[P1800] P1800-Meeting-Minutes-Approved-10-10-05.doc

From: Brophy, Dennis <dennisb_at_.....>
Date: Thu Jan 19 2006 - 23:36:23 PST
The approved meeting minutes from 10 October 2005 as approved at the 19
January 2006 meeting. 
________________________________


IEEE P1800 SystemVerilog Working Group

APPROVED MINUTES

10 October 2005

10:00 a.m. - 2:45 p.m. PDT

 

 

1.     CALL TO ORDER

 

Chair Srouji called the meeting to order at 10:15AM PDT. Attendance is
shown below.

 

 

 

Key:

 

 

 

 

 

DR= Designated Representative 

DRA= Designated Representative Alternate

O=Observer

S-Staff

 

 

 

 

 

 

 

 

 

Present

Entity Name

Representative

Name

Representative

Designation

IEEE-SA

Member

Voting Member

X

Accellera

Karen Pieper

DR

Yes

Yes

X


IBM

Johny Srouji

DR

Yes

Yes

 

X

X

Mentor Graphics

Dennis Brophy

Dave Rich

Doug Warmke

DR

DRA

O

Yes

Yes

X

Sutherland HDL

Stu Sutherland

DR

Yes

Yes

X

Cadence

Victor Berman

DR

Yes

Yes

X

Synopsys

Oz Levia

DR

Yes

Yes

X 

Fintronics

Alec Stanculescu 

DR

Yes


Yes

X

Synopsys

Brad Pierce

O

 

 

X

Sunburst Design

Cliff Cummings

O

 

 

X 

Sun

Neil Korpusik

DR

Yes

Yes

X

Intel

Yossi Levi

DR

Yes 

 Yes

X

IEEE-SA

Noelle Humenick

S

 

 

 

 

 

 

 

 

 

 

 

2.     Approve agenda

 

A motion (Victor Berman, Stu Sutherland) was made to approve the agenda.
The motion approve unanimously.

 

 

3.     CALL FOR PATENTS

 

Noelle Humenick read the patent slides.

 

 

4.     Approval OF 29 JULY 2005 meeting minutes 

 

A motion (Victor Berman, Karen Pieper) was made to approve the meeting
minutes.  The motion approved unanimously.

 

 

5.     Action Items Review and Follow-up

 

Johny reported that all action items have been closed, except for
Letters of Assurance (pending and on track) 

 

 

6.     Update from RevCom

 

Johny explained that P1800 and P1364 submittal packages were prepared
and submitted to RevCom. Both Standards postponed pending inclusion of
names of negative balloters for both ballots. 

 

Noelle provided a report indicating that recirculation for P1800 is now
closed and acceptable to Dave Ringle-RevCom. Recirculation for P1364 is
now closed and also acceptable. Next RevCom meeting will occur on
October 27 or 28, where both standards will be considered on that date.
Publication for P1800 is targeted for November 9, 2005 (funded project).
Publication date for P1364 is March 2006 (unfunded project and not
eligible for accelerated publication).

 

ACTION REQUEST - Stu & Noelle have an action item to get Tapati Basu's
name from sv-cc committee on the P1800 Standard.

 

 

7.     Financial Report 

 

Noelle presented that the working Group will have used all IEEE-paid
funds as of the 9th of November. No outstanding invoices are expected.
No expenses are anticipated beyond what is currently budgeted. (Thanks,
Victor and Oz!)

 

 

8.     Review overall schedule of the P1800 WG

 

P1800 approval date is November 3rd and a press release expected 1-2
days after P1800 release. The working group anticipates submitting the
P1800 Standard to the IEC

 

 

 

 

9.     Two year future roadmap discussion for 1800 & 1364 

 

Johny presented the following document regarding possible items for
discussion for future activities related to SystemVerilog: "P1800 and
P1364 Long Term Plans; Structure and Operation"

 

Possible future activities voiced by Oz

*        Support

*        Bug-fixing

*        Merging of 1800 & 1364

*        Enhancements

*        Further standardization into International Organizations

Alec noted IEC Standardization is very important as part of the original
P1800/P1364 PARs

 

Perhaps Analog Mixed Signal (AMS) should be included in a future version
of the SystemVerilog standard.

