Minutes from the October 2, 2019 IBIS Interconnect Task group meeting are
attached.
Regards,
Justin
================================================================================
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 October 2, 2019 Meeting (* means attended at least using audio)
ANSYS Curtis Clark
Cadence Design Systems Bradley Brim
Intel Corp. Michael Mirmak*
Keysight Technologies Radek Biernacki
Mentor, A Siemens Business Arpad Muranyi*
Micron Technology Justin Butterfield*, Randy Wolff*
SiSoft Walter Katz*, Mike LaBonte
Teraspeed Labs Bob Ross*
Michael Mirmak convened the meeting. No patents were declared.
Justin Butterfield took minutes.
Review of Minutes:
- Michael called for review of the minutes from the September 25, 2019 meeting.
Randy Wolff moved to approve the minutes. Bob Ross seconded. The minutes
were approved without objection.
Review of ARs:
- Walter Katz to add the [End EMD Designator Map] keyword.
- Walter reported this is done.
- Bob to rework the rules in the designator section.
- Bob reported this is in progress, and we need to discuss these changes
further.
- Walter to send out EMD draft 21.
- Michael noted this was done.
- Randy to put together a stacked die model example for EMD.
- Randy reported he created an example to get started. Arpad Muranyi asked
if we
have an example that includes the power rails as well. Randy agreed this
would
be good idea to add later as a follow on.
Opens:
- None.
EMD Draft 22 Review:
Arpad noted he sent out some comments in relation to draft 21, but Bob had sent
out
a draft 22. He asked if his comments should be applied to draft 21 or 22. Bob
stated he made changes and sent out a draft 22, and suggested to start with
editing
draft 22. Arpad stated most of his comments were editorial and others were
seeking
clarification. Walter suggested to go through Bob's comments, then we can go
through the email comments from Arpad and others.
Bob shared draft 22. He made some editorial fixes to the hierarchy diagram and
clarifications to the keyword introduction section. Bob moved some of the
keyword
definition text. He also added bus_label in many places where signal_name is
referenced. Bob noted it is important to list all the IO pins in the
[Designator
Pin List]. Randy asked if this was necessary. Arpad asked what is the purpose
of
having the [Designator Pin List] and why the EDA tool cannot take this
information
from the model. He also asked about the naming differences between the
signal_names. Walter noted the editorial changes are okay, but we need to be
careful in the [Designator Pin List] section. A pin_name in the [Designator
Pin
List] must match exactly the pin_name in IBIS component model. Bob stated the
reason is the pin_name must be "plugged in" between EMD and component. Walter
noted
the signal_names in the designator are meaningless, as the pin_names define the
connectivity. The signal_names that are the same across the designator map the
connections. Bob commented we should be careful with the words "connected" and
"shorted". Walter noted connected does not mean shorted. If two pins have the
same
signal_name in the EMD, they are connected. Bob suggested to use bus_label for
these. Walter noted, if two rail connections have the same signal_name, they
are
connected, where if we use bus_label, there is a short and a single terminal
can be
used. Michael asked if we have definitions of shorted and connected in the
BIRD
draft or in IBIS. Arpad stated he had the same question. Randy noted we
removed
the definition of the Nyquist connection. Bob noted this does not apply to
rails.
Bob noted we need to be careful about what we consider a short vs. a
connection.
Bob asked how this is done in Randy's example where the signals are the only
way to
determine the signal paths. He noted Aggressor_only would not work in this
case,
since the signal_names are not the same. Walter disagreed that the
Aggressor_only
cannot be set for this as, the terminals are mapped with the pin_name only.
Michael
suggested to go through Randy's example before discussing this.
Bob suggested to be more consistent with our terminology such as pins vs.
pin_name.
And, we need to clarify when we are talking about I/Os or rails.
Bob noted in the example you might have an EMD model where pins may be
connected
based on the signal_names. Walter stated they are connected but not shorted.
Walter will go through Bob's changes and accept the changes he approves of and
send
out a draft 22 [AR]. He will try to go through Arpad's and Randy's comments,
and we
can discuss the other changes. Bob commented he made some corrections to the
examples. Bob was concerned about using the syntax of "U1.1,U1.2" to jump
across
multiple terminals, and he would like to raise this as an issue.
Stacked Die EMD Example:
Randy shared his stacked die EMD example. He noted that this example is a two
x8
die to create a x16 package, where the DQs only go to one die each and the
Address
and Command signals go to multiple die. He started from an EBD model with a U0
and
a U1 for the bare die. Walter stated we might consider moving the
[Manufacture] and
[Description] keywords. Bob stated he would oppose this change. Randy noted,
for
the [Designator Pin List], he used the signal_names out of the die IBIS file.
He
asked if these signal_names should be the top-level names. Walter stated this
is
correct. Bob commented this is what is meant by renaming the signal_names, but
the
pin_names should match.
Randy noted the example has an [EMD Group] with only one [EMD Set], and he
plans to
add a second Group for a fully coupled model. The [EMD Set] has [EMD Model]s
for
each of the signals. The DQs are single die connections. Walter asked if we
could
make the DQS model differential to show this capability. Randy replied this
could
be done. Walter asked if there was a separate VSS and VSSQ. Randy noted they
are
shorted in this case.
Next Meeting:
The next meeting will be October 9.
Arpad moved to adjourn. Randy seconded. The meeting adjourned without
objection.
================================================================================
Bin List:
EMD Comments to be Resolved:
1. Should the [Define Module] keyword be renamed? - RESOLVED
2. Documentation of CAD nets, extended nets and signal names definitions. -
RESOLVED
3. Add bus_labels as possible Terminal_type_qualifiers. - RESOLVED
4. Add [End EMD Pin List], [End Designator Pin List] to keyword hierarchy. -
RESOLVED
5. Remove [Number of EMD Pins] keyword? - CLOSED
6. Add definition of "Nyquist". - CLOSED
IBIS-ISS Parser:
- IBIS-ISS parser scope document