Minutes from the 26 January ibis-atm meeting are attached.
IBIS Macromodel Task Group
Meeting date: 26 January 2021
Members (asterisk for those attending):
Achronix Semiconductor: Hansel Dsilva
ANSYS: Curtis Clark
Wei-hsing Huang
Cadence Design Systems: * Ambrish Varma
Ken Willis
Jared James
Google: Zhiping Yang
Intel: * Michael Mirmak
Kinger Cai
Alaeddin Aydiner
Keysight Technologies: * Fangyi Rao
* Radek Biernacki
Ming Yan
* Todd Bermensolo
Rui Yang
Luminous Computing: David Banas
Marvell: Steve Parker
Mentor, A Siemens Business: * Arpad Muranyi
Micron Technology: * Randy Wolff
* Justin Butterfield
SAE ITC Jose Godoy
SiSoft (Mathworks): * Walter Katz
Mike LaBonte
Teraspeed Labs: * Bob Ross
Zuken USA: Lance Wang
The meeting was led by Arpad Muranyi. Justin Butterfield took the minutes.
--------------------------------------------------------------------------------
Opens:
- None.
-------------
Review of ARs:
- Fangyi to email his "AMI Redriver Flow" presentation to the ATM list.
- Done. Arpad noted this was sent out with the corrections we discussed
last week.
- Arpad to ask Steve Parker to post the presentation to the ATM archives
- Done. Arpad noted Fangyi did this in his email, but it is not posted on
the website yet.
- Curtis to start an email thread about whether the BIRD204 modifications
can be handled in Editorial.
- Done. Arpad noted we will discuss this more today.
--------------------------
Call for patent disclosure:
- None.
-------------------------
Review of Meeting Minutes:
Arpad asked for any comments or corrections to the minutes of the January 19th
meeting. Michael moved to approve the minutes. Randy seconded the motion.
There were no objections.
-------------
New Discussion:
Potential issues with BIRD204:
Arpad stated, if the DLL requires clock ticks, the clock DLL needs to provide
them. The first proposal uses the word "shall". This is a technical change,
since the previous text stated "should", and this change would make it a
requirement. He asked if we can get away with making this as an editorial
change. The second proposal is to write additional statements advising the
model maker to return clock_times.
Arpad asked if we need to create a new BIRD for this. Bob suggested we need a
new BIRD for this. Randy asked if we have ever done a new version of an
approved BIRD and voted on it. Arpad thought we had not. Bob stated we
reopened a BIRD to make a change to it. Arpad asked if that was an approved
BIRD. Bob replied that it was approved, but he could not remember the BIRD
number. Arpad stated we have done the BIRD158 to BIRD194 change, where we
superseded an approved BIRD with a new BIRD. Randy thought, in most of those
cases, the changes were significant. He suggested we could create a BIRD204.1
with this minor change. Michael suggested an editorial BIRD that could
include these small changes. Bob suggested, going forward, we should
supersede a BIRD with a new one, that way we keep a clean record. He would
like to avoid having an errata type list for BIRDs.
Arpad asked if we are all in agreement that we should write a new BIRD. Randy
was in favor of this, although he would like to have it submitted to the Open
Forum by this Friday's meeting. Arpad will take an AR to write the new BIRD.
He suggested to follow the first proposal.
Radek asked if we should change "DLL model" to "executable model". Arpad
replied we have a to-do list item to resolve this throughout the entire
specification. Randy commented we also have a picture in this BIRD with "DLL
model" in it, and we will need to resolve this. Randy noted it is in the IBIS
7.0 known issues list.
Bob asked if this is a new requirement that the parser will have to check.
Arpad replied the parser cannot check this, as it does not check the DLL input
and output.
Redriver Flow Issues:
Arpad stated, in our previous meeting, we discussed the combined version of
Variant 1 and 2, which we called Variant 3. He thought we were mostly in
agreement to go with Variant 3. Arpad would like to make a decision on which
variant and find who would like to write the BIRD.
Michael asked if we want to have any discussion before the straw poll. Arpad
replied we should have the discussion before the poll. Michael asked if we
are assuming that all redriver models have to have both the Init and GetWave,
and the Init_Returns_Impulse flag is true. He also asked if we are excluding
the case where we have an Init only and GetWave only in the same simulation
flow. Arpad replied the proposal addresses the Init simulation flow only,
because the GetWave flow was not flawed. Michael asked if there would be
anything to prevent a case where the wrong data is presented and the
simulation case is impossible. Ambrish commented, if you have an Init only
model, it will work for both simulation types. A GetWave only model is where
you would have an issue. Ambrish thought this case is considered in the
slides, where a GetWave only model will not work in a statistical simulation.
He noted this was in his BIRD. We should make this obvious, but he was not
sure how to check this with the parser. Arpad stated we can go through all
the combinations when we write the BIRD. He suggested we should have a truth
table to show this in the BIRD.
Ambrish asked, on slide 6 of Fangyi's slides, if we can specify whats new for
the specification vs. what is already in there today. This would be helpful
for readers later. Arpad suggested we can mention this in the BIRD.
Straw poll:
Ambrish - Variant 1
Michael - Variant 3
Radek - Variant 3
Todd - Abstain
Arpad - Variant 3
Randy - Abstain
Justin - Abstain
Walter - Variant 3
Bob - Abstain
Curtis (by mail) - Variant 3
The results of the straw poll were:
Variant 1 - 1
Variant 2 - 0
Variant 3 - 5
Abstain - 4
Arpad asked Ambrish if he would be okay with Variant 3. Ambrish replied, yes,
but he prefers Variant 1.
Arpad asked who would be willing to draft a BIRD on this. Ambrish suggested
Fangyi might be the best person to write the BIRD. Fangyi stated he can start
to draft the BIRD.
Fangyi stated he had thought of a simplification to Variant 3. We have two
new reserved parameters in Variant 3. Possibly, we can simplify to one new
parameter, if we always require both the upstream cumulative impulse and the
filter place holder. Arpad asked if this makes it more complex for the model
makers. Fangyi thought it is not much more complicated to add both
parameters. Arpad suggested we could also have one new parameter with
multiple values. Fangyi stated his proposal would add only one new flow.
Arpad asked whether this simplification would impact performance. Fangyi
replied the changes would not significantly affect performance.
Arpad asked if model makers would be okay with this. Arpad asked if anyone
was opposed to this. Fangyi will add this new variant to his slides.
Ambrish asked Fangyi if he can add notes about which parameters are new vs.
existing to slide 6. Fangyi agreed to add this.
- Bob: Motion to adjourn.
- Radek: Second.
- Arpad: Thank you all for joining.
AR: Arpad to write a new BIRD to supersede BIRD204 with the proposed changes.
AR: Fangyi to begin a Redriver Flow BIRD draft.
AR: Fangyi to add additional variant and comments about which parameters are
new vs. existing to his slides.
-------------
Next meeting: 02 February 2021 12:00pm PT
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
