Minutes from the February 2, 2022 IBIS Interconnect Task Group meeting are
attached.
Thanks,
Randy
================================================================================
IBIS INTERCONNECT TASK GROUP
http://www.ibis.org/interconnect_wip/ ;
Mailing list: ibis-interconnect@xxxxxxxxxxxxx
Archives at //www.freelists.org/archive/ibis-interconn/ ;
================================================================================
Attendees from February 2, 2022 Meeting (* means attended at least using audio)
ANSYS Curtis Clark
Intel Corp. Michael Mirmak*
Marvell Steve Parker*
MathWorks Walter Katz
Micron Technology Justin Butterfield
Randy Wolff*
Siemens EDA Arpad Muranyi*
ST Microelectronics Aurora Sanna
Teraspeed Labs Bob Ross*
University of Illinois Jose Schutt-Aine
Zuken USA Lance Wang*
Michael Mirmak convened the meeting. No patents were declared.
Randy Wolff took minutes.
Steve Parker had not attended in the past. He said Marvell is interested in
Touchstone 3 and saw the recent conversation about pole-residue formats.
Steve mentioned he did a presentation with Adge Hawes in 2017 on pole-residue
modeling at DesignCon. Steve is joining to be a part of the conversation.
Steve's paper can be found in the 2017 DesignCon proceedings. The paper title
is "Mind Your P's and R's: Pole-residue CTLE filtering in 56G PAM4 Model".
Review of Minutes:
- Michael called for review of the minutes from the January 26, 2022 meeting.
Michael displayed the minutes. Arpad moved to approve the minutes.
Randy seconded. The minutes were approved without objection.
Review of ARs:
- None
Opens:
- Arpad noted he sent out by email the summary of his conversations with
Vladimir about port reduction with pole-zero format. This is available for
review.
- Michael reported he was in contact recently with the IEEE 370 group. They
are interested in following along with our work on Touchstone. They may have
an officer or participants call in to the Interconnect task group. 370 covers
some interconnect measurement procedures. They port a lot into their
specification by way of Touchstone. Michael will make sure they are invited
to participate and share any formal presentations.
- Michael filed a TSIRD regarding the [End] keyword in Touchstone 2.0. About
1/3 of the examples in the specification are incorrect, missing the [End]
keyword along with other problems. Randy asked if Michael would like a review
in the Interconnect task group before official posting. Michael is adding the
TSIRD review to the bin list.
- Bob mentioned Walter shared a re-scaling proposal by email, and Bob had some
comments to share. Michael added it to the bin list for later discussion when
Walter was in attendance.
Discussion:
Designator Pin List Relaxation BIRD Draft Review:
Arpad shared Draft 7. He added text related to allowing NC as a signal_type
in the Designator Pin List. Arpad realized his earlier concern about the
addition of NC conflicting with the rule for I/O pins being allowed to have
only two columns (pin_name and signal_name) would not require any changes.
When we have a NC pin, it is not an I/O pin but will be a special case "NC",
like POWER or GND.
Arpad did not use the term "NC pins" in the BIRD, instead saying "designator
pins which have nothing connected to them from any of the EMD Models".
Randy commented that he felt a statement should be added to clarify why a NC
pin would be added, such as for complete documentation of pin lists. Arpad
thought this could be added after the sentence "Designator pin names which are
not referenced by any terminal line Qualifier_entry in any of the EMD Models
may optionally be listed under the [Designator Pin List] keyword."
Bob noted he did not want to imply a restriction that routing to a NC pin is
not allowed. Arpad thought if a pin was a named NC pin designed for some
future expansion, then it should not have the 3rd column as NC. The NC should
be removed from the 3rd column if a pin has routing and is intended to be
simulated. Bob thought NC has special meaning. Michael asked if NC has a
special meaning of "reserved" or NC means "not connected anywhere" versus "not
documented". Bob thought it could mean a reserved pin. Bob added that you
could have one side of an EMD board designated as NC that physically exists.
You could say that no routing is expected. Arpad said he wrote the specific
definition in the BIRD that if we don't have a single EMD model making a
connection to the pin, that's when it is a NC pin. Arpad said we don't have a
clear definition of what we mean by NC in IBIS. Bob said there are several
definitions. We should decide what it means for EMD.
Lance asked about the case of if we have routing, like a pin connected to
ground with routing, then we change the designator pin to NC, is this
allowed. Bob said we don't allow NC in the current specification. Arpad
noted the parser could check for conditions when the designator pins do have
something connected to them in the EMD models but are listed as NC. Michael
suggested Arpad send out the latest draft for review.
Touchstone 3 Survey Results:
Michael shared the survey results posted at:
https://ibis.org/interconnect_wip/2019%20IBIS%20Touchstone%20Survey%20Report.pdf
Michael noted the features most wanted in the next version of Touchstone were
port names, sparse matrices, binary format and reduced order models, in that
order of importance.
Bob noted that the IBIS committee has issued 8 TSCHK2 parser licenses. Bob
added that if adding port identification, that should be optional.
Michael noted there is a [Begin Information]/[End Information] section in
Touchstone 2.0 where port information could go. Bob suggested you'd really
want a dedicated new keyword. Michael said 370 may be interested in putting
their own keywords in that area.
Michael asked if there are other issues with Touchstone that need to be part
of Touchstone 3. Bob stated pole-residue is a good way to go but is branching
from Touchstone. There is no standard format, so do we make a standard that
all tools will accept as a standard from their private formats? Randy
suggested we need to see what formats from several vendors look like. There
may not be a clear starting place like we had with IBIS-ISS using HSPICE syntax.
Arpad noted there was a concern expressed by Jose Schutt-Aine about it being
difficult for model makers to create quality pole-residue models. Randy noted
we already have quality issues with S-parameters, so we still rely upon model
makers to create both a quality S-parameter and a quality pole-residue fit
model.
Michael asked EDA vendors to provide a format example from their tool if
allowed, but at least the requirements of what is needed for their tool [AR].
Bob noted Jose provided a format example that he created.
Lance asked Michael to send an email to the reflector asking for EDA tool
vendor feedback on the pole-residue format [AR].
Bob motioned to adjourn. Arpad seconded. The meeting adjourned.
Next Meeting:
The next meeting will be February 9.
================================================================================
Bin List:
1. Touchstone 3
2. Pole-zero support for Touchstone
3. New version of Touchstone to include 3 TSIRDs
4. Clarification BIRD drafts from ATM
5. Review Touchstone survey results
6. TSIRD(4)
7. Re-scaling proposal by Walter Katz
