Meeting date: 02 may 2006
Members (asterisk for those attending):
*Arpad Muranyi,
*Bob Ross,
Todd Westerhoff,
*Mike LaBonte,
Paul Fernando,
Barry Katz,
Walter Katz,
Ken Willis,
*Ian Dodd
Lance Wang
Richard Ward
*Doug White
*Sanjeev Gupta
*Joe Abler

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

- Arpad compile email responses
  - done

- Mike investigate commercial solution
  - done

- Mike ask Cadence (Dovich) to join next call
  - done, can't make it this week or the next

- Richard check if he can share the receiver model
  - email response: "We plan on writing a clean room style DFE
    model. We'll have to place basic approximations for the non-linearities
    and impairments for the front-end."

- Todd, Ian, Richard meet this Wednesday
  - OK
-------------

Library issues:
- Arpad: What is the order of events when a siulator begins?
  - Affects usage of break statements, which can cause
    1st derivative discontinuities when not used correctly.
  - Break statements are used in the PWL and divider elements.
  - DC operating conditions are sometimes wrong.
  - Hard to do this right in VHDL-AMS:
    - Assignments are made last at each simulation event.
    - Analog solvers are called in the middle, get stale data.
    - Arpad si still working on this.
- Simple cut-and-paste typo was found in the macro library.

AR: Mike send begin/end marker usage to Arpad

Phase 1 vs. Phase 2 Library Enhancements
- We are entering a new phase to support models beyond pre-emphasis
  drivers.
- How to make plans for coding this?
  - Arpad looking for someone else to share the work.
  - Maybe TI?

Recap of our history for new member Joe Abler:
- IBIS running out of steam for very fast buffers.
- Proposal: Berkely SPICE extensions
  - Not very robust.
- Counterproposal: add HSPICE-like special sources and B element.
- Analog portion of AMS languages would be good:
  - Write macros that correspond to these elements:
  - Top level netlist only calls element modules:
    - Netlists look about the same in any language.
- Joe has seen Richard Ward's presentation.
- Now we find that Matlab is widely used for advanced modeling
  - It has user-friendly functions.
  - If our macro library had a PLL that would be useful.

Ideas for new modules to build:
- PLL
- FIR filter
- Gary Pratt's models
- IEEE proposals (see below)

AR: Ian ask Gary Pratt for public domain models

AR: Joe send Mike info on IEEE proposals for minutes

Done, Joe sent the following:
- At 10Gbps, IEEE802.3ap is defining a 3tap FFE transmitter with specific
  requirements for tap settings, output waveforms, etc.  They are also
  defining an informative channel model.
- At 6Gbps & 11Gbps OIF CEI has defined reference configurations for
  serdes and a compliant channel model.  The serdes reference configurations
  are defined in StatEye scripts and channel compliance is determined by
  simulating a specific channel against the reference scripts.  Info on
  StatEye can be obtained at stateye.org
- Other standards in process which may define specific advanced
  equalization requirements or channel characteristics are 8.5Gbps FCS
  and 6Gbps SAS.  It's still a bit early to see the specific direction
  with these.

-------------
Next meeting: Tuesday 09 may 2006 12:00pm PT