======================================================================
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 April 7, 2010:

Attendees:
----------
(* denotes present)
Agilent                    - Radek Biernacki*, John Moore, Ken Wong
Ansoft                     - Denis Soldo
Cadence Design Systems     - Terry Jernberg, Brad Griffin
Green Streak Programs      - Lynne Green
Hewlett-Packard            - Rob Elliott
IBM			   - Greg Edlund*, Lijun Jiang, Yaping Zhou
Intel                      - Michael Mirmak*
I/O Methodology		   - Lance Wang
Mentor Graphics Corp.      - John Angulo*, Vladimir Dmitriev-Zdorov
Micron Technology          - Justin Butterfield*, Randy Wolff*
Sigrity                    - Sam Chitwood, Raymond Y. Chen, Tao Su, Brad Brim
SiSoft                     - Walter Katz*
Teraspeed Consulting Group - Bob Ross*

========================================================================
No patents were announced.  

During opens, Michael Mirmak thanked Bob Ross for serving as chair of 
the meeting, and also thanked SiSoft for their donation of LiveMeeting*
resources.

Greg Edlund noted that the TSIRD for binary formatting was reviewed in 
the last meeting. A volunteer is needed to wordsmith first three
paragraphs.  Michael was volunteered to help.

Bob Ross noted that the [Binary] keyword was decided in last meeting as the 
keyword to indicate file format.  The actual section division of the
TSIRD is to be decided.

Bob reviewed his comments on the file.  He suggested [Binary] is to 
apply to TS2 and above.  [End] is still under discussion and is likely
to be left off at end of binary conversion.

John Angulo asked whether the [End] keyword starts on a new line and
whether it can be omitted in binary files.  Walter Katz suggested that 
no point existed in having the [End] keyword in binary files.

Michael noted that [End] came from IBIS.  Bob added that it was not 
needed for Touchstone; adding it was a consistency issue with IBIS.

Michael asked whether any other technical issues existed that need 
closure, aside from [End]. Bob added that he would like to assume 
noise data restriction/elimination and omission of the [End] keyword.

Walter observed, per a question from Michael, that CPU time for reading 
a file in binary is reduced up to 3x with 2-3x file size compression.
John added that there was likely to be no big impact from [End] 
keyword either way

Walter asked whether the parser will be enhanced to read in and write 
out Touchstone2 files.  Would IBM be willing to donate this code?  
Walter noted that the parser already converts v1 to v2, suggesting
binary conversion would encourage source code acceptance and income
by IBIS Open Forum.  Bob supported this approach.

Radek responded that this goes beyond current parser definitions; 
reading is parsing, whie conversion is more than parsing.  

John commented that binary reading is within the scope of the 
standard.  Radek agrees with the reading part.  Conversion is still 
under discussion.

Radek asked whether the document should include the exact meaning 
of the format, including definitions of little endians, double-floats,
etc.  This is not ambiguous now, but based on common knowledge.  

The team discussed whether binary formats cause issues across operating
systems. Greg stated that he does not have test cases across different 
architectures.

Walter noted that architecture, not OS, is the problem. IEEE standards 
govern how bits are stored.  Any 32-bit or 64-bit machine would be 
OK, but an older 60-bit machine might cause issues.  Radek's point 
about formats relates here.  Quad bit format is still being 
discussed.

Bob added additional comments: 

- example 1 should use bracketed keyword
- a readable HEX- or character-encoded example should be added

Greg asked if we can remove the [Binary] DFB in the example.

Radek observed that the example is inconsistent; we need to have 
ASCII and binary side-by-side or in independent examples.  Michael added
that whether the binary is "inline" or covers the entire file is 
somewhat ambiguous.  Bob aske whether the binary code should show 
up as a separate section in an ASCII file?

The team agreed to reconvene on April 14th.

ARs
Michael, Bob: to work on TSIRD formatting and links