Banca de DEFESA: ALEX FABIANO DE ARAÚJO FURTUNATO

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : ALEX FABIANO DE ARAÚJO FURTUNATO
DATE: 09/02/2021
TIME: 09:00
LOCAL: https://meet.google.com/nmm-yrvi-rtp
TITLE:

Analytical modeling for performance prediction of parallel applications in symmetric architectures considering variations in the data access delay


KEY WORDS:

Analytical modeling, Performance Modeling, Parallel Systems, Data Access Delay, Speedup, Memory Wall


PAGES: 118
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUMMARY:

Several analytical models created since Amdahl's pioneering work have explored aspects such as variation in the size of the problem, memory size, communication overhead, and synchronization overhead, but delays in accessing data are considered constant. However, such delays can vary, for example, according to the number of cores used, the relationship between processor and memory frequencies, and the problem's size. Given different problem sizes and the large number of possible configurations of operational frequency and number of cores than current architectures can offer, speedup models suitable for describing such variations among these configurations are quite desirable for offline or online scheduling decisions. A new analytical performance model that considers variations in the average data access delay to describe the limiting effect of the memory wall in parallel applications in homogeneous shared memory architectures is presented in this thesis. The experimental results indicate that the proposed modeling captures the application's behavior well. Besides, we show that considering parameters that reflect the applications' intrinsic characteristics, the proposal presented in this work has significant advantages over statistical models such as those based on machine learning. Our experiments also show that conventional machine learning modeling may need about an order of magnitude more measurements than the proposed model to achieve the same level of precision achieved by the proposed model.


BANKING MEMBERS:
Presidente - 1673543 - SAMUEL XAVIER DE SOUZA
Interno - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Interno - 1543191 - LUIZ FELIPE DE QUEIROZ SILVEIRA
Externo à Instituição - Arthur Francisco Lorenzon - UFP
Externo à Instituição - EDSON BORIN - UNICAMP
Notícia cadastrada em: 12/01/2021 19:56
SIGAA | Superintendência de Tecnologia da Informação - (84) 3342 2210 | Copyright © 2006-2024 - UFRN - sigaa02-producao.info.ufrn.br.sigaa02-producao