.S Akopyan
  • Conference Article
  • Citations7

Toward comprehensive and accurate simulation performance prediction of parallel file systems

  • Jun 01, 2012
  • Miguel A Erazo +3
  • Research Article
  • Citations9

Timing parallel programs that use message passing

  • Jan 01, 1992
  • Journal of Parallel and Distributed Computing
  • Nicholas T Karonis
  • Conference Article
  • Citations3

Ant Colony Algorithm for Scheduling Parallel Program Based on DAG Graph Heuristics

  • Jan 01, 2006
  • Xiaohong Kong +1
  • Research Article
  • Citations49

Using parallel program characteristics in dynamic processor allocation policies

  • Oct 01, 1996
  • Performance Evaluation
  • T Brecht
  • Research Article
  • Citations23

Using parallel program characteristics in dynamic processor allocation policies

  • Oct 01, 1996
  • Performance Evaluation
  • Timothy B Brecht +1
  • Book Chapter

A New Approach to Programming and Prototyping Parallel Systems

  • Jan 01, 2005
  • Kunle Olukotun
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.