Date: Tue, 29 Mar 94 17:44:18 PST


X-MX-Warning: Warning -- Invalid "To" header.

Subject: 1364 Working Group Meeting Minutes

                   IEEE 1364 WORKING GROUP MEETING (VERILOG)
                      MINUTES OF MEETING HELD ON 03/14/94

Attendees: Maqsoodul Mannan, Mike MacNamara for John Sanguinetti, Andrew {*filter*}
           for Peter Eichenberger, Joseph Skudlarek, Alec Stanculescu, Alex
           Zanfirecsu for Gabe Moretti, Lynn Horobin, Yatin Trivedi,
           Leigh Brady, Brian Griffin, Andrew Guyler, Jen-jen Lung for  
           J. Bhasker, Joel Paston, Cliff Cummings, Phil Moorby.

1. Approval of the minutes from the February 11, 1993 meeting:  The following
   amendments were made to the minutes:

   a) Amendment to section #3, Item c : Sentence to read "Joseph Skudlarek
      submitted a list of issues for discussion at the end of the meeting.
      These items included identifier length and maximum number of ports."
      [Maq: Is this all he wanted listed ?]

      Motion: Approve minutes as amended
      Proposed:  Mike MacNamara
      Seconded:  Yatin Trivedi
      Motion was passed by unanimous vote.

2. Agenda Changes and Additions:

   a) Joseph Skudlarek suggested that a discussion about "mapping a game plan"
      for producing a specification by July be added to the agenda.

   b) Yatin Trivedi suggested that the group be flexible about when the PLI
      discussion takes place due to Rajeev Madhavan's afternoon tutorial at
      the IVC.

   c) Maq Mannan said that the DASC has requested that the IEEE 1364 Working
      group hold a meeting at DAC in San Diego. It has already been decided
      that this was not generally convenient for a lengthy working session due
      to other DAC committments and that some of the members will not be
      attending DAC. Maq proposed that a 1-2 hour "informative" meeting be
      held and that members attending DAC should be there.  

      Motion: Schedule a "non-working" informational meeting at DAC
      Proposed: Mike MacNamara
      Seconded: Yatin Trivedi
      Motion was passed by unanimous vote.

3. Electronic LRM copies: The OVI Board has agreed to release copies of the LRM
   on tape to working group members who request them.An agreement must be signed
   stating that the tape will be used for IEEE 1364 work only, not to be copied
   or distributed for any other purpose. The Board has also asked for $40.00 to
   cover the tape reproduction, handling and shipping costs. Members who need a
   tape are to contact Lynn Horobin.

4. John Mancini has agreed to help with the formatting of the 1364 specification
   according to IEEE requirements, and has contacted Mary Lynne Nielsen at the
   IEEE. He will submit a report on his findings.

   a) John Mancini also requested that in preparation for the July balloting,
      each member of the Working Group submit a list of people to him by e-mail
      to be added to the balloting group. Maq asked that each working group
      member take an action item to submit their list by the next meeting to
      John Mancini.  

5. SDF: It was decided at the first 1364 meeting that SDF would not be
   worked on now, but would be considered after the initial Verilog specifica-
   tion was complete. Leigh Brady thought that it would take 2 to 6 months
   to study and prepare the SDF. Lynn Horobin said that the OVI Library
   Modeling Committee has completed SDF version 2.1. It was also pointed out
   that OVI still "owns" SDF and that VITAL also has it's own version. It was  
   suggested that a task force be formed to study all the versions of SDF and  
   report it's findings to the 1364 working group.

   Motion: Form a task force to look at the SDF. Leigh Brady will chair, Bob
           Sahai of Fintronic and Peter Eichenberger will participate.
   Proposed: Joseph Skudlarek
   Seconded: Alec Stanculescu
   Passed unanimously by voice vote.

   a) Lynn Horobin will send copies of the SDF 2.1 to the task force members
      when she recieves it.

