Panel For Example Panel For Example Panel For Example

VCS Verification Flow with VSO.ai

Author : Adrian September 29, 2025

VCS Verification Flow with VSO.ai

Introduction

The recent surge in AI has driven greater attention from semiconductor vendors and EDA companies, which have introduced various AI-powered EDA tools to improve PPA in chip design. Vendors such as Cadence, Synopsys, and Mentor have released competitive AI-enabled products.

Synopsys added a perceptive AI feature to VCS in the 2023.03 release: VSO.ai (Verification Space Optimization). Although VSO was previously available as a separate product, its direct integration into VCS is significant. Existing designs generally require no code changes; enabling VSO often only requires adding a few command-line options.

About VSO.ai

Synopsys VSO.ai provides an AI-driven verification solution intended to help verification teams reach coverage convergence faster and with higher quality. The system operates autonomously to achieve coverage goals as quickly and cost-effectively as possible while aiming for high-quality results. Machine learning techniques are used to identify and remove redundant regressions, automatically perform root-cause analysis on coverage, and infer coverage from RTL and stimuli to identify coverage gaps and provide guidance for improving coverage.

Coverage Convergence Challenges

Design, constraints, and coverage goals can change multiple times during the development flow. Manual regression runs often struggle to improve coverage or may fail to make progress. Test cases may contain substantial redundant coverage, which complicates analysis and consumes engineering time. Relying solely on manual analysis to converge coverage is time-consuming and may not achieve the expected results.

Traditional coverage verification requires significant human effort and repetitive work, and it is difficult to make progress on the most challenging coverage points.

VSO.ai Approach

VSO.ai automates several key steps in the flow to improve efficiency. A coverage inference step runs during compilation, using automatically generated coverage to complement structural and user-specified coverage metrics.

In short: VSO.ai guides designers and verification engineers toward faster, higher-quality coverage convergence, allowing engineers to focus on fixing bugs rather than spending time locating them. Automating labor-intensive tasks accelerates coverage closure.

Outlook

Current AI-focused verification products of interest include Synopsys VSO.ai, DSO.ai, TSO.ai, and Synopsys.ai Copilot. These offerings can be deployed locally or in the cloud, depending on the vendor's support model. Cadence has also released multiple AI-enabled EDA tools, with notable progress in backend and board-level AI capabilities.

Since the introduction of ChatGPT, AI has been integrated increasingly quickly into various stages of IC design. Using AI to accelerate design and verification and to improve PPA is becoming a common trend. IC industry professionals should monitor AI developments relevant to their work.

Statement

The content and images were provided by contributing authors or authorized partner sites. The views expressed are those of the authors and do not represent the position of the publishing site. The article and images are for engineers' learning purposes only. If there are copyright or other issues, please contact the publishing site for resolution.