Customer Reviews

Functional Verification Coverage Measurement and Analysis (Information Technology: Transmission, Processing and Storage)

5 star:		(2)
4 star:		(1)
3 star:		(1)
2 star:		(1)
1 star:		(0)

 

 

This book covers features of related tools extensively without talking about specific tools.

Before buying this book, it is a good idea to read the discussion between the author and others at http://verificationguild.com/modules.php?name=Forums&file=viewtopic&p=185#185 if you are interested in using coverage measurement. There, none of the experts were able to address the following question:

Coverage tools are often used to set goals for simulation efforts. The flexible functional coverage allows such goals to be set too high or too low. Has any user made such mistakes? What can people do to get back on track quickly when the goals are too ambitious or too conservative?

This book also suggests using functional coverage for setting goals for simulation, but it also avoids the above question. The recent answers from some other authors are available:

Authors of http://www.research.ibm.com/journal/rd/494/victor.html set the coverage goals conservatively and assure each coverage point got hit many times.

A.Adir, H.Azatchi, E.Bin, O.Peled, and K.Shoikhet of "A generic micro-architectural test plan approach for microprocessor verification" in Proceedings of the 42nd Design Automation Conference (paper 47.1 from www.dac.com) suggested setting functional coverage goals according to project budgets.

Peet James' book (ISBN: 1402076193) seems to recommend using functional coverage but not to use it for setting goals.

Andreas Meyer's book (ISBN: 0750676175) provides an option to avoid using coverage because it can be too much work to explain the coverage data.

After reading some of the above, it can be easier to decide how useful this book is.

First, this book is short for the price, it only weighs in at 135 pages, excluding a unnecessary BNF description of the "e" language, and the more useful definitions. Furthermore, the book wasn't typeset well, it has widow and orphan lines all over - while this probably won't bother most, I find it jarring to my reading.

I'm somewhat familiar with coverage analysis, but no expert. I think the book serves as a OK introduction to the subject, but I found it wanting in concrete suggestions. When should I choose one technique over another? What coverage methods uncover the most bugs? Can I expect to get 100% line coverage? How do I make expression coverage more valuable and reduce false negatives? I'd like statistics, more what-to-do when, and more coverage of recent techniques based on PSL/OVL assertions - currently limited to two pages.

To summarize, it lacks the detail I'd want to make it worth the price. If you desire an introduction to coverage, I think you are better off with a recent book on general verification tecniques.

In the field of functional verification, coverage has been traditionally seen as a luxury activity which is done if time permits (and we all know what that means !). In recent years, along with the exponential growth of design complexity, this situation has drastically changed. Coverage is now typically a mandatory part of the verification process. However, in most cases, it is still effectuated in an ad-hoc manner, since there is no agreed upon methodology in this field yet. Andy's book comes in time to fill this gap.

The book's content spans from the basics of functional verification and functional coverage up to state of the art topics such as coverage-driven verification and usage of Hybrid models to improve coverage fidelity. In between, Andy reviews the different types of coverage, namely functional, code, and assertion coverage.

There is no assumption of prior knowledge, and the concepts are very clearly articulated. Examples are mostly shown using the e language (Andy works for the late Verisity), but it is done in a generic manner which allows to implement the core ideas in any environment.

I must admit only reading the first few chapters, as all I am interested in is functional coverage. I will some day read the rest of the book (I hope).

Commenting on the parts that I did read so far - I found this book an excellent learning tool. I had to learn how to write functional coverage in e, and this book was very valuable in getting me started, both with basic concepts, as well as examples. The example involving the wood stove is excellent. Unlike the documentation available from (Cadence), which is highly fragmented, Andy Piziali did an excellent job providing examples of complete - start - to -finish solutions.

In his forward to Piziali's book, Harry Foster, of Jasper Design Automation, says that this is "the first book to introduce a useful taxonomy for coverage metric classification." I agree that "useful taxonomy" describes this book well. I think it would be hard to find a more clear and complete introduction to functional verification coverage concepts and methods.

As an added bonus, this book is a good read. Piziali has a passion for verification coverage and a passion for language. The result is prose that not only is clean and easy to follow, but that also infects you with the excitement Piziali feels for his subject.