======================================================================
IBIS FUTURES/COOKBOOK TASK GROUP MEETING MINUTES
Date: February 2, 2006

http://www.eda.org/ibis/futures/
http://www.eda.org/ibis/cookbook/

Attendees:
----------
Agilent                 - Sanjeev Gupta 
Cisco Systems           - Syed Huq
Intel                   - Michael Mirmak, Arpad Muranyi
Mentor Graphics         - John Angulo, Ian Dodd
Micron                  - Randy Wolff
Teraspeed               - Bob Ross

======================================================================

Next Meeting:
Thursday, February 16, 2006
9 AM - 10 AM US Pacific Time 
(no meeting Feb. 9)

               Telephone     Bridge   Passcode
              916-356-2663     2      589-9198

Agenda:

9:00 - 9:10 AM     Opens
                   
9:10 - 9:30 AM     IBIS-ICM Simple Link, continued

9:30 - 9:50 AM     Simple Proposal on User-Defined Measurements
                   
========================================================================

Michael reviewed a simple proposal for linking ICM and IBIS.  Most of
the suggested additions to the [Pin] keyword, to support complex IBIS
3.2/4.0 packages, are removed in the simplified proposal.  Additionally,
[Pin Mapping] is essentially unchanged.  Only 1-to-1 mapping of nodes
and pins is supported, unless IBIS 4.1 [External Circuit] is used.

Arpad Muranyi inquired whether positional rather than name matching
could be implemented.  Bob Ross and Michael both noted that matching 
the sequence between name and position would be very difficult in this
case.

Arpad also inquired whether [External Circuit] can support the 
multiple-path links removed from the [Pin] keyword in this proposal. 
Michael responded, yes.

Sanjeev Gupta asked whether the ICM/IBIS links can be considered both 
"horizontal and vertical" - in other words, whether the parsing is
conducted in the [Pin] list line-by-line or across a single line.  
The specifications allow some aspects of both.

John Angulo noted that power buses need very clear connections to
enable power delivery simulations.  Does the new proposal assume
the die nodes are linked in a 1-1 manner? 

Michael noted that [Pin Mapping] may take care of some of these linking
concerns.  As each buffer is identified by pin number, and 1-to-1 
linking is assumed for [Model], few problems should arise with
power supply connections.

Bob inquired whether [Pin Mapping] can permit partial pin maps.  Michael
observed that the current definition does not allow this, but that
a relaxation might be possible.

The team agreed that a general review of the proposal would be eased
when an actual BIRD draft is available for examination.

Michael briefly mentioned a suggested user-definable measurement
scheme where AMS ports are used to pass measurement information to
the tool.  The team agreed to take this proposal up after the DesignCon
summit, where this is likely to be discussed in-depth.