Minutes, IBIS Quality Task Group

13 August 2019

11:00-12:00 EST (08:00-09:00 PST)

ROLL CALL

Intel Corp.                         Michael Mirmak
Mentor, a Siemens Business          Arpad Muranyi
Signal Integrity Software:        * Mike LaBonte
SPISim                              Wei-hsing Huang
Teraspeed Labs:                   * Bob Ross
Zuken USA:                        * Lance Wang

Everyone in attendance marked by *

NOTE: "AR" = Action Required.

-----------------------MINUTES ---------------------------
Mike LaBonte conducted the meeting.

Call for IBIS related patent disclosures:

- None

Call for opens:

- None

Review of previous meeting minutes:
Minutes from the August 06 meeting were reviewed.  Lance Wang moved to approve the minutes.
Bob Ross seconded.  Without objection the minutes were approved.


ARs:
- None


IBISCHK700 parser development:
Bob Ross had provided additional feedback to the developer.  An issue regarding
Aggressor_Only is scheduled to be discussed in the ATM Task Group.  There was confusion
about overlap rules, involving the word "can", which could mean the error condition
"can" appear, or that it is allowed and not an error.

Mike LaBonte showed Aggressor_Only2docx, from Bob Ross.  Bob pointed out that in
example 1, only 1 terminal had Aggressor_Only.  Mike said the Interconnect Task Group
had repeatedly discussed that, and it had been agreed that only one Aggressor_Only was
needed to consider the entire path as Aggressor_Only.  Bob read a sentence from page
306 of IBIS ver7.0.  Mike agreed that the word "identifier" was used in two different
contexts in the same sentence, and that was confusing.

Bob described a situation in which two Interconnect Models in the same Interconnect
Model Set, having pin A1 on the same pair of interfaces, neither having Aggressor_Only,
would not be allowed.  Mike questioned why that would be the case. He described a
case where a single-line pin A1 model and a coupled pin A1 were both in a set. If
the user requested an uncoupled simulation, the uncoupled model would be chosen by
the tool. For an uncoupled simulation the single-line model would be chosen. He said
tools would implement a scoring system to choose the best model for any given simulation.

We looked at the rule on IBIS page 308. Mike agreed that Bob's conclusion that IBIS
prohibited having multiple models in the same Interconnect Model Group, in which a
pin appears in more than one model WITHOUT Aggressor_Only, seemed fair given the text.

Bob asked if the two bottom examples in Fig 47 could exist in the same set.  Mike said
they could, the tool could easily choose which one to use because they have different
victim pins, as denoted by the absence of Aggressor_Only.

Bob read from IBIS page 33. The second bullet on the page prohibited the same pin
appearing twice in a group unless one is not Aggressor_Only.  Mike said that would
allow the tool to identify exactly one model for a particular victim.


IBISCHK7 source code license:
Bob Ross reported no change from last week. We have three signed agreements. One
has paid.  It is under legal review at one company, at least.  Bob described a problem
scenario for companies accepting outside software, to explain the need for legal review.


New IBISCHK bugs:
No new bugs.


IBIS-ISS parser:
Mike LaBonte noted that there had been discussion about an IBIS-ISS parser in the
previous ATM Task Group meeting.  The group had voted to request quotes for a level
1 and level 4 IBIS-ISS standalone parser.  Bob Ross said he would wait until after
IBISCHK700 is accepted before requesting the quotes.  He said we will need to write
descriptions for the levels.

AR: Mike LaBonte to write descriptions of IBIS-ISS parser levels for quotes


Bob Ross moved to adjourn. Lance Wang seconded. Without objection the meeting ended.

Meeting ended: 11:59 ET

Next meeting August 20, 2019


