======================================================================
IBIS INTERCONNECT MODELING AD HOC TASK GROUP MEETING MINUTES AND AGENDA

http://www.eda.org/ibis/adhoc/interconnect/

Mailing list: ibis-interconn@freelists.org

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

Minutes from July 15

Attendees (incomplete):
----------
(* denotes present)
Agilent                    - Radek Biernacki*, John Moore, Ken Wong
Ansoft                     - Denis Soldo
Cadence Design Systems     - Terry Jernberg, Brad Griffin
Cisco Systems              - Mike LaBonte
Green Streak Programs      - Lynne Green
Hewlett-Packard            - Rob Elliott
Intel                      - Michael Mirmak*
Mentor Graphics Corp.      - John Angulo, Vladimir Dmitriev-Zdorov*
Micron Technology          - Randy Wolff
Sigrity                    - Sam Chitwood, Raymond Y. Chen, Tao Su, Brad Brim*
SiSoft                     - Walter Katz*
Teraspeed Consulting Group - Bob Ross*

======================================================================
No patents were declared.

Radek reviewed his e-mail proposal for simplified sparse matrix format.  A single header, in the form of row and column pairs, would be used in the proposed format, with as many pairs as present in the data section.  A new optional keyword, [Implicit Sparse Matrix Entries], would be defined.

Walter asked for clarification whether this proposal was essentially equivalent to Vladimir's and Michael's proposals, where [Sparse Matrix Entries] specified locations (1,1) (1,5) (3,1) in the full matrix, but the order of the pairs referred to the order of the data under the sparse keyword. 

Walter asked what engineers would do, as this would help direct the team between the equivalent proposals.

Bob suggested that some sort of symbolic mapping is needed, to associate "one to the many."

Michael asked about [Number of Ports] and whether it would still refer to the full matrix or the reduced matrix.  The team confirmed it would refer to the full matrix, but matching would not take place between the keyword argument and the data section size.  

Walter suggested that "transfer function" rather than "vector" be used to describe the keyword data.  Further, he proposed giving each transfer function a specific name or identifier.  This would prevent confusion in mapping row, column entries and package pins.

Walter stated that readability would be an issue, unless row vs. pin associations were clearly noted.  Michael added, with Walter's agreement, that this would be similar to the approach used in ICM.

Bob disagreed, stating that no Touchstone file is readable today, and that readability is not necessarily a benefit to users.  Radek added that his own proposal was created to be easy to understand.  

The team also agreed that the sparse matrix mapping would not be frequency-dependent.  Bob, Vladimir and Radek disagreed as to whether mixed-mode and sparse matrix data would interact (Bob suggested they would, with Vladimir and Radek suggesting they would not).   

Brad asked What sparse matrices without redundancy would look like, assuming zeroes are not necessarily assumed to be redundant.  The meeting ended with an AR to MM to create a mixed-mode example showing sparse-matrix interaction.

AR:
- develop example of mixed-mode under sparse format (MM)