Analyzing Dynamic Task-Based Applications on Hybrid Platforms: An Agile Scripting Approach

Vinicius Garcia Pinto 1, 2 Luka Stanisic 3, 4 Arnaud Legrand 5, 1 Lucas Mello Schnorr 2 Samuel Thibault 3 Vincent Danjean 1, 5
3 STORM - STatic Optimizations, Runtime Methods
LaBRI - Laboratoire Bordelais de Recherche en Informatique, Inria Bordeaux - Sud-Ouest
4 HiePACS - High-End Parallel Algorithms for Challenging Numerical Simulations
LaBRI - Laboratoire Bordelais de Recherche en Informatique, Inria Bordeaux - Sud-Ouest
5 POLARIS - Performance analysis and optimization of LARge Infrastructures and Systems
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
Abstract : In this paper, we present visual analysis techniques to evaluate the performance of HPC task-based applications on hybrid architectures. Our approach is based on composing modern data analysis tools (pjdump, R, ggplot2, plotly), enabling an agile and flexible scripting framework with minor development cost. We validate our proposal by analyzing traces from the full-fledged implementation of the Cholesky decomposition available in the MORSE library running on a hybrid (CPU/GPU) platform. The analysis compares two different workloads and three different task schedulers from the StarPU runtime system. Our analysis based on composite views allows to identify allocation mistakes, priority problems in scheduling decisions, GPU tasks anomalies causing bad performance, and critical path issues.
Complete list of metadatas

Cited literature [24 references]  Display  Hide  Download

https://hal.inria.fr/hal-01353962
Contributor : Vinicius Garcia Pinto <>
Submitted on : Tuesday, July 17, 2018 - 3:59:52 PM
Last modification on : Friday, October 25, 2019 - 4:56:06 PM
Long-term archiving on : Thursday, October 18, 2018 - 3:46:19 PM

File

vpa2016-submitted.pdf
Files produced by the author(s)

Identifiers

Citation

Vinicius Garcia Pinto, Luka Stanisic, Arnaud Legrand, Lucas Mello Schnorr, Samuel Thibault, et al.. Analyzing Dynamic Task-Based Applications on Hybrid Platforms: An Agile Scripting Approach. VPA 2016 - 3rd Workshop on Visual Performance Analysis, Nov 2016, Salt Lake City, United States. pp.17-24, ⟨10.1109/VPA.2016.008⟩. ⟨hal-01353962v2⟩

Share

Metrics

Record views

611

Files downloads

445