Energy Prediction of OpenMP Applications Using Random Forest Modeling Approach

Benedict, Shajulin; Rejitha, R. S.; Gschwandtner, Philipp; Prodan, Radu; Fahringer, Thomas

doi:10.1109/ipdpsw.2015.12

Cited by 21 publications

(9 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These decision trees are aggregated into a random forest ensemble that combines their input. Then, results are aggregated, so it can outperform any individual decision tree's output [13].…”

Section: B Performance Modelsmentioning

confidence: 99%

Improving Collective I/O Performance with Machine Learning Supported Auto-tuning

Bagbaba

2020

2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

View full text Add to dashboard Cite

Collective Input and output (I/O) is an essential approach in high performance computing (HPC) applications. The achievement of effective collective I/O is a nontrivial job due to the complex interdependencies between the layers of I/O stack. These layers provide the best possible I/O performance through a number of tunable parameters. Sadly, the correct combination of parameters depends on diverse applications and HPC platforms. When a configuration space gets larger, it becomes difficult for humans to monitor the interactions between the configuration options. Engineers has no time or experience for exploring good configuration parameters for each problem because of long benchmarking phase. In most cases, the default settings are implemented, often leading to poor I/O efficiency. I/O profiling tools can not tell the optimal default setups without too much effort to analyzing the tracing results. In this case, an auto-tuning solution for optimizing collective I/O requests and providing system administrators or engineers the statistic information is strongly required. In this paper, a study of the machine learning supported collective I/O auto-tuning including the architecture and software stack is performed. Random forest regression model is used to develop a performance predictor model that can capture parallel I/O behavior as a function of application and file system characteristics. The modeling approach can provide insights into the metrics that impact I/O performance significantly.

show abstract

“…These decision trees are aggregated into a random forest ensemble that combines their input. Then, results are aggregated, so it can outperform any individual decision tree's output [13].…”

Section: B Performance Modelsmentioning

confidence: 99%

Improving Collective I/O Performance with Machine Learning Supported Auto-tuning

Bagbaba

2020

2020 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

View full text Add to dashboard Cite

show abstract

“…As an example, authors in [13], [14] assign an average current/power cost to every opcode in the target ISA and applies it to the execution trace of the program to estimate the energy consumption. The authors of [7] combine performance counters values (monitored during the application run) and a random forests model for predicting the energy consumed by parallel OpenMP applications.…”

Section: Energy Estimation From Source Codementioning

confidence: 99%

“…However, these techniques have never focused on energy, nor they targeted ultra-low-power embedded architectures. From the other side, previous research work investigated power and energy modelling of parallel architectures using features extracted from code execution profiling [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Source Code Classification for Energy Efficiency in Parallel Ultra Low-Power Microcontrollers

Parisi

Barchi

Bartolini

et al. 2021

2021 Design, Automation &Amp; Test in Europe Conference &Amp; Exhibition (DATE)

View full text Add to dashboard Cite

The analysis of source code through machine learning techniques is an increasingly explored research topic aiming at increasing smartness in the software toolchain to exploit modern architectures in the best possible way. In the case of lowpower, parallel embedded architectures, this means finding the configuration, for instance in terms of the number of cores, leading to minimum energy consumption. Depending on the kernel to be executed, the energy optimal scaling configuration is not trivial. While recent work has focused on general-purpose systems to learn and predict the best execution target in terms of the execution time of a snippet of code or kernel (e.g. offload OpenCL kernel on multicore CPU or GPU), in this work we focus on static compile-time features to assess if they can be successfully used to predict the minimum energy configuration on PULP, an ultra-low-power architecture featuring an on-chip cluster of RISC-V processors. Experiments show that using machine learning models on the source code to select the best energy scaling configuration automatically is viable and has the potential to be used in the context of automatic system configuration for energy minimisation.

show abstract

“…An energy prediction mechanism for OpenMP applications using a Random Forest Modeling (RFM) approach in compilers is proposed in (BENEDICT et al, 2015).…”

Section: Approaches With No Runtime Adaptation and No Transparency Tomentioning

confidence: 99%

Aurora: Seamless Optimization of OpenMP Applications

Lorenzon

Oliveira

Souza

et al. 2019

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

Primeiramente, gostaria de agradecer a Deus pela saúde, coragem e persistência para alcançar os objetivos. Ao Prof. "Caco", que, mostrou ser mais que um orientador de mestrado e doutorado. Se tornou um amigo e orientador de decisões importantes da minha vida. Sou e serei muito grato por todos os teus ensinamentos, puxões de orelha, discussões e momentos em que pude aprender contigo. Seus ensinamentos serão passados adiante. Obrigado por apostar em um guri lá de "cacimbinhas". À Prof. Márcia Cristina Cera, que, durante sua passagem contribuiu e foi muito importante para minha formação pessoal e profisisonal. Obrigado por ter me ensinado o caminho e ter me dito "vai" quando hesitei. Seu legado será passado adiante. Ao Prof. Stephan Wong, pelos ensinamentos durante os dois períodos em que fui aluno de doutorado visitante na TU-Delft. Estendo os cumprimentos aos demais amigos que lá fiz e companheiros de futebol. Ao corpo docente e quadro de funcionários do Instituto de Informática da UFRGS, em especial aos professores Luigi Carro, Flávio Rech, Gabriel Nazar e Paolo Rech. Ás agências financiadoras CAPES e CNPq, pelo apoio financeiro. Ao Prof. e amigo Fábio Rossi, que foi o responsável por instigar a vontade e desejo de um jovem de cidade pequena se tornar Dr. em Computação. Talvez tu não compreenda a grandeza de seu gesto e a gratidão que eu tenho por tu ter feito parte deste processo. Aos colegas do Laboratório de Sistemas Embarcados que essa longa caminhada me possibilitou encontrar. Em especial, os que foram importantes para minha formação:

show abstract

Energy Prediction of OpenMP Applications Using Random Forest Modeling Approach

Cited by 21 publications

References 20 publications

Improving Collective I/O Performance with Machine Learning Supported Auto-tuning

Improving Collective I/O Performance with Machine Learning Supported Auto-tuning

Source Code Classification for Energy Efficiency in Parallel Ultra Low-Power Microcontrollers

Aurora: Seamless Optimization of OpenMP Applications

Contact Info

Product

Resources

About