Generation of Executable Runtime Constrained Random Programs Functional Processor Verification

Bauer W, Kohl J, Bäsig J, Rübesam S, Fey D (2016)


Publication Type: Conference contribution

Publication year: 2016

Publisher: EUROSIS

Edited Volumes: European Simulation and Modelling Conference 2016, ESM 2016

Book Volume: 30

Pages Range: 256-263

Conference Proceedings Title: ESM 2016

Event location: Gran Canaria, Spain

ISBN: 9789077381953

Abstract

The functional verification of multi-issue deep pipeline processor architectures exceeds the complexity level manageable with static test methods. A randomized approach including automatic test program generation leads to instruction sequences not anticipated and reaches the verification requirements. The challenge is to generate programs that on the one hand are deterministic and terminate for all possible inputs and on the other hand cover the entire instruction set. Moreover for out of order architectures the dependencies between instructions must be taken into account and therefore the instruction sequencing must be part of the coverage. In this paper we present a random program generation technique that produces executable runtime self-constraining deterministic random programs for processor verification. The core elements are a template based program flow generation, program termination certainty and runtime error protection. To avoid unpredictable or illegal situations during program execution value constraining blocks are added to program areas with uncertain behavior.

Authors with CRIS profile

How to cite

APA:

Bauer, W., Kohl, J., Bäsig, J., Rübesam, S., & Fey, D. (2016). Generation of Executable Runtime Constrained Random Programs Functional Processor Verification. In ESM 2016 (pp. 256-263). Gran Canaria, Spain: EUROSIS.

MLA:

Bauer, Wolfgang, et al. "Generation of Executable Runtime Constrained Random Programs Functional Processor Verification." Proceedings of the The European Simulation and Modelling Conferences, Gran Canaria, Spain EUROSIS, 2016. 256-263.

BibTeX: Download