[sv-ac] Minutes from SV-AC Meetin 9/21/2010

Minutes from SV-AC Committee Meeting
Date: 2010-09-21
Time: 16:00 UTC (9:00 PDT)
Duration: 1.5 hours

Dial-in information:
--------------------
Meeting ID: 38198

Phone Number(s):
1-888-813-5316 Toll Free within North America

Agenda:
-------
- Reminder of IEEE patent policy.
See: http://standards.ieee.org/board/pat/pat-slideset.ppt

- Minutes approval

- F2F meeting (discussion)

- Email ballot results
2476: Need clarification about system functions $onehot, etc
Passed: 13y/0n/0a

- New issues

- Enhancement progress update
Checker usability enhancements

- Issue resolution/discussion
3008: In $past BNF, "expression" should be "expression1"
1853: BNF for calls to $rose and other sample value system functions.
2904: Clarify when disable iff condition must occur relative to starting
and ending of an attempt
3134: sequence and property range parameters are erroneously defined
3135: Verbal explanation of nexttime and always is misleading for
multiple clocks
1678: Clarify that rewriting algorithm doesn't replace name resolution
2571: confusing assertion clock inference rule
2386: Rename 16.9 to "Local variables"?
3117: make it clear that rewriting algorithm (F.4.1) applies to checker
and let

- Opens

Attendance Record:
------------------
          Legend:
                  x = attended
                  - = missed
                  r = represented
                  . = not yet a member
                  v = valid voter (2 out of last 3 or 3/4 overall)
                  n = not a valid voter
                  t = chair eligible to vote only to make or break a tie

Attendance re-initialized on 2010-07-06:

   v[x-xxxxx--xxx] Laurence Bisht (Intel)
   v[xxxxxxxxxxx-] Eduard Cerny (Synopsys)
   v[xxxxx-xxxxxx] Ben Cohen
   v[x-x--xxxxxxx] Surrendra Dudani (Synopsys)
   v[xxx---x-xxxx] Dana Fisman (Synopsys)
   v[x-x-xxxxxxxx] John Havlicek (Freescale)
   v[xxxxxxxxxxxx] Tapan Kapoor (Cadence)
   t[xxxxxxxxxxxx] Dmitry Korchemny (Intel ¿ Chair)
   v[xxx-xxxxxxxx] Scott Little (Freescale)
   v[xxxx-xxxxxxx] Manisha Kulshrestha (Mentor Graphics)
   v[xxxxxxxxxxxx] Anupam Prabhakar (Mentor Graphics)
   v[-xxxxxxx-xxx] Erik Seligman (Intel)
   v[xxx-xxxxxxx.] Samik Sengupta (Synopsys)
   v[xxxxxxxx-xxx] Tom Thatcher (Oracle ¿ Co-Chair)
     |- attendance on 2010-09-21
   |--- voting eligibility on 2010-09-21

Minutes:
--------

- Reminder of IEEE patent policy.
See: http://standards.ieee.org/board/pat/pat-slideset.ppt
Participants were reminded of the policy.

- Minutes approval
Scott: Move to approve minutes
Tapan: Second
        Vote results: 12y, 0n, 0a

- F2F meeting (discussion)
     Silicon Valley: Tom, Anupam, Erik, Dmitry
     Ed, Surrendra: Suggested Boston

Dmitry: Would it help to move to coincide with DV-CON? (Feb 28--Mar 3)
Ed: May help, but can't know in advance. Would also have more things
        to discuss at this time.
Tom: Agree that we would have more to discuss in DV-con time frame.
Dmitry: Also, Nov date would coincide with IC_CAD as well.
Dmitry: Will conduct one more poll on the time-frame, Decide next time.

- Email ballot results
2476: Need clarification about system functions $onehot, etc
Passed: 13y/0n/0a
Dmitry: Has implemented two friendly amendments.
Dmitry: This proposal takes care of 3008 as well.

- Issue resolution/discussion
3008
Scott: Move to close 3008 as resolved by 2476.
Tom: Second
Vote Results: 12y, 0n, 0a

- Enhancements progress update
Checker usability

Dmitry: Discuss definition of sampling.
Dmitry: Current proposal: "Sampled" defined differently for different data:
        Example: checker rand variables, $sampled returns value assigned
                at beginning of Observed
                Automatic variables: $sampled(x) = x
Manisha: Definition seems fine: just confused about automatic variables.
        They have short life span. They may not exist at end of simualtion
        tick
Dmitry: Then it may be illegal to call $past on an automatic.
        May also be hard to define $past for local variables.

Dmitry: Continuous assignments and always_comb/blocking assignments.
Tom: Make things consistent. Continuous assignments in checkers always
        update in the Reactive region. Even if a checker input changes in the
        active region, the assignment would be scheduled for the Reactive
        region.
John: Will have to think about examples of interacting agents.
Manisha: Are you sure that design continuous assignments always update in
        the Active region?
Dmitry: Pg 31 of standard: 4.9.1
        We can ask other committees to be sure.
John: Will try to think of examples.
Dmitry: Who would like to work on proposal?
Anupam: 3034, currently assigned to me.
Manisha: Can volunteer to help.

Checker outputs
Lawrence: BNF of checker outputs. Need to specify input/output of each
port.
        To maintain backward compatibility, if direction not specified, first
        port would be considered an input.
Dmitry: But it would be inconsistent with everything else?
Ben: Can't it be like a module?
Lawrence: In modules, you need to specify direction.
Ben: But some people like to put outputs first in the port list.
Dmitry: You can do it, you just need to specify "output" for the direction
        of the first port.

Lawrence: Do we want to allow untyped outputs for checker?
Dmitry: Why not?

Scott: For modules: the LRM says if the direction is unspecified, the port
        type defaults to inout. this proposal is inconsistent with this.
Dmitry: We are not intending to introduce inout ports for checkers right
now.
Manisha: Would prefer to default to input. inout ports cause more overhead.
Anupam: So we plan to make input the default direction, and we are not
        proposing inout ports.
Ben: Dmitry was talking about using a checker as a model. For this
        application, It may need tri-state ports, inout ports, etc.

- Issue resolution/discussion

2904
Anupam: Simulation should not need to retract something that has already
        been evaluated. Proposed to change the line Dana had modified.
Anupam: In a single-clocked case, should be consistent. In multi-clocked
        case there would be a race, but don't want to deal with that.
        Proposed change is in e-mail
Dmitry: Should you mention Observed region?
Anupam: Sentence mentions assertion evaluation. This occurs in Observed
        region. Shouldn't be necessary to explicity mention Observed
        region.
John: You are not specifying an ordering between disable iff condition and
        the assertion evaluation?
Anupam: Correct. No ordering is implied.

Meeting adjourned.

- Enhancement progress update
Checker usability enhancements

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.