IBIS Macromodel Task Group

Meeting date: 5 May 2008

Members (asterisk for those attending):
  Ambrish Varma, Cadence Design Systems
* Anders Ekholm, Ericsson
* Arpad Muranyi, Mentor Graphics Corp.
  Barry Katz, SiSoft
* Bob Ross, Teraspeed Consulting Group
* Brad Brim, Sigrity
* Brad Griffin, Cadence Design Systems
  David Banas, Xilinx
  Donald Telian, consultant
  Doug White, Cisco Systems
  Essaid Bensoudane, ST Microelectronics
* Fangyi Rao, Agilent
  Ganesh Narayanaswamy, ST Micro
  Gang Kang, Sigrity
  Hemant Shah, Cadence Design Systems
* Ian Dodd
  Joe Abler, IBM
* John Angulo, Mentor Graphics
  John Shields, Mentor Graphics
  Ken Willis, Cadence Design Systems
  Kumar
  Lance Wang, Cadence Design Systems
  Luis Boluna, Cisco
  Michael Mirmak, Intel Corp.
  Mike LaBonte, Cisco Systems
  Mike Steinberger, SiSoft
  Mustansir Fanaswalla, Xilinx
  Patrick O'Halloran, Tiburon Design Automation
  Paul Fernando, NCSU
* Radek Biernacki, Agilent (EESof)
* Randy Wolff, Micron Technology
  Ray Comeau, Cadence Design Systems
  Richard Mellitz, Intel
  Richard Ward, Texas Instruments
* Sam Chitwood, Sigrity
  Sanjeev Gupta, Agilent
  Shangli Wu, Cadence
  Sid Singh, Extreme Networks
  Stephen Scearce, Cisco Systems
  Steve Pytel, Ansoft
  Syed Huq, Cisco Systems
  Syed Sadeghi, ST Micro
* Terry Jernberg, Cadence Design Systems
  Todd Westerhoff, SiSoft
  Vikas Gupta, Xilinx
  Vuk Borich, Agilent
* Walter Katz, SiSoft
  Zhen Mu, Cadence


-----
Opens:

Arpad Muranyi has returned as chair.  Randy Wolff is taking minutes in Mike 
Labonte's absence.


--------------------------
Call for patent disclosure:

- No one declared a patent.


-------------
Review of ARs:

- Walter: Send "ICM vs EMC" to Mike for web posting
  -Done, should also list EMD, not EMC.

- Mike M: Prepare a list with pros and cons about ICM, EMD, and
          our goals (etc...) for discussion in the next meeting
  - not done

Old ARs:

- David Banas: Report Xilinx position on LTI assumption for SerDes
  - No update

- Arpad:  Write parameter passing syntax proposal (BIRD draft)
          for *-AMS models in IBIS that is consistent with the
          parameter passing syntax of the AMI models
          - not done

- TBD:    Propose a parameter passing syntax for the SPICE
          [External ...] also?
  - TBD

- Arpad:  Review the documentation (annotation) in the macro libraries.
  - Deferred until a demand arises or we have nothing else to do

-------------
New Discussion:

Arpad asked for a short summary on Kumar's comments and how it was
handled in the last meeting ("Kumar is no longer affiliated with
Cadence" in the minutes doesn't explain much...).

Terry: Questioned if Kumar was in agreement to the addition to the AMI BIRD.  
He didn't think Kumar was opposed to the added switch keyword, he just didn't 
think it was necessary.  The added switch was a compromise between Candence 
and SiSoft.  Not being affiliated with Cadence referenced how Kumar did not
have background on the compromise because he was not present in the meetings
between Cadence and SiSoft.

Arpad: Are there any concerns with BIRD 107.2 and 108?
- The consensus was there are no concerns to address.

Arpad: Need any discussion on Bob's editorial changes in BIRD 107.2 and 108?
- Bob: No, BIRD107.2 is up for a vote in the next IBIS teleconference.  BIRD108
  is up for introduction.

Arpad: Still delaying IBIS version 5.0?
- Bob: Yes, need to delay it to get these editorial BIRDs included.


Continue discussion on Electrical Module Description
- Repeat with emphasis:  Need to discuss where we are heading with
  EMD and ICM, and decide what our goals are that we want to achieve
  BEFORE going into the technical details of EMD.  It would be nice
  to Have Mike Mirmak's list or something similar ready for the
  meeting as a guideline for our discussion...

Walter: His presentation should help answer this question.  He showed the 
presentation "ICM vs. EMD - A 9 pin wire bond BGA example". He noted that the 
first picture of a 300 pin connector is well suited to ICM for RLGC 
representations.  He showed a package and a DIMM module - two examples of what
we need to be able to model.  He showed the ICM method and keywords for 
modeling a 9 pin package.  

Brad Brim: What do we want to do in ICM that we can't do currently?
- Walter: Continued to present same 9 pin BGA example in EMD format.  Then he 
  noted that trying to model a DIMM in ICM is very complicated.
- Bob: A tool could automate generation of a DIMM model in ICM with many 
  sections, coupled and uncoupled.

Walter: S-parameters are needed for SerDes applications.  He showed an example
of S-parameter usage in EMD.  This can also be done in ICM.  But, sparse 
S-parameter files are needed.  ICM falls short with packages of any real 
complexity.

Arpad: So far all examples shown can be done in ICM and EMD.  Sparse 
S-parameters are the only new thing. 
- Walter: No, he wants to mix S-parameters with RLGC.  Also, one may want to 
  include impulse responses or pole and zero responses.  He wants to deal with
  this with a netlist structure that is easy to work with.

Arpad: Sees that these ICM limitations exist.  Why not fix ICM instead of 
abandoning it?
- Walter: Any volunteers?
- Arpad: Thinks it will be more work to come up with a new spec than fixing ICM.
- Walter: Committee needs to decide this exact question.

Brad Brim: What are the requirements needed in any spec?  Making a Spice-like
netlist parser is a big deal.

Arpad: Sees usefulness to [Pin Mapping].
- Walter: Likes Verilog and VHDL that include node description at beginning of
  model.
- John Angulo: ICM format makes parsing easier by being explicit in its Pin 
  Mapping.

Walter: ICM is too strict in what it is designed to model.  What about a 
situation to model power delivery with different numbers of pins on each side
of a structure?
- Arpad: Thinks ICM can handle this, for instance modeling cables with splits. 
  He challenged Walter to show specific examples of where ICM can not model a 
  structure.

Anders: What about multi die packages and trying to model this in ICM?
- Bob: This is very difficult, but could be done.

John: Discussion showing ICM is verbose and complicated.
 
Arpad: When mixing RLGC and S-parameters, don't you lose coupling information?

Sam Chitwood: Part of the reason ICM hasn't been improved in the last two years 
is because the group came to the realization that using Touchstone was an 
issue. Touchstone needed to be cleaned up first.  Sparse S-parameter capability
would be very useful, but it should go into the Touchstone specification.

Arpad: Out of time.  We will continue the discussion next week.  Walter, can 
you bring an example of real limitations in ICM?
- Walter: Will put some thought into it.

-----------
    
Next meeting: 13 May 2008 12:00pm PT

-----------
