IBIS Macromodel Task Group

Meeting date: 02 Oct 2007

Members (asterisk for those attending):
* Ambrish Varma, Cadence Design Systems
* Arpad Muranyi, Mentor Graphics Corp.
  Barry Katz, SiSoft
* Bob Ross, Teraspeed Consulting Group
  David Banas, Xilinx
  Donald Telian, consultant
  Doug White, Cisco Systems
  Essaid Bensoudane, ST Microelectronics
  Ganesh Narayanaswamy, ST Micro
  Hemant Shah, Cadence Design Systems
  Ian Dodd
  Joe Abler, IBM
* John Angulo, Mentor Graphics
  John Shields, Mentor Graphics
  Ken Willis, Cadence Design Systems
* Kumar, Cadence Design Systems
  Lance Wang, Cadence Design Systems
  Luis Boluna, Cisco
* Michael Mirmak, Intel Corp.
* Mike LaBonte, Cisco Systems
  Mike Steinberger, SiSoft
  Patrick O'Halloran, Tiburon Design Automation
  Paul Fernando, NCSU
  Radek Biernacki, Agilent (EESof)
  Randy Wolff, Micron Technology
  Ray Comeau, Cadence Design Systems
  Richard Ward, Texas Instruments
  Sanjeev Gupta, Agilent
  Shangli Wu, Cadence
  Stephen Scearce, Cisco Systems
  Syed Huq, Cisco Systems
  Syed Sadeghi, ST Micro
* Terry Jernberg, Cadence Design Systems
  Todd Westerhoff, SiSoft
  Vikas Gupta, Xilinx
  Vuk Borich, Agilent
* Walter Katz, SiSoft

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

No one declared a patent.

-----
Opens:

- Bob and Arpad brought up availability of model DLLs for Windows.
- SiSoft has provided Windows DLLs already in their toolkit.
- Cadence has provided only a Linux version in their toolkit.
  - Cadence will provide a Windows version in a few days.
- Bob said someone from Huawei was interested in reviewing the BIRD.
  - He would like to be added to our mailing list.
  - Bob will send him the latest version of the draft BIRD.
- Bob asked about speeds that the models work at.
  - SiSoft's model is configured for 5GHz.
  - Cadence's model is configured for 6.25GHz.
  - They will run at any speed.
  - Arpad suggested using the same default speed for both models.

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

- SiSoft will propose a new reserved_parameter keyword table
  - Done

- Write parameter passing syntax proposal for a possible BIRD
  a) Propose a parameter passing syntax for AMI models
  b) IBIS spec should allow parameter value assignment for
     *-AMS using a syntax that is consistent with the
     parameter passing syntax of the AMI models
  c) Allow parameter passing for SPICE [External ...] also?
     - Berkeley SPICE does not handle parameters at all.
     - Walter proposed handling this with a preprocessor

- Ian: send relevant papers on related IEEE work
  - TBD

- Doug will attempt to get Steve Pytel (Ansoft) and/or 
  Rich Mellitz to discuss upcoming paper regarding LTI
  assumptions wrt Serdes
  - TBD

- Arpad: Review the new macro library files.
  - TBD

- Radek write new proposal for Platform_Compiler_Bits

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

Ambrish presented the latest AMI_BIRD_DRAFT_2.1:
- Page 5:
- Ambrish asked if the description of [Algorithmic Model] could merely
  reference [External Model].
  - We decided against this.
- Michael Mirmak suggested changing "used with [Pins]" to "used with pins".
- If an [Algorithmic Model] is used with a [Model] that is used by a [Diff Pin]
  pin, the spec takes no position on whether the AMI model needs to be
  instantiated once or twice.
  - For pseudo-differential IBIS, models are usually instantiated twice.
  - Can the EDA tool figure it out?
  - Walter suggested that it is instantiated only once.
    - [External Model] requires that the [Model] assigned to the active
      high pin is the single model for the diff pair.
  - Bob suggested that Model_type Output_diff would indicate a single
    instantiation.
  - Walter suggested changing [External Model] so that it is specific,
    and AMI could follow.
    - Michael M. said we can't change that spec now.
  - Arpad asked whether the waveform that is processed by the AMI model
    is always a single waveform.
    - The answer was Yes
  - Arpad suggested that the usage rules should be based on this fact
    and state that "if the [Model] is differential then the AMI model
    will process the difference waveform".
    - This was agreed.
  - Mike L. asked if these models can work without knowing common mode voltage.
    - Kumar said models have to handle this.
  - The final statement specifies that the [AMI] always processes a single
    signal regardless of whether the model is single or differential.
    When the model is differential the waveform pass to the [AMI] must be
    a difference waveform.
- Page 3:
  - Ambrish added a specification of reserved words that can not be used
    in parameters.
- Page 8:
  - Ambrish added examples sent by Mike S.

AR: Ambrish will send AMI_BIRD_2.2 to the list
AR: All review AMI_BIRD_2.2 by Friday
AR: Ambrish complete MSWindows* version of Cadence RX executable

Arpad asked what we will do next:
- Walter suggested continuing what Mike S. started.
- We might not meet every week.
- Bob said the committee needs to remain active to respond to BIRD questions.

Arpad asked how we submit the BIRD:
- Do we need to take a vote on submitting the BIRD?
  - The BIRD is essentially approved by the task force.
  - The current efforts are mostly editorial.
- Needs to be converted to text

AR: Mike L. will post AMI BIRD Drafts 2.0, 2.1, and 2.2 to the web

AR: Arpad will convert the Word document to text on Monday and we will
    have a final review in the next meeting.  Bob requested a copy before
    the meeting for review.
	 

Next meeting: 09 Oct 2007 12:00pm PT