IBIS Macromodel Task Group

Meeting date: 31 October 2023

Members (asterisk for those attending):
Achronix Semiconductor:       Hansel Dsilva
Amazon:                       John Yan
ANSYS:                      * Curtis Clark
                              Wei-hsing Huang
Aurora System:                Dian Yang
Cadence Design Systems:     * Ambrish Varma
                              Jared James
Google:                       Hanfeng Wang
                              GaWon Kim
Intel:                      * Michael Mirmak
                            * Kinger Cai
                              Chi-te Chen
                              Liwei Zhao
                              Alaeddin Aydiner
Keysight Technologies:        Fangyi Rao
                              Majid Ahadi Dolatsara
                              Stephen Slater
                              Ming Yan
                              Rui Yang
Marvell:                      Steve Parker
Mathworks (SiSoft):         * Walter Katz
                              Graham Kus
Micron Technology:            Justin Butterfield
Missouri S&T:                 Chulsoon Hwang
                              Yifan Ding
                              Zhiping Yang
Rivos:                        Yansheng Wang
SAE ITC:                      Michael McNair
Siemens EDA (Mentor):       * Arpad Muranyi
                            * Randy Wolff
Teraspeed Labs:             * Bob Ross
Zuken USA:                    Lance Wang

The meeting was led by Arpad Muranyi.  Curtis Clark took the minutes.

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

- None.
  
-------------
Review of ARs:

Michael: Send out draft14 of the AMI Test Data proposal including more
         feedback from the ATM meeting.
         - Done.
         
Michael: Develop a full syntactically complete example demonstrating the
         [AMI Test Data] proposal.
         - In progress.

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

- None.

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

Arpad asked for any comments or corrections to the minutes of the October 24th
meeting.  Curtis moved to approve the minutes.  Randy seconded the motion.
There were no objections.

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

BIRD220 and BIRD226:
Arpad reported that at the most recent IBIS Open Forum meeting Randy had asked
that we renew our focus on getting people to review these so we can vote on
them.  BIRD226, "PSIJ Sensitivity" was submitted on August 8, 2023.  BIRD220,
"Pre-driver PSIJ Sensitivity Keyword" was submitted over a year ago, on October
20, 2022.

Kinger said BIRD220 does not capture all jitter effects.  Randy agreed that
BIRD220 is only for the final pre-driver stage.  It captures the effects of
pulling power through the package to supply the final pre-driver stage.  He
said BIRD220 allows a b-element transient simulation to get better edge rate
alignment with a full SPICE simulation when the power supply is non-ideal.
Arpad agreed that it is an enhancement to the existing ISSO and Composite
Current keywords.

Randy said one question was whether there might be any interaction between
jitter included in the IBIS-AMI flows vs. a regular IBIS model.  Arpad said he
thought BIRD220's new keyword says how much an edge is delayed given power
supply changes.  When we do channel characterization for AMI, it is a single
rising edge.  There's no difference whether the delay is in the channel or due
to PSIJ.  It doesn't really affect the slope or amplitude of the edge, so it
won't affect the AMI simulation.  Randy wondered whether the multiple-edge AMI
flows used by some tools for single-ended AMI scenarios might see an effect
from BIRD220.  If, for example, power supply noise to the Tx was modeled during
the computation of the multiple edges, the edges might see different variations
in timing.  Randy said he wasn't sure that was something anyone would try to do,
but we might want to include some language addressing the possibility.

Kinger said he thought BIRD220 and BIRD226 were complementary and could co-exist
without fear of double counting.  He said BIRD226 envisions a two-step process.
First an SI simulation is run under ideal power conditions, then the effects of
PSIJ are added in as a post-processing step.  Since BIRD220 would have no effect
on the ideal power supply SI simulation, it would not cause double counting of
jitter effects in a BIRD226 type simulation.

Kinger took an AR to prepare some slides to review BIRD220 and BIRD226 and how
they might co-exist.

AMI Ignore Block Feature:
Arpad said that Alaeddin had sent out draft2 incorporating the feedback from the
ATM group.  Arpad said that he thought it was ready to submit to the Open Forum.
He asked for questions and comments.  Walter agreed with Arpad.  Walter moved
that we recommend Alaeddin submit his proposal to the Open Forum.  Michael
seconded.  There were no objections.  Michael took an AR to ask Alaeddin to
submit his proposal to Lance (IBIS Chair) for posting as an official BIRD.

AMI Test Data proposal:
Michael reviewed draft14.  He said it now contains a single keyword [AMI Test
Configuration] scoped by [Algorithmic Model].  He said a new Executable_index
subparameter refers to the specific configuration under that Algorithmic Model
that was used to generate the data included in the [AMI Test Configuration].

Ambrish noted that page 8 mentions the "Executable" line, and he asked whether
we should also mention Executable_Tx and Executable_Rx.  Michael agreed that 
we should mention all of them.  Arpad asked whether the Executable_index
subparameter considers the direction.  Michael said it is a single index into
all of the configuration lines of the [Algorithmic Model].  Michael said it
might enable a simple parser check to make sure the Direction subparameter of
[AMI Test Configuration] matches the direction of the configuration referred to
by the Executable_index.

Michael said the proposal now states that a Type Time_domain simulation with a
Golden_waveform comparison may be specified for an AMI model that does not
contain AMI_GetWave.  In this case, the tool would provide a pseudo-GetWave
by convolving with the modified impulse response returned by the model.

Bob said he would provide Michael with a revised hierarchy tree diagram.
Michael said he would send out draft15 incorporating the changes discussed.

- Curtis: Motion to adjourn.
- Michael: Second.
- Arpad: Thank you all for joining.

New ARs:

Kinger: Prepare some slides to aide in the discussion of BIRD220 and BIRD226 and
        their interaction
        
Michael: Ask Alaeddin to submit his Ignore Block proposal to the Open Forum as
         an official BIRD

Michael: Send out draft15 of the AMI Test Data proposal including more feedback
         from the ATM meeting

-------------
Next meeting: 07 November 2023 12:00pm PT
-------------

IBIS Interconnect SPICE Wish List:

1) Simulator directives