 

 

User comments:

 

Yossi - Intel

Intel is pleased with the SystemVerilog work. Need to have live
committees to report problems and request enhancements. Some
enhancements were not added to P1800 due to accelerated schedule. Yossi
expressed the desire to consider extensions going forward. A two year
turn of the standards is a good idea. More than two years might promote
diverging standards. A desire to merge 1800, 1364 and AMS was also
expressed. Also would like syncing of PSL & SVA. Enhancements are the
most important issue.

 

Neil - Sun 

Sun does not currently use SystemVerilog but will be using SVA shortly -
limiting factor is tool availability. Next capability of interest is the
design subset. No big need for AMS (not actively used) - merging LRMs is
a good idea (not sure if two years is doable for the merge) - would like
a new update in two years or less. Perhaps merge committees (1364 &
1800). Clarifications and bug-fixes are most important.

 

Cliff - Sunburst Design

Users are looking for live support for bug fixes and clarifications.
Users are initially interested in SystemVerilog verification. Users
would like enhancements like AMS, 2-State, and other incremental
improvements to new SystemVerilog features. Merging is considered a good
idea - some of the SystemVerilog sections can be moved into a new
document without any merging effort.

 

 

Stu - Sutherland HDL

Seconds preceding comments - suggested separated volumes or
dot-standards for modeling, testbench, SVA, and API. Take PLI into the
API document. AMS could be another volume or dot-standard. Live document
with errata and clarifications (highest priority) should be available
until the next release. Enhancements are also important. If we stop
doing enhancements, the language dies.

 

Victor - Cadence

Live standard with bug fixes and clarifications is the most important
activity and must have consistent releases of addendums. Combining 1364
& 1800 could be more difficult than anticipated. It may make sense to
re-write the LRM to put definitions in a single place and to make
standard more concise. Note - re-write could cause difficulties with
existing tools. Enhancements are nice but not the top priority. Groups
should remain pretty much the same except merge 1364 and 1800 groups.

 

Dave Rich - Mentor

1800 & 1364 should be merged (important to users and vendors) - 1800
spec is not currently up to par and needs at least one round of updates
and clarifications before merging documents. 1364 & 1800 committees
could be merged, PLI/API/CC-partition, Bug-fixes/Modeling. A forum is
needed to make clarifications and corrections that will be firm and
web-available for the user community to view. Generic interoperability
standard would be good (not just but including VHDL). Role of Accellera
should be to form the support group for SystemVerilog.

 

Brad Pierce - Synopsys

Personally feels that merging standards is a waste of time.
Interoperability with VHDL could be standardized. Enhancements are going
to happen so it is matter of whether they will be captured or not (more
important than fixes).

 

Doug Warmke - Mentor

Priority should be clarifications and bug-fixes - a forum to clarify and
document fixes is required. Future enhancements need to be prioritized.

 

Oz Levia - Synopsys

EDA vendors want to make a profit and sell tools - SystemVerilog should
be a viable business. Integration with other standards like VHDL is
extremely useful. Priorities are: 

#1 - bug fixes & clarifications

#2 - integration with other interesting standards

#3 - future extensions and enhancements to SystemVerilog (selecting
enhancements needs to be carefully considered and accepted). 

#4 - merged LRM is important, but a lot of work. Current organization is
good, even the separate 1800 and 1364 committees.

 

Alec - Fintronics

The LRM is very large - there should be different books for different
parts of the standard and the PLI should be part of the standard.
Deprecation of features is a bad idea. Putting tf & acc features could
be part of a separate book. Interoperability is an important issue -
both in general and in particular with SystemVerilog, VHDL & SystemC.
Most important interoperability is between Verilog & SystemVerilog.
Support of committees should be done through the IEEE.

