Banca de QUALIFICAÇÃO: ELISA GABRIELA MACHADO DE LUCENA

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : ELISA GABRIELA MACHADO DE LUCENA
DATE: 20/02/2026
TIME: 14:00
LOCAL: Sala de reuniões DCA
TITLE:

Study of Machine Learning Models for Accelerating Scalability Analysis in the PaScal Suite

KEY WORDS:

Scalability analysis, High-performance computing, Machine learning

PAGES: 38
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Matemática da Computação
SUMMARY:

Performance and scalability analysis of parallel applications is a fundamental step in high-performance computing software development; however, it requires multiple executions, resulting in high computational, energy, and time costs. In this context, this work proposes two machine learning–based approaches to support scalability analysis by integrating predictive models into the PaScal Analyzer. The proposed approach relies on regression models trained adaptively from real measurements collected during the execution of parallel applications. The workflow is based on cross-validation, enabling the assessment of measurement quality and the automatic selection of model hyperparameters before final training. As a result, execution times for unmeasured configurations can be estimated, reducing the waiting time required to analyze application scalability.
Initial experiments were conducted in a high-performance computing environment, validating the integration of the experimental infrastructure. The results obtained at this stage indicate the feasibility of the proposed approach and provide a foundation for more in-depth comparative analyses between the adopted models, as well as for the expansion of experimental scenarios.

COMMITTEE MEMBERS:
Interno - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Interno - 1153006 - LUIZ AFFONSO HENDERSON GUEDES DE OLIVEIRA
Interno - 1673543 - SAMUEL XAVIER DE SOUZA