Zocalo Tech, a startup concentrating on assertion based verification, has been active recently in terms of their articles, blogs and other activities in the media in an attempt to bring attention to the value of assertions in a verification methodology. In general they have had a refreshingly honest tone – especially in terms of the way in which they see this technology being adopted.

They have just released a new whitepaper titled Enabling Assertion Based Verification, (ABV) which is available for download from their website. Just like their previous materials, they start with the kind of statement that you would not expect to see from an EDA company “Assertion Based Verification is a difficult technology to implement and is perceived as marginally cost effective.” Of course, they then go on to try and dispel these perceptions. They also make an observation which they have made before, but I think is worth repeating “Most surveys on assertions reflect assertion use as opposed to ABV.”

So this indicates that the problems are not so much with the assertion languages – since they are being used in an ad-hoc manner, but that the methodologies do not make it easy to incorporate assertions in a way that they can become a core part of the verification strategy. As you read the paper, you will see that the previous assumption is not strictly correct. The assertion languages are easy enough to use if they are only used for simple assertions, but when used for more complex cases – the ones that would perhaps have a much higher value in the verification process, there seem to be few engineers out there with the requisite skills.

The paper concentrates on three main areas – coding, debug and control infrastructure and illustrates the technologies developed by Zocalo that can help to alleviate these problems. Perhaps one of the most exciting aspects is a visual way of putting assertions together. As they put it “Just as Visual Basic did for the BASIC programming language, VSVA raises the level of abstraction for describing properties and completely hides the complexities of using the SVA language.”

This paper is a good read and well worth the time. My only complaint is that the example they use does not in any way show the power of this visual language as it only uses a trivial assertion that has no temporal complexity – but then that may well be the meat of their next white paper. One can only hope so.

2 thumbs up on this one.

-------------------------------------------------
Brian Bailey – keeping you covered