IBIS Macromodel Task Group

Meeting date: 16 February 2016

Members (asterisk for those attending):
ANSYS:                      * Dan Dvorscak
                            * Curtis Clark
Broadcom (Avago):             Xingdong Dai
                              Bob Miller
Cadence Design Systems:     * Ambrish Varma
                              Brad Brim
                              Kumar Keshavan
                              Ken Willis
Cisco:                        Seungyong (Brian) Baek
eASIC:                        David Banas
                              Marc Kowalski
Ericsson:                     Anders Ekholm
GlobalFoundries:              Steve Parker
Intel:                      * Michael Mirmak
Keysight Technologies:      * Fangyi Rao
                            * Radek Biernacki
Maxim Integrated Products:    Hassan Rafat
Mentor Graphics:              John Angulo
                            * Arpad Muranyi
Micron Technology:          * Randy Wolff
                              Justin Butterfield
QLogic Corp.:                 James Zhou
                              Andy Joy
SiSoft:                     * Walter Katz
                              Todd Westerhoff
                            * Mike LaBonte
Synopsys:                     Rita Horner
Teraspeed Consulting Group:   Scott McMorrow
Teraspeed Labs:             * Bob Ross
TI:                           Alfred Chong


The meeting was led by Arpad Muranyi.

--------------------------------------------------------------------------------
Opens:

- None.

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

- None.

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

- Michael M. to send out a kickoff email for editorial task group meetings
  starting on March 4th.
  - Done.
  
- Arpad to review the spec and find all instances of language stating that the
  typ value should serve as the default for missing min max values.
  - Not yet done.

-------------------------
Review of Meeting Minutes:

- Arpad: Does anyone have any comments or corrections? [none]
- Arpad: Motion to approve the minutes.
- Michael M.: Second.
- Arpad: Anyone opposed?  [none]

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

New Redriver Flow BIRD:
- Fangyi: Recent discussions regarding improving the AMI redriver flow lead to
          some new ideas for improving the normal AMI flows as well.
  - Vladimir had a suggestion that I liked a lot, so I decided to delve into
    that possibility and work out this proposal for an improved flow.
- Fangyi: [Sharing his presentation "AMI Simulation Flow Round 3"]
  - Motivations:
    - Handle Init() only Rx properly for time domain and statistical flows and
      for normal and redriver scenarios.
    - Enhance Rx AMI Init() optimization by passing it the full channel IR.
      This is essential for Init() only models, but could also be useful for
      GetWave() models.
    - Eliminate the need for deconvolution in any flow.
  - Summary:
    - No change to the Tx Init().
    - Rx Init() function's IR matrix gets two new columns.
      - One is used to provide the model with the total IR of the channel (from
        the terminal (initial) Tx up to the Rx in question).
      - The second column is used for the model to return the DFE impulse
        (Vladimir's idea).
    - Rx, in partial IR - Legacy IR, "partial" because it only represents the
      most recent upstream channel.
    - Rx, in total - New full channel IR up to the Rx.
    - Rx, in DFE - Placeholder for model to use to return Rx DFE Impulse.
    - Rx, in xtalk - Legacy IRs from aggressors to the Rx.
- Discussion: Fangyi reviewed the entire presentation.  It details the various
  flows in combination with normal or repeater scenarios, how the new IR
  matrix columns are used, and how the Rx model may apply its DFE portion, its
  non-DFE portion, or both to the various IRs it returns.  Fangyi mentioned that
  one reason to use the full impulse instead of just the tap coefficients for
  DFE is that you can broaden the DFE spikes to a certain degree to model the
  slicer switching, and having the full impulse column allows you to do that.
  Mike/Radek/Fangyi noted that this new functionality would require changes in
  the tool and the model, and might be controlled/indicated by the presence of a
  new Reserved keyword.  If both the simulator and model indicated that they
  supported the new flow, then it could proceed.  Otherwise the old flow would
  be the fallback.
- Discussion: Arpad asked about the logistics of this proposal becoming a BIRD.
  Fangyi said he expected Vladimir, Walter, Arpad and others might be co-
  authors.  Fangyi said he would prefer another few iterations of discussion
  before drafting a formal BIRD.  Walter said he felt Fangyi's proposal would be
  a new BIRD independent of his own.  He said this his original proposal
  addressed the Rx Init() only repeater flow problem without the addition of new
  columns, and Fangyi had presented a totally different approach.  

- Arpad: Thank you all for joining.
-------------
Next meeting: 23 February 2016 12:00pm PT
-------------

IBIS Interconnect SPICE Wish List:

1) Simulator directives