Publication detail

Program Generation Through a Probabilistic Constrained Grammar

ČEKAN, O. PODIVÍNSKÝ, J. KOTÁSEK, Z.

Original Title

Program Generation Through a Probabilistic Constrained Grammar

Type

conference paper

Language

English

Original Abstract

The paper introduces a probabilistic constrained grammar which is a newly formed grammar system for use in the area of test stimuli generation. The grammar extends the existing probabilistic context-free grammar and establishes constraints for grammar limitations. Stimuli obtained through the proposed principle are used in the functional verification of a RISC processor and coverage metrics are evaluated. The detailed information about the construction of an assembly code for processors is described, as well as the experimantal results with the implemented generator. Experiments show the expressive power of the probabilistic constrained grammar and achieved code coverage in the verification of the processor. The grammar system demonstrates that is very suitable for an assembly code generation and universal use in the area of test stimuli.

Keywords

Probabilistic Constrained Grammar, Probabilistic Context-Free Grammar, Stimulus, Constraint, Functional Verification

Authors

ČEKAN, O.; PODIVÍNSKÝ, J.; KOTÁSEK, Z.

Released

29. 8. 2018

Publisher

IEEE Computer Society

Location

Praha

ISBN

978-1-5386-7376-8

Book

Proceedings - 21st Euromicro Conference on Digital System Design, DSD 2018

Pages from

214

Pages to

220

Pages count

7

URL

https://www.fit.vut.cz/research/publication/11709/

BibTex

@inproceedings{BUT155034,
  author="Ondřej {Čekan} and Jakub {Podivínský} and Zdeněk {Kotásek}",
  title="Program Generation Through a Probabilistic Constrained Grammar",
  booktitle="Proceedings - 21st Euromicro Conference on Digital System Design, DSD 2018",
  year="2018",
  pages="214--220",
  publisher="IEEE Computer Society",
  address="Praha",
  doi="10.1109/DSD.2018.00049",
  isbn="978-1-5386-7376-8",
  url="https://www.fit.vut.cz/research/publication/11709/"
}

Documents

program_generation.pdf