6. Status of Specification: Yatin Trivedi gave an update on the work done to the
   LRM to date, as follows:

   II. Vectored & Scalared - The vote was to remove references to vectored and
   scalared and define them as reserved words.  Work is completed.

   III. Attributes - This section of the proposal still needs to be discussed.

   IV. Signed/unsigned - The vote was to remove all references to signed/
   unsigned. Work is completed.

   V. Macro Arguments - The vote was to leave macro arguments in specification.
   Left as is.

   VI. Compiler Directives - The proposal suggested the removal of references to
   "accelerate" and noaccelerate"; "protect and endprotect"; "protected and
   endprotected" and define as reserved words. No agreement was reached. Left as
   is.

   VII. Synthesizeable Constructs - The proposal recommends the removal of
   Appendix E. Work is completed.

   VIII. Appendix F reserved list - The proposal recommended the removal of
   attribute, scalared, vectored and endattribute, and the addition of real.
   Work is completed.

   IX) Yatin has also corrected some typographical errors, and will continue
   to look for errors.

   X) There was some discussion about how and why some of the decisions were
   made at the last meeting. Maq Mannan pointed out that it would be very
   unproductive to re-open issues that have already been decided and voted on.
   If the group is positively convinced that it wants to re-visit a certain
   issue, then it will be put to a vote.  

6. Mapping out a "Game Plan":  Maq Mannan reviewed the four goals established
   at the last two meetings and the three task forces that have been set up.
   The timeline is as follows:

   April 22 : Group should have a good idea of what will stay in and what
              is out of the spec.  
              John Mancini will have made contact with the IEEE editor.
              Ballot lists are due to John Mancini from all Working Group
              members.

   May 25   : Yatin will send the draft specification out for review/approval
              by the working group.
              PLI task force to complete PLI draft.
              Upon approval, PLI & LRM draft will go to IEEE editor.

   July 13  : Specification should be finalized by IEEE editor.
              Working Group to finalize all additions and changes to the LRM.
              Submit to IEEE for balloting.
              Begin discussion of next steps.

   a) There was some discussion about how realistic this timeline is based on
   some of the issues at hand, such as the PLI. It was already agreed that
   the group would start with PLI 1.0, and Rajeev Madhavan's task force must
   attend the meeting and submit their proposal and discuss the issue. Drew
   {*filter*} offered to put together a list of inadequacies and clarifications
   needed in the PLI 1.0. (See line item 12 of these minutes for more on PLI)

   b) Alex Zanfirecsu asked if there was a "plan B", or a way to intensify the
   effort. Maq Mannan said that plan B is that the specification will be late.
   Maq also stressed that the group must function as a team with a common goal.
   Everyone agreed that the timeline is aggressive, but the goal is attainable.
   In relation to the PLI, the best case is that the PLI task force presents a
   good proposal and it is accepted. The worst case is that no one is willing
   to work on the PLI, and PLI 1.0 will be submitted as is.

   Motion: Accept the timeline as discussed.
   Proposed: Alex Zanfirescu
   Seconded: Mike MacNamara
   Passed by majority vote, with 1 abstaining.

7. Attributes: Leigh Brady recommended that attributes be eliminated from the
   1364 specification. She said that not enough consideration was given when
   attributes were originally proposed. Cadence is the only company to implement
   attributes, and no one else is using them. There is not time to prepare a
   well-thought-out proposal, and this will hold things up.

   Motion: Remove attributes from specification and reserve the work
           "attributes" for the future.  
   Proposed: Leigh Brady
   Seconded: Mike McNamara
   Passed by unanimous vote.

8. Discussion on "vectored and scalared": Leigh Brady explained the implications
   of changing and/or removing these words, and opened a discussion on how
   different simulators respond and implement vectored and scalared. Leigh
   recommended that these words not be removed. There was continued discussion
   on this issue and how to resolve. Leigh Brady will make a proposal  and
   submit to the group for E-mail discussion and vote at the next meeting. Also
   included in this proposal will be performance, signal strengths, individual
   bit access, PLI, net connections, enforcement: warning or error.

