Support for Coverage Directed Test Generation

[Chrisdb1]

Hi all,

This is a question rather than an issue - please direct me if I'm asking it in the wrong place!

I am considering using CVA6 in research focused on coverage-directed test methods to efficiently achieve coverage closure, especially functional coverage.

My question is whether the current CVA6 verification environment is capable of generating random instructions and reporting the coverage points hit per test case (or instruction sequence). If so, how extensive is the support for this style of verification?

In 2018, a similar question was asked regarding the collection of functional and code coverage (issue #6). At that time, functional coverage-based verification was still in its early stages. However, the dashboards suggest that there have been significant advancements since then. Additionally, in the core-v-verif project, the CVA6 was reported as being in early stages of development in 2022 (CORE-V-VERIF Quick Start Guide). But recent issues suggest riscv-dv is being used successfully for random generation (Issue #1400).

Thanks, Chris