Coverage Fulfillment Automation in Hardware Functional Verification Using Genetic Algorithms

Abstract

The functional verification process is one of the most expensive steps in integrated circuit manufacturing. Functional coverage is the most important metric in the entire verification process. By running multiple simulations, different situations of DUT functionality can be encountered, and in this way, functional coverage fulfillment can be improved. However, in many cases it is difficult to reach specific functional situations because it is not easy to correlate the required input stimuli with the expected behavior of the digital design. Therefore, both industry and academia seek solutions to automate the generation of stimuli to reach all the functionalities of interest with less human effort and in less time. In this paper, several approaches inspired by genetic algorithms were developed and tested using three different designs. In all situations, the percentage of stimulus sets generated using well-performing genetic algorithms approaches was higher than the values that resulted when random simulations were employed. In addition, in most cases the genetic algorithm approach reached a higher coverage value per test compared to the random simulation outcome. The results confirmed that in many cases genetic algorithms can outperform constrained random generation of stimuli, that is employed in the classical way of doing verification, considering coverage fulfillment level per verification test.

Cost-Efficient Approaches for Fulfillment of Functional Coverage during Verification of Digital Designs

“…The SPEA2 algorithm is found to perform similarly to CPA, considering the first obtained high-performing stimulus sequence and the first-generation index containing 10 high-performing sequences. However, the number of performing data sequences generated by the SPEA2-based approach is larger than the number of sequences generated by the approaches developed during the present work, which are based on the implementation of the simple genetic algorithm in [ 20 ]. The main observation is that the CPA implementation is simpler and easier to understand than the NSGA-II and SPEA2 algorithms.…”

“…In order to develop sequences of input stimuli that can achieve high coverage, general principles of genetic algorithms are applied. Based on the Simple Genetic Algorithm (SGA) described in [ 20 ] (which was also addressed in [ 21 ]), the steps shown in Figure 6 were implemented, and the initial implementation was subsequently adapted to binary number processing and was improved to be more suitable for solving the current problem. In the first instance, several random data sets (having the structure required to be driven to the DUT inputs) are generated.…”

Reinforcement Learning Made Affordable for Hardware Verification Engineers

“…Its inputs and outputs are described in Table 2. This DUT has also been used to test design and verification automation mechanisms in [24,25]. The problem proposed in this paper is to make DUT reach a certain state (which is equivalent to generating a certain value on the "result" signal) by performing a minimum number of operations, whatever its initial state is.…”