Meeting date: 09 may 2006
Members (asterisk for those attending):
Arpad Muranyi, Intel Corp.
*Bob Ross, Teraspeed Consulting Group
*Todd Westerhoff, Cisco Systems
*Mike LaBonte, Cisco Systems
Paul Fernando, NCSU
Barry Katz, SiSoft
Walter Katz, SiSoft
Ken Willis, Cadence Design Systems
Ian Dodd, Mentor Graphics
Lance Wang, Cadence Design Systems
*Richard Ward, Texas Instruments
Doug White, Cisco Systems
Sanjeev Gupta, Agilent
Joe Abler, IBM

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

- Mike send begin/end marker usage to Arpad
  - Done.

- Ian ask Gary Pratt for public domain models
  - Done, Gary asked to join this meeting.

- Joe send Mike info on IEEE proposals for minutes
  - Done, included in minutes.

-------------

TI:
- TI must deliver models that have sufficient:
  - performance
  - accuracy
    - like to run million bit simulations.
- Displaying data for receiver models:
  - Receiver model has a quality-of-service indicator that must be
    watched, for example.
  - AMS tools have ways to monitor and display anything.
  - In classic SI tools the measurements are somewhat fixed.
    - Waveforms may be displayed, but no analysis.
- Compiled Matlab model is license free.
  - Released under NDA to somewhat protect behavior details.
- Getting good results from Verilog

Encryption:
- Bob contacted John Shields to talk about encryption
- Mike contacted Steve Dovich
  - Steve is away for a few weeks, but may be able to join us
    at some point.

Meeting of Richard, Gary, Ian, and Todd in SJ:
- Contention about what VHDL-AMS can/can't do.
- Having IBIS wrappers around models is helpful.
- Discussion about what is not being modeled in transmitters.
- Making transmitters more accurate, but need to work on
  receivers.
- Reflections
- Gary asked Richard to perform a sweep of drive characteristics.
  - Probably need to run some patterns.
- Gary contends a digital/analog language is required.
  - Richard has not found this yet.

Moving to phase 2:
- People needed:
  - Describe the model requirements - semiconductor companies
  - Modeling implementor - EDA

Matlab as a platform:
- Connecting IBM model to TI model:
  - Can be done with compiled Matlab models
  - TI algorithm yields large arrays to implement the transfer functions.
  - Matlab code incorporates the simulator
  - Key problem is defining the interface
- Competitors: MathCAD? Mathematica? Maple?
- IBIS has hooks for external languages
  - Could we add [External Circuit] Matlab?
  - EDA companies would define the interface
- Compiled Matlab:
  - License to compile, not to run
  - Can any vendor use the DLLs?
  - Cadence working to integrate compiled Matlab
- How to make compiled models interoperate:
  - Port interface for models must be standardized.
  - Three layer simulation system:
          - Compiled models
          - Simulation "backplane" (written in Matlab)
	  - Matlab
- Google search: interfacing matlab dlls

-------------
Next meeting: Tuesday 16 may 2006 12:00pm PT