Subject: Re: Assertion Committee Goals Revisited.
From: Steve Grout (grouts@earthlink.net)
Date: Tue Mar 19 2002 - 19:21:36 PST
Vassilios -
Responding to the several areas you requested feedback on...
...
> c- Usage Model: Currently, OVL is a model-based library. A more powerful and flexible capability will be required to allow an improved way for building and using such high level constructs. The plan is for SystemVerilog 3.1 to provide such a mechanism (through a task or a function) that is non-blocking.
In addition to the above task/function 'call-to-assertion' approach,
I would like also propose constructing a macro capability for each
assertion that allows a direct insertion, with such macros each to
be done in a free-standing way so as to not require the use of any
particular environment.
- GnuEmacs definitely could be instrumented to produce complete
correct specific assertions for each design instance/occurrence
of the assertion. This implementation take the form of a
macro-with-args as well as 'verilog-language-assertion-smart-completion'
approach with in the several verilog LSE editing modes that
are available.
- Ditto with the venerable UNIX M4 macro processor (though I
personally prefer an Emacs solution.)
- Ditto with VBA in a windows environment (though, you might no
it, I would personally first construct a working macro template,
and then port it (via an emacs macro translator) to windows VBA.
>
> 3- Action Item: It is important that every assertion member provide a position statement on OVL. The position statement should indicate the following:
> a- Will you support OVL (support is not just mere word, but clear involvement to improve or correct the current OVL).
I, of course now being retired and not working for a company, have no
company position to assert. I personally however strongly believe the
OVL is both a very important aspect of fully modeling the complete
semantics of a design and its requirements, as well as a very well
constructed work-in-progress of very useful and proven
assertion constructs. I have been reviewing each edition as they
have come out.
> b- If OVL is not adequate, what modification or additional capability required to help change your position.
The current OVL is, in my technical view, quite adequate for all the
specific applications and semantics it covers. As I noted some time
ago, there are other application areas that the OVL should be expanded
to include, but with no delay in its releases. These roughly can be
described as including more assertions / constraints in the area of
system level, requirements, specifications, environment, initialization,
power moding, and, as we begin to understand the general need, for specific
design applications, such as SERDES, RF, etc.
- I fully agree, though, with the current scope of coverage as being
that of the main set needed for many, many generally-digital design
applications, and very much applaud the work done todate.
> Best Regards
>
> Vassilios
--
--Steve Grout
Design Verification, CAD Methodology/R&D, Manager, Individual Contributor -
CAD System Infrastructure, Database/API, Flows, Tools, Integration,
and Support for both Analog/Mixed-Signal and Digital Design Teams.
11306 Musgrove Mill Drive, Spring Hill, FL 34609
Phone: 352-683-3298 or Cell: 352-428-8661,
email: grouts@earthlink.net
http://www.earthlink.net/~jsgrout/Personal/resume20225.pdf (or doc,rtf,txt)
This archive was generated by hypermail 2b28 : Tue Mar 19 2002 - 19:23:20 PST