 

Karen - Synopsys

Agrees with Brad & Oz's comments - Should keep committees structured
much as they are, perhaps with some minor merging between 1364 & 1800.
Merging errata & enhancement committees would be a good idea because
they have similar knowledge and skills. Personal opinion - prioritize
errata and enhancements. Interoperability with VHDL is a good idea. No
great return for merging multiple standards.

 

Johny - IBM

Prioritize between balanced set of enhancements and bug fixes. Fix
important bug fixes. Add enhancements where divergence could happen to
satisfy customer requests. Current committee structure is pretty good as
is. Two years time-frame for next approved LRM is a good timeframe, but
we do not need to fix this schedule. Errata and clarifications and a
live standard, perhaps every 6 months would be a good idea. The #1
priority is to make sure there are no incompatibilities between
SystemVerilog and Verilog. Merged LRM is a good idea but not as high of
a priority and the effort will be large.

 


SUMMARY


 

We do not see major enhancements in the next 2-year period. We do see a
need for prioritized bug-fixes and enhancements. Set a goal to complete
the next IEEE SystemVerilog Standard in two years (completion could be
anything from end of technical work to actual approved and published
IEEE Standard).

 

We need an approved mechanism to make corrections and revisions for
release at least every six months.

 

 


Merging Discussion


 

Stu believes it is essential to merge the 1364 & 1800 documents to (1)
make tools support the full merged language, (2) users can write code
without `ifdef's,  (3) the working group already said we would do it.

Victor would like a re-written and merged LRM but does not believe it is
critical to merge the documents.

Oz would like a merged LRM but is concerned that if merging has priority
that the next revision will be significantly delayed.

Karen is concerned that it will require an engineering-year to edit the
document, and 6 months - 1 year for a committee to review the edited
document (committee of ~6 engineers).

Brad fears that nobody will want to participate on the merging
committee.

 

Points of disagreement:

(1)   The effort required to merge the documents

(2)   The quality of the merged document

 


Discussion of Operation and Structure of Technical Committee


 

At this time, keep the committees as they are and Karen has the action
item to return and propose changes if she deems the changes are
beneficial.

 


Concerning 1364 Committees


BTF has been folded into the BC.

 

Encryption committee is now dissolved and errata will be addressed by
the BC.

 

1364 Database will be merged with the SystemVerilog database.

 

Proposal (Karen - 2nd Neil - unanimous passed) - move the encryption and
errata committees into SV-BC.

 

 


Action Requests


 

Johny to look into requirements and rules related to making
SystemVerilog corrections and revisions available for public release.
Johny to find the IEEE-proper method to possibly make enhancements to
the SystemVerilog Standard.

 

Stu to do sizing and estimates related to merging the documents and
report back to the committee at the next meeting.

 

Karen & Victor to also do sizing of the review effort and report back to
the committee at the next meeting.

 

Karen has the action item to return and propose changes to the
committees if she deems the changes are beneficial.

 

Proposal (motion: Victor - 2nd Stu - unanimous passed) this committee
authorizes Karen to establish the P1800 errata committee to begin
addressing issues related to the P1800 standard. Karen to report back to
the P1800 working group with a proposed organizational structure to
address errata in the next working group meeting.

 

10.             Future P1800 Interface Work 

 

Desire to enable interfaces to VHDL, AMS and SystemC.

 

First interface to address should be VHDL.

 

Action Item: Johny and Karen to facilitate first meeting between
SystemVerilog and VHDL groups.

 

11.             NEXT MEETINGS

 

Potential dates and places for the next meeting (November time frame)
will be proposed by chair Srouji.

 

 

12.             ADJOUNMENT

 

A motion (Stu Sutherland) was made to adjourn the meeting.  The meeting
adjourned at 2:45 p.m. PDT.
Received on Thu Jan 19 23:37:01 2006

This archive was generated by hypermail 2.1.8 : Thu Jan 19 2006 - 23:37:49 PST