IBIS Macromodel Task Group Meeting date: 01 March 2011 Members (asterisk for those attending): Agilent: * Fangyi Rao * Radek Biernacki Ansoft: Chris Herrick