                                             
IBIS Macromodel Task Group

Meeting date: 12 Jul 2011

Members (asterisk for those attending):
Agilent:                    * Fangyi Rao
                            * Radek Biernacki
Ansys:                        Samuel Mertens
                            * Dan Dvorscak
                            * Curtis Clark
Arrow Electronics:            Ian Dodd
Cadence Design Systems:       Terry Jernberg
                            * Ambrish Varma
Celsionix:                    Kellee Crisafulli
Cisco Systems:              * Mike LaBonte
                              Ashwin Vasudevan
			      Syed Huq
Ericsson:                     Anders Ekholm
IBM:                          Greg Edlund
Intel:                        Michael Mirmak
LSI Logic:                    Wenyi Jin
Mentor Graphics:            * John Angulo
                              Zhen Mu
                            * Arpad Muranyi
Micron Technology:            Randy Wolff
NetLogic Microsystems:        Ryan Couts
Nokia-Siemens Networks:     * Eckhard Lenski
Sigrity:                      Brad Brim
                            * Kumar Keshavan
                            * Ken Willis
SiSoft:                     * Walter Katz
                            * Todd Westerhoff
			      Doug Burns
Snowbush IP:                * Marcus Van Ierssel
ST Micro:                     Syed Sadeghi
Teraspeed Consulting Group:   Scott McMorrow
                            * Bob Ross
TI:                           Casey Morrison
                              Alfred Chong
Vitesse Semiconductor:        Eric Sweetman
Xilinx:                       Mustansir Fanaswalla

The meeting was lead by Arpad Muranyi

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

- Walter will not join us next week

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

- None

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

- Arpad submit Correcting the rules for AMI_Close BIRD draft to open forum
  - Done

- All check task list to verify completion of items
  - Done

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

Arpad: Ken had sent documents to Mike for posting:
- Mike: These are not posted yet

Arpad showed a backchannel presentation from Ken:
- Ken: This is an update of a previous presentation
- Slide 4:
  - AMI_parameters_out will be used
- Slide 5:
  - New Training param (on/off)
  - New Backchannel_Protocol param
    - Points to a .bci file
- Slide 6 & 7:
  - Examples of new params
- Slide 8:
  - New Reserved params go in .bci file
  - .bci uses same format as .ami
- Slide 9 & 10:
  - Example of .bci file
- Slide 11:
  - This is about defining bit patterns
  - Walter: Extra quote error on the LFSR line
    - It is done as Table to keep them in sequence
  - Fangyi: This tells when it is in training mode?
  - Walter: It goes through preamble, post-amble states, etc.
  - Fangyi: The EDA tool will have to parse the .bci file
  - Walter: Yes, except the Model_Specific params are passed through to the DLL
  - Ambrish: These all use standard patterns?
  - Ken: The model knows the patterns it should get
    - The way it is done is generic
    - Multiple protocols can be handled
  - Walter: The standards give pre and post-ambles
  - Kumar: 802.3 defines a frame very specifically
    - The different standards use different terminology
  - Ken: TX and RX and will get the same .bci file
- Slide 13:
  - This is a training flow diagram
  - SI OK = Training_Done
  - Fangyi: Why turn off training?
  - Kumar: Sometimes you want to see what happens with no training
  - Fangyi: The user can set Training = Off
  - Arpad: Does the messaging go through the tool or is it direct?
  - Kumar: Through the EDA tool
  - Marcus: The models have pointers to the shared data
  - Todd: How is training turned off?
  - Ken: There is a parameter "Training" in the .ami file
  USAGE - Arpad: It is Usage In so it goes to the model
  - Walter: But the EDA tool sees it too
  - Marcus: The TX does not need to know this
  - Walter: The TX needs to know if it should generate the needed data
  - Fangyi: If a model doesn't support training it sets Training off?
  - Walter: It would not have the parameter
  - Radek: Both TX and RX must have training and it must be on
  - Arpad: This is a dual use file
    - Do we need a new Usage type?
  - Walter: It is Reserved and we can specify in text how it works
  - Todd: We can't put Training in a 5.0 model
  - Fangyi: How does the model know training is done?
  - Ken: Training_Done is set by the RX
  - Marcus: The tool also may want to have a timeout
- Slide 14:
  - Walter: There is an example flow 
  - Marcus: How does the tool tell the models it is terminating training?
  - Walter: Either RX or TX can say training is done
  - Kumar: There could be a Max_Training_Time param
  - Marcus: The RX model should behave the same with training on or off
  - Fangyi: Some may not want the performance trade-off
  - Mike: We are implementing a shared state variable using messaging
    - Sometimes that doesn't work well
  - Marcus: Maybe the RX should be in charge
  - Radek: The EDA tool has to know when training is off

Fangyi: Both models need the .bci file?
- Walter: The model only needs to know which protocol is selected
  - They are written sometimes to handle multiple protocols
- Kumar: It is good practice to read the files
- Fangyi: The .bci is provided by the model vendor?
- Walter: A higher authority might be needed
  - I don't think the standards organizations want to do this
- Ambrish: IBIS could do it
- Walter: We own the Reserved_Parameters section
  - We can decide from the Model_Specific implementation what to standardize
  - The vendors must work with us on this
- Ken: They should be posted on the IBIS web site

Mike: If the .bci files can have Model_Specific sections each model needs it's own .bci
- Fangyi: There should be no Model_Specific in .bci
- Marcus: It should be in the .ami file
- Walter: We could change it to Protocol_Specific
- Fangyi: Then the EDA tool will have to do more
  - The .ami has Model_Specific
- Walter: That would not be used
  - The EDA tool needs to merely pass the Protocol_Specific section
- Fangyi: How about Training_Done?
- Walter: That is in Reserved
- Ken: I like Protocol_Specific
- Walter: The .bci file is a convenience
  - The protocol is really hardcoded into tools and models
- Ambrish: We should check with standards groups on this

Arpad showed the text version of the BIRD:
- Arpad: Do we need more discussion?
- Ken: The authors need to iron out some details
  - The draft will be revised
- Walter: It should be .doc, not .txt
  - IBIS is going in that direction anyway
  - It will be easier to take it to other committees
- Bob: We will need a list of Reserved params allowed in .ami and .bci
  - And which are not allowed

AR: Marcus, Ken and Walter update Backchannel BIRD draft

Meeting ended.

-------------
Next meeting: 19 July 2011 12:00pm PT

Next agenda:
1) Task list item discussions

-------------
IBIS Interconnect SPICE Wish List:

1) Simulator directives
