Obstacles to adopting Assertion-Based Verification

Newsletter · Original Contribution Joined: Dec 08, 2003 Posts: 1107

(Originally from Issue 4.11, Item 2.0)

From: Ben Cohen

First I want to congratulate Harry Foster, Adam Krolnik, and David
Lacey on an excellent job in the writing of the book "Assertion-Based
Design." This book addresses, among other things, the assertion
methodology, properties and assertion patterns for PSL, OVL, and
PLI-based assertions. I highly recommend it.

An interesting discussion in the book is why assertions are not used,
and misconceptions among engineers and managers about ABV. I would
like to add a few more reasons and observations:

1) "assertion what? ... never heard of it".

2) "ABV is a paradigm shift. We stick to what we normally do".
Incidentally, the paradigm shift from graphical design entry and
manual synthesis to RTL design and use of synthesis tools met with
a similar challenges.

3) "As management, I am not technical, and I am proud of it. I am not
responsible for any mistake made by engineering. However, I
allocate funds. Design is an engineering problem, not a
management issue. ABV may add costs, particularly if I have to
buy tools, and I have a bottom line". Of course, there are 'false
bottoms'.

4) "To really assess the value of an ABV methodology, we need to start
an IR&D effort. That means that we need a 30-page document
explaining the value of such a task. This will allow us to decide
if we need to allocate some money, which can be taken away at any
time, for 'next year funding' to do such research. A committee
will evaluate the results of this IR&D effort on the use of ABV,
and then we will ponder for a while on the recommendations. If
new tools are required, then the planning and money allocation
needs to be made a year in advance. In the meantime, business as
usual in the field of design definition and verification".

The above is laughable but, sadly enough, it's no joke. After getting
involved with ABV, and writing a book on "Using PSL", I made the
paradigm shift and see ABV as an invaluable methodology to enhance the
success of designs at reduced design (clearer definitions of
requirements and designs) and verification costs (white box, tests
against requirements, early error detection,...).

I was impressed at DAC the emphasis that many vendors have taken to
develop tools that uses ABV and support of PSL.

I would like to hear issues that this audience has with regards to
ABV.

- Ben Cohen