Learning from Limited Data in VLSI CAD

Wang, Li-C.

doi:10.1007/978-3-030-04666-8_13

Cited by 1 publication

(1 citation statement)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alternatively, the relationship might not be learnable because the function to be approximated is too complex. The lack of positive training examples [10] is also problematic in CDS because, by aiming to exercise hard-to-reach and yet-uncovered DUT functionality, we are trying to make predictions about events that have occurred rarely, if at all, in the past.…”

Section: Constraints Verification Engineermentioning

confidence: 99%

Hybrid Intelligent Testing in Simulation-Based Verification

Masamba,

Eder,

Blackmore

2022

Preprint

View full text Add to dashboard Cite

Efficient and effective testing for simulation-based hardware verification is challenging. Using constrained random test generation, several millions of tests may be required to achieve coverage goals. The vast majority of tests do not contribute to coverage progress, yet they consume verification resources. In this paper, we propose a hybrid intelligent testing approach combining two methods that have previously been treated separately, namely Coverage-Directed Test Selection and Novelty-Driven Verification. Coverage-Directed Test Selection learns from coverage feedback to bias testing towards the most effective tests. Novelty-Driven Verification learns to identify and simulate stimuli that differ from previous stimuli, thereby reducing the number of simulations and increasing testing efficiency. We discuss the strengths and limitations of each method, and we show how our approach addresses each method's limitations, leading to hardware testing that is both efficient and effective.

show abstract

Section: Constraints Verification Engineermentioning

confidence: 99%