Minutes from OVL Meeting 9 May 2002


Subject: Minutes from OVL Meeting 9 May 2002
From: David Lacey (dlacey@rsn.hp.com)
Date: Thu May 09 2002 - 11:03:08 PDT


Attending:
Jason Andrews - Axis
Henry Cox - Co-Design
Gail Dagan - Intel
Harry Foster - Verplex
Peter Flake - Co-Design
Adam Krolnick - LSI
Erich Marschner - Cadence
Joe Richards - SGI
Sean Smith - Cisco
Richard Stolzman - Verplex
Bassam Tabbara - Novus
Curt Takara - 0-in

Updates for the July Release of OVL
- Sean's suggestion for defining severity levels and allowing
  customization of reporting and control of the assertions.
  (see attached file)
  - We will align the severity levels with DAS
  - Adam will send out suggestion on how to also include
    ability to define a threshold of assertion firings
    that will end the simulation. Approval will be via
    email.
- Joe's and Adam's suggestion of standardizing implementation of
  window asserts (assert_window, assert_win_change,
  assert_win_unchange)
  - The current implementations are not consisten with each
    other. Some continue to check the test expression through
    the last cycle while some stop checking upon hitting the
    end event condition.
  - Will add a way to specify how checking is done on the last
    cycle. Three options for doing this:
    - add a new parameter to these modules (preferred method)
      This option may not work with existing code
    - make use of the "option" parameter that is already in
      OVL. This would not impact existing code.
    - Create a new set of window assertions
  - Sugar has a method of specifying how the final check is
    done. There should be consistentcy between OVL and Sugar
    here.
  - Harry will send out examples of the options and we will
    decide via email.
- Including DAS in the OVL. Some of this work is already done
  and we can show at DAC the direction we are going with OVL
  and DAS. But the final integration will not be done until the
  July release. Harry is driving this.
- assert_set: Harry will send out this new assertion and we
  will decide via email whether to include it in July's release.
- Chaining of assertions. Harry has come up with a method to
  allow the chaining of assertions. However, it was determined
  this this type of change was extensive enough that it should
  be part of the consideration for the 2.0 release of OVL in
  December.

We discussed the need to coordinated efforts between the formal
property language, DAS, and OVL. We should look at how all
three are working together in our current and future work.

OVL 2.0 Investigation Areas
- Next major release of OVL targetted for December 2002
- In this release, we should look at OVL from the ground up.
  This doesn't mean we have to change it all, but need to
  determine the best and consistent approach to OVL and its
  future.
  - We will start this discussion by looking at requirements
    for assertion methodology. Everyone is charged with sending
    ideas to the group on this subject over the next month.
- Need to include both concurrent and procedural assertions
  in this release. Again, being sure to align with DAS and
  Sugar.
- Formally define the semantics of the new library release.
  This will allow us to move forward with presenting this for
  offical standardization.
- Define deliverables
  - The library itself
  - users manual
  - How to ensure quality of the OVL. This could include
    defined formal checks or a test suite. The actual test
    suite may not be the deliverable, but our documentation
    of how we are ensuring that OVL does what it says it
    does.
- What needs to be done to enable migrating from OVL 1.0 to
  OVL 2.0?

The approach to define OVL 2.0 will be to start with
assertion methodology. Please generate your thoughts on
how you are using assertions, how you control them, etc.
and send them to the email reflector. These thoughts will
be merged and a combined list will be sent out before
our next meeting. We are targetting the next meeting for
after DAC on June 20. Details will be sent out later.

VHDL subcommittee is being formed to look at issues related
to the VHDL version of OVL. Contact Harry of David if you
are interested in paritcipated in this subcommittee.

There will be a Verification in the Univerity meeting the
Sunday befor DAC (June 9). This group is finding ways
to get industry supporting the creation of new classes
that will address topics that are currently important.
Sean will send out information on this meeting.

-- 
David J. Lacey
Email: dlacey@rsn.hp.com
Phone: 972-497-4114



This archive was generated by hypermail 2b28 : Thu May 09 2002 - 11:02:10 PDT