9. Task force IV Proposal: Elliot Mednick was not present to explain his
   proposal. Leigh Brady attempted to decipher. It was generally agreed that
   Elliot needs to be present for discussion of this proposal.

   Motion: Ask task force IV to clarify section 1, items 1 - 5.
   Proposed: Mike McNamara
   Seconded: Joseph Skudlarek
   Passed by majority.

   Motion: Skip section 1, item 6
   Proposed: Alex Zanfirescu
   Seconded: Leigh Brady and Alec Stanculescu
   Passed by majority, 1 against, 1 abstain

   Motion: Section 2, Part A: change wording so that "input/output" and inout
   declarations are there to detect inconsistent port pairs. Implementation may
   enforce the declaration or allow coersion with warnings.

   Proposed: Joseph Skudlarek
   Seconded: Mike McNamara
   Passed by majority, 1 abstaining

   Motion: Section 2, Part B: If an object is declared inconsistent within the
           module or primitive, it is a warning.
   Proposed: Joseph Skudlarek
   Seconded: Alex Zanfirescu
   3 for, 2 against, 1 abstain. No majority, motion not passed.

   Motion: Reject Section 3
   Proposed: Joseph Skudlarek
   Seconded: Leigh Brady
   Passed by majority, 1 abstaining

   The remainder of this proposal from Section 4 on will be deferred for
   discussion by email.

10. Net Shorts Proposal:

    Motion: Verilog specification must model net shorts.
    Proposed: Joseph Skudlarek
    Seconded: Brian Griffin
    5 against, 2 in favor. Motion rejected.

    Motion: Analyze items 1 - 5 of Task Force IV proposal - at  same time check
    if net shorts is feasible, and report back to group. Joseph will work with
    Elliot.
    Proposed: Alex Zanfirescu
    Seconded: Mike McNamara
    Passed by unanimous vote.

11) Bounded Limits Proposal:

    Motion: Accept Item 1, part 1 as is.
    Proposed: Joseph Skudlarek
    Seconded: Mike McNamara
    Passed by unanimous vote.

    Motion: Accept Item 1, part 2 ammended as follows:  The maximum length of
    an identifier <MIN_LIMIT> is implementation -dependent but will be at least
    65536 characters. p. 1-5; p. 2-8.
    Proposed: Joseph Skudlarek
    Seconded: Mike McNamara
    Motion passed by majority, 1 against, 2 abstain.

    Rest of proposal discussion to be continued at next meeting.

12. PLI Discussion: Joel Paston was present at Rajeev Madhavan's request and
    noted that he represented himself only and was expressing his own opinions.
    The PLI task force has not yet met. He stated that PLI 1.0 is not a complete
    functional specification. Someone must go through it to define it as an
    acceptable specification. The PLI 2.0 is a functional superset of PLI 1.0,
    and work on PLI 1.0 will result in a version similar to 2.0. We need to
    identify a basis to start from. There was a discussion that some simulators
    are not able to implement version 1.0. Cadence has implemented about 50%
    of the PLI 2.0. There are no customers implementing 2.0. The discussion
    continued about which version to use as a basis, how much work was involved
    and that the working group not discount all the work done in OVI on the PLI.

    It was decided to let the PLI task force to come up with recommendations
    on PLI at the next meeting.  If the PLI task force cannot come up with a
    recommendation than PLI 1.0 will be submitted for standardization. The
    PLI task force should :

       * Get PLI 1.0 fully defined
       * Get PLI 2.0 fully defined and
       * Map PLI 1.0 in terms of PLI 2.0.

    The meeting was adjourned at 3:15PM.

Next Meeting : April 22, 1994, 9:00AM to 3:00PM, at Cadence (San Jose &
               Chelmsford). Exact location and agenda will be sent out later.  

--
Elliot Mednick                                      P.O. Box 150
Wellspring Solutions, Inc.                          Sutton, MA.  01590
                                                    (508) 865-7271