Attendees:
Mark Hahn, Cadence
Tom Dewey, Mentor
Bob Dilly, IBM
Amlatta Gupta, IBM
Kerim Kalafala, IBM
Vikas Sharma, IBM
Alex Suess, IBM
Jim Swift, IBM
New action items:
Who When What
---------- ------ --------
1. Mark 8/10 Update table of contents
2. Mark 8/10 Prepare sample of draft format for
data_arrival_time
3. Mark 8/17 Discuss what documentation tool to
use for the draft spec with the technical
editor
Open action items:
Who When What
---------- ------ --------
1. Bob, Vikas 5/25 Review the GCF 2.0 draft on operating conditions
-> 7/6
-> 8/10
2. Mark 3/2 Add operating conditions to the taxonomy
-> 7/13
-> 8/17
3. Bob 7/27 Send description of additional open issues
encountered during DAC demo development
4. Mark 7/27 Prepare Microsoft Project schedule for the
second half of '99.
Closed action items:
Who When What
---------- ------ --------
1. Tom 7/27 Send description of additional open issues
encountered during DAC demo development
2. Tom 7/27 Send example of mixed BNF/man page documentation
style
3. Mark 7/27 Prepare table of contents
Next Meeting:
The next meeting will be a teleconference on 8/17
from 9-11 am (PDT).
Details:
1. Review action items
Tom sent an example to Mark of a Mentor command reference
for defining a clock, which used the man page style. It
didn't include a BNF description; instead valid combinations
of options were covered in the semantics discussion.
We discussed the issues that IBM and Mentor ran into during
DAC demo development.
Mentor's issues (additional types of objects besides pins for
multi_cycle_path, jitter, and skew) are already addressed in
the taxonomy but were intentionally excluded from the demo
subset.
IBM would like to see a way to specify the ideal slew at
register clock input pins (already in the taxonomy). There
is an outstanding problem with identifying operating points
by name, as in DCL/OLA, or by a corresponding numerical value
for process, as in most textual library formats. There was
also some confusion about waveform definitions, where early/late
and lead/trail don't imply which edge is rise and which edge
is fall. This was resolved during the DAC development with
a -inverted option, where the default is the leading edge is
rise, and -inverted changes it to fall.
2. Review table of contents
Mark completed a first draft of the table of contents.
We discussed a number of things that should be added:
- inheritance of constraints
- user extensions
- reserved words
- case sensitivity
- convention for comments
- include files
- pragmas for "persistent comments", preserved through
read/write cycles
- history, similar to an early concept in SPEF, each
tool that processes the constraint data adds a history
entry
We talked about whether it should be legal to specify
units and other general controls such as design name space
multiple times in a file. We agreed that this is necessary
to support hierarchical constraints, where different portions
of a design are likely to be developed by different groups
and companies in different languages.
3. Review schedule
Mark wasn't able to complete the Microsoft Project schedule.
Mentor would like to have a milestone for when the syntax
will be complete, so that tool development can begin.
4. Discuss operating conditions
Bob and Vikas weren't able to review the GCF 2.0 draft description
of operating conditions yet, so we deferred this discussion.
Thanks,
Mark
--