Use este identificador para citar o ir al link de este elemento:
http://hdl.handle.net/1843/43133
Tipo: | Dissertação |
Título: | Autoparbench: a framework for parallel code verification |
Título(s) alternativo(s): | Autoparbench: uma ferramenta para verificação de código paralelo |
Autor(es): | Gleison Souza Diniz Mendonça |
primer Tutor: | Fernando Magno Quintão Pereira |
primer Co-tutor: | Liao Chunhua |
primer miembro del tribunal : | Dorgival Olavo Guedes Neto |
Segundo miembro del tribunal: | Pei-Hung Lin |
Resumen: | There exist presently many parallelization tools based on the automatic insertion of OpenMP pragmas into programs. However, it is challenging to automatically and quantitatively compare these tools for their strengths and limitations, due to the diverse choices to parallelize a program. This work describes AutoParBench, a test framework aimed to mitigate this problem. AutoParBench consists of benchmarks and a verifier. Benchmarks currently include 99 programs with 1,579 loops. A procedure is defined to allow quick and easy additions of new programs. The verifier consists of a common intermediate representation, based on JSON, plus all the machinery necessary to convert OpenMP programs into a format henceforth called a JSON snapshot. The snapshots produced by different tools enable automatic semantics-aware comparison of syntactically different parallelization results. AutoParBench is an effective bug-finding instrument. By investigating differences in snapshots produced by separate sources, i.e., tool versus tool or tool versus human, we have reported bugs in selected parallelization tools such as ICC, Cetus, AutoPar and DawnCC, all of which have been confirmed. |
Abstract: | There exist presently many parallelization tools based on the automatic insertion of OpenMP pragmas into programs. However, it is challenging to automatically and quantitatively compare these tools for their strengths and limitations, due to the diverse choices to parallelize a program. This work describes AutoParBench, a test framework aimed to mitigate this problem. AutoParBench consists of benchmarks and a verifier. Benchmarks currently include 99 programs with 1,579 loops. A procedure is defined to allow quick and easy additions of new programs. The verifier consists of a common intermediate representation, based on JSON, plus all the machinery necessary to convert OpenMP programs into a format henceforth called a JSON snapshot. The snapshots produced by different tools enable automatic semantics-aware comparison of syntactically different parallelization results. AutoParBench is an effective bug-finding instrument. By investigating differences in snapshots produced by separate sources, i.e., tool versus tool or tool versus human, we have reported bugs in selected parallelization tools such as ICC, Cetus, AutoPar and DawnCC, all of which have been confirmed. |
Asunto: | Computação – Teses Benchmarking – Teses Paralelização – Teses Software – Verificação – Teses |
Idioma: | eng |
País: | Brasil |
Editor: | Universidade Federal de Minas Gerais |
Sigla da Institución: | UFMG |
Departamento: | ICX - DEPARTAMENTO DE CIÊNCIA DA COMPUTAÇÃO |
Curso: | Programa de Pós-Graduação em Ciência da Computação |
Tipo de acceso: | Acesso Aberto |
URI: | http://hdl.handle.net/1843/43133 |
Fecha del documento: | 28-feb-2020 |
Aparece en las colecciones: | Dissertações de Mestrado |
archivos asociados a este elemento:
archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
gleison_dissertation_v2.pdf | AUTOPARBENCH: A FRAMEWORK FOR PARALLEL CODE VERIFICATION | 1.02 MB | Adobe PDF | Visualizar/Abrir |
Los elementos en el repositorio están protegidos por copyright, con todos los derechos reservados, salvo cuando es indicado lo contrario.