IBIS Macromodel Task Group

Meeting date: 18 Mar 2008

Members (asterisk for those attending):
  Ambrish Varma, Cadence Design Systems
* Arpad Muranyi, Mentor Graphics Corp.
  Barry Katz, SiSoft
* Bob Ross, Teraspeed Consulting Group
  Brad Griffin, Cadence Design Systems
* David Banas, Xilinx
  Donald Telian, consultant
  Doug White, Cisco Systems
  Essaid Bensoudane, ST Microelectronics
  Ganesh Narayanaswamy, ST Micro
* Hemant Shah, Cadence Design Systems
  Ian Dodd
  Joe Abler, IBM
* John Angulo, Mentor Graphics
  John Shields, Mentor Graphics
  Ken Willis, Cadence Design Systems
  Kumar, Cadence Design Systems
  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

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

- No one declared a patent.

-----
Opens:

Michael M: The draft IBIS 5.0 specification has been posted for review.

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

- John create document to demonstrate example of extending ICM
  - TBD, need more time
  - Still fuzzy on appropriate way to connect power and ground ports.

- Arpad send EMD presentation to Mike L for posting
  - Done

- David Banas report Xilinx position on LTI assumption for SerDes
  - Clarification: are buffer non-linearities small enough so that AMI
    models are OK?
  - Walter: nothing is linear and everything is linear (to 1st order)
    - Whether the LTI assumption is valid depends on engineering judgement.
  - Mustansir: we will have to match with spice once we have models.

- AMI BIRD authors meet to prove AMI BIRD can handle Mellitz cases
  - Todd: we need real part data
  - We decided to leave this AR on the list.

- Walter coordinate completion of true diff BIRD draft
  - In progress

- Michael M. draft Eye keywords for true diff BIRD
  - In progress

- Randy draft derating keywords for true diff BIRD
  - Randy and Walter need to meet on this
  - The Tvac keyword is figured out
  - Still in "off-line" phase

- 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
  - TBD

- 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:

Advanced connectivity approach: short term or long term solution?
- Arpad:
  - John showed us a more flexible solution that may take longer to implement.
  - Michael M. proposed something that can be done quickly
    - Some people use HSPICE circuits for all connectivity
  - Arpad: We may be stretched thin if we try to do both
- David: Doesn't PKG have coupling?
  - Michael M: IBIS gives 4 choices:
    - lumped across 3 corners [Package]
    - [Pin] RLC
    - [Define Package Model]
      - distributed single line, DC R (cascading)
      - matrix includes coupling
      - We can't have both
    - We want to have via and ball as separate sections
- Sam: Proprietary solutions today are essentially SPICE based
- We can treat IBIS as die only, and all PKG stuff is external
- Arpad: we may miss 5.0; do we need a spec change?
  - Michael M: 5.1 is supposed to be a cleanup with no major features
  - This will take a lengthy BIRD
    - Touches both IBIS and ICM
    - Still not sure about connections
    - Do we need an integrated parser for both?
  - Michael M: Not sure about the level of support for this.
    - In this case it may be better to wait for the long term solution
  - Arpad: is this too big a change for 5.1?
    - Michael M: We have done it before
- Arpad: tool vendors can provide hookup facilities on their own
- Summary:
  - It's too late for something new to go into IBIS 5.0
  - Tools will have to hook ICM to IBIS on their own
- John: Some vendor might implement something that then becomes 5.1
  - It looks easy at first glance

John shared circuit_call_to_model.txt
- [Circuit Call] si needed to bring in pad nodes beyond 1 pin and 1 pad
  - But it can only instantiate a [Model]
- Walter: There should be a model for each piece of silicon
  - Mixing that into an IBIS file will cause problems
  - There must be an independent scheme
  - ICM can't daisy chain disparate elements
- Michael M: Conceptually there can be multiple [Define Package Model]s
  - Tools can offer a choice as with [Model Selector] or [Component]
  - Arpad: Does the [Component] effectively include the package?
- John: Can't have an IBIS file with just a package
  - Walter: We can have a stub [Component]
    - Would be need to use same [Component] with many packages
    - Mike L: [Pin] names refer to external pin names
    - Walter's proposal handles that mapping
- Walter: Crosstalk can be complicated when component models have interconnect

Arpad: so do we go for the short term or long term solution?
- Randy: We need a long term solution
- Radek: agree
- Bob: agree, short term solutions can be pursued by others
- We will focus on Walter's proposal
  - This will be reported to the open forum
  - No action required there
- Randy will continue work on the measurement BIRD
  - He may have something in 1 or 2 meetings

Next meeting: 25 Mar 2007 12:00pm PT

-----------