Minutes from the 31 May ibis-atm meeting are attached.
IBIS Macromodel Task Group
Meeting date: 31 May 2022
Members (asterisk for those attending):
Achronix Semiconductor: Hansel Dsilva
Amazon: John Yan
ANSYS: * Curtis Clark
* Wei-hsing Huang
Cadence Design Systems: * Ambrish Varma
* Jared James
Google: Zhiping Yang
Intel: Michael Mirmak
* Kinger Cai
Alaeddin Aydiner
Keysight Technologies: Fangyi Rao
Majid Ahadi Dolatsara
Ming Yan
Radek Biernacki
Rui Yang
Luminous Computing David Banas
Marvell Steve Parker
Mathworks (SiSoft): * Walter Katz
Mike LaBonte
Micron Technology: * Randy Wolff
Justin Butterfield
Missouri S&T Chulsoon Hwang
SAE ITC Michael McNair
Siemens EDA (Mentor): * Arpad Muranyi
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:
- Arpad to send BIRD213.1 draft 26 to the ATM list.
- Done.
--------------------------
Call for patent disclosure:
- None.
-------------------------
Review of Meeting Minutes:
Arpad asked for any comments or corrections to the minutes of the May 24th
meeting. Randy moved to approve the minutes. Walter seconded the motion.
There were no objections.
-------------
New Discussion:
PAMn BIRD 213.1 draft 26 (now 27b):
Arpad recapped the history of the document since the previous meeting. Arpad
had sent out draft 26 per his AR. Fangyi had responded with a draft 27 further
clarifying the cases. Arpad then observed that the precedence rules listed for
the statistical flow were confusing (because of a missing sub-heading for the
case when neither Rx_Decision_Time nor Rx_Clock_Recovery_Mean are specified).
Fangyi had suggested adding the missing sub-heading, but he had then observed
that the precedence rule sections were essentially copies of what was contained
in the enumerated cases. Arpad had prepared a new version that removed the
precedence rules section (essentially creating a 27b).
Randy and Walter said they agreed with what had been done to create draft 27b.
Ambrish agreed too, but he again suggested we replace existing instances of
"actual_time" in the Rx_Clock_Recovery_XXX and Rx Jitter parameters with
the new "nominal_sample_time". Arpad recalled that some had suggested this
would best be handled by a new BIRD. This meeting, however, the group reviewed
the suggestion. Ten instances of "actual_time" were identified, and the group
agreed that all ten could be changed to "nominal_sample_time". Arpad added
language stating that all ten instances should be replaced.
Randy moved to submit this version to the Open Forum as BIRD213.1. Walter
seconded. There were no objections. Randy noted that he would adjust the
indentation of the case statements prior to submitting it.
Upcoming PI presentation:
Kinger said that at the next meeting he would have a presentation on supporting
PI modeling in IBIS. He said IBIS provides chip level models to facilitate
platform SI design, and the goal of a new PI modeling proposal is to provide
chip level models to facilitate platform PI design. He said they are working on
a syntax proposal inspired by the .pkg format in IBIS. They are considering the
extension .spi (for standard power model). He said several EDA vendors and end
user companies are already supporting and testing the new format. The goal will
be to present the proposal in ATM and the Open Forum and get feedback from
stakeholders.
Multi-level analog modeling:
Arpad recalled the multi-level analog I/V curve model discussion he had recently
started. He said he had received feedback from several people who supported the
idea of extending standard IBIS models to address PAMn devices. He said his
current plan was to follow some of the concepts from a joint presentation he and
Randy had done on a GDDR6X investigation. He said that approach had used
Verilog-A models to implement the multi-level driver. He said the goals is to
use multiple I/V and V/T tables with some state-machine logic based on the bit
pattern. He's considering a similar concept to the Verilog-A implementation,
except that new IBIS keywords would be added so that EDA tools could support the
flow natively and not rely on a multi-lingual model.
- Ambrish: Motion to adjourn.
- Curtis: Second.
- Arpad: Thank you all for joining.
AR: Arpad to send today's draft of BIRD213.1 to Randy.
AR: Randy to submit today's draft to the Open Forum as BIRD213.1.
-------------
Next meeting: 07 June 2022 12:00pm PT
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
