IBIS Macromodel Task Group

Meeting date: 07 April 2015

Members (asterisk for those attending):
Altera:                     * David Banas
ANSYS:                      * Dan Dvorscak
                            * Curtis Clark
Avago (LSI)                   Xingdong Dai
Cadence Design Systems:     * Ambrish Varma
                              Brad Brim
                              Kumar Keshavan
                              Ken Willis
eASIC                         David Banas
Ericsson:                     Anders Ekholm
IBM                           Steve Parker
Intel:                        Michael Mirmak
Keysight Technologies:      * Fangyi Rao
                            * Radek Biernacki
			    * Nicholas Tzou
Maxim Integrated Products:    Hassan Rafat
Mentor Graphics:            * John Angulo
                            * Arpad Muranyi
Micron Technology:          * Randy Wolff
                              Justin Butterfield
QLogic Corp.                  James Zhou
                              Andy Joy
eASIC                         Marc Kowalski
SiSoft:                     * Walter Katz
                              Todd Westerhoff
                            * Mike LaBonte
Synopsys                      Rita Horner
Teraspeed Consulting Group:   Scott McMorrow
Teraspeed Labs:             * Bob Ross

(Note: Agilent has changed to Keysight)

The meeting was led by Arpad Muranyi.

------------------------------------------------------------------------
Opens:

- None

--------------------------
Call for patent disclosure:

- None


-------------
Review of ARs:

- Michael M update AMI Directionality BIRD
  - No update

- Walter and Randy produce C_comp BIRD.
  - Arpad: Some work has been done and this has been posted.

- Arpad to review IBIS specification for min max issues.
  - In progress.


-------------
New Discussion:

Pin merging:

- Arpad: I had proposed some changes.
- Arpad showed an example for a proposed change.
- Arpad: Bob had proposed a new keyword instead of a subparameter.
  - This shows [Merged Pins], prepared by Randy.
- Randy: Line wrapping needs to be allowed.
- Arpad: In the second format the main pin follows the keyword.
  - Other connected pins are on the next line.

- Radek: How does this relate to IBIS line continuation rules?
- Bob: I prefer the two line format.
  - You can comment them that way.
- Walter: These are not consistent with the pin list above.
  - For example K9 is absent.
- Arpad: That is because there is no model for K9.
- Radek: But it must be on the component.
- Walter: These issues will go away with the new package model.

- Bob: This doesn't force using signal names for merged pins.
- Arpad: We found that [Pin Mapping] did not work the same way in each tool.
  - But this BIRD addresses just the ambiguity for pins that are not in [Pin Numbers].
  - The new package model capability will fully address it.

- John: This can create a set of shorted pins that is not the same set as [Pin Mapping].
- Arpad: It's the same capability.
- Bob: This can be used only with [Pin Mapping] present.
- John: You can have two buffers on the die side using the same group of shorted pins.
- Arpad: That is ruled out.
- Bob: We want to avoid shorting power and ground.
- Arpad: Near the die features become smaller.
  - It makes more sense to combine those.

PAM4 BIRD:

- Walter: The BIRD draft is up to revision 14 now.
  - I will send it with any changes to Mike for posting.
  - We should be able to forward it to the Open Forum with a recommendation.
  - It will not be ready for this Friday however.
- Arpad: We should review the changes.

- Walter showed the draft PAM4 BIRD.
- Walter: The Modulation parameter can be Info if there is only one value.
  - It has to be In to have multiple values.
- Bob: The Default must be <string_literal>.
- Radek: Why does the Default have to be here at all?
- Bob: That's how our rules work.
- Walter: For List format Default makes sense, but not for Value.

- Walter: PAM4_Mapping maps levels to symbols.
  - Grey code "0132" is the default.
- Bob: Again the Default is a <string_literal>.
  - Even though it is Value only.

- Walter: We have three thresholds for PAM4.
- Arpad: We do not assume center for the middle threshold?
- Walter: That is up to the model maker.
  - We allow InOut in case the model chooses it.
  - If not specified the EDA tool must determine thresholds.
- Arpad: Is it specified how to do that?
- Walter: It is up to the tool.
  - The thresholds really should be given.
  - There is also a PAM4_Receiver_Sensitivity, only for the middle threshold.
- Bob: The Default should be <numeric_literal>.
  - We should support Dep usage.
- Radek: Dep should be allowed for anything that can be Out.

- Walter: We have timing offsets for the upper and lower eyes.
- Bob: The Default must be described in the text.

- Arpad: Scott McMorrow was asking about non-linearity.
  - Fangyi had said AMI_GetWave could handle it.
  - That should be explained.
- Walter: It depends on what causes the non-linearity.
  - It will not happen in the RX analog section.
  - Voltage levels there are small, it will be linear.
  - In the TX there could be a voltage lookup table based on previous symbols.
  - It does not have to be a classic FFE.
  - Any non-linearity there probably is not about saturation.
- Arpad: Fangyi said it might be a saturation effect.
- Walter: Maybe only the impedance changes, but that's not voltage.
- Arpad: This assumes the impulse response is done with an NRZ level.
- Walter: It can be thought of as pulse or step too.

- Fangyi: Should we allow Out for PAM4_Mapping?
- Walter added a note for that.
  - We need to ask IC vendors about this.

- Fangyi: We need to discuss the bit_time argument in AMI_Init
- Walter: In PAM4 we have a symbol, not a bit.
  - For PAM4 it will be symbol time.
- Mike: We could rename it to symbol_time
- Arpad: We could say bit_time is scaled by symbols/bit.
- Ambrish: Other places in the IBIS specification might need to change for this.

AR: Walter send updated PAM4 BIRD to Mike for posting.

C_comp:

- Randy showed C_Comp Model BIRD draft revision 6.
- Randy: ISS has been changed to IBIS-ISS.
  - An example has been reformatted to insert a format column.
  - A figure has been updated.
- Radek: There is no A_gnd in there, we will need that.
- David: Is the numbering of names required?
- Randy: That is done to avoid conflicts.
  - The unnumbered nodes are at predefined locations.
- Bob: Corner has to be described as a model corner.

AR: Randy send updated C_comp BIRD to Mike for posting.

-------------
Next meeting: 14 Apr 2015 12:00pm PT
-------------

IBIS Interconnect SPICE Wish List:

1) Simulator directives
