20× Better CDC Results & 95% Less Review Burden – Microarchitecture Insights

Srinivas: Hi Prakash Narayan, nice to meet you and great to listen to your talk. It was wonderful. Thank you for speaking with us. For the next few minutes, I’d like to ask you some questions about your product and its unique features. I want to focus especially on microarchitecture analysis, because that seems distinctive.

Prakash Narayan: Digital design ultimately comes down to bits — zeros and ones. Most traditional tools operate at the bit level or Boolean level (0, 1, and X). Designers, however, don’t think that way. They create higher-level components and assemble them to build functionality.

The belief behind microarchitecture analysis is that this structure already exists in the design. If we can extract those components, we can apply far more efficient reasoning and analysis techniques. Microarchitecture analysis extracts the building blocks of the design and then reasons on that higher-level view. As a result, it can be many times — often orders of magnitude — more efficient than pure bit-level or Boolean techniques.

Srinivas: Excellent — that’s exactly what I wanted to hear. You cited 20× more precise CDC violation reports and a 95% reduction in designer review burden. What was the baseline methodology you compared against, and how repeatable is that number across designs?

Prakash Narayan: This was in the context of CDC static sign-off. The traditional state of the art produces noisy results that users must review exhaustively before signing off. Tools mainly try to make that review easier by providing more context. The problem is that the volume of information has become so large that people make mistakes. This is a fundamental limitation of current CDC methodology, and it exists because those tools operate at the bit level.