Parallel steganography benchmark for ValiPar concurrent program testing tool

  • Silvia M. D. Diaz
  • Paulo S. L. de Souza

Resumo


The validation and verification of concurrent programs is a major concern in software testing and parallel programming areas. Testing tools have been an essential support for V V tasks, and there are practices that allow analysing and validating their different characteristics and limitations. Here is where benchmarks come up with a solution with the goal of evaluating their behavior. ValiPar is a tool that conducts structural testing for both shared and distributed memory concurrent programs, and there is a set of benchmarks that were developed to verify several aspects. However, it is required the evaluation of the tool in regards of scalability, and the purpose of this work is to implement a real problem with a greater size algorithm to be tested in ValiPar, using both memory access models and a larger number of interactions among processes. This paper presents the results of a parallel steganography program development and the results of its evaluation, finding that the tool achieved to instrument and evaluate the parallel code but showed some limitations in the execution of test cases.
Publicado
10/04/2017
DIAZ, Silvia M. D.; SOUZA, Paulo S. L. de. Parallel steganography benchmark for ValiPar concurrent program testing tool. In: ESCOLA REGIONAL DE ALTO DESEMPENHO DE SÃO PAULO (ERAD-SP) , 2017, São Carlos. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2017 . p. 45 - 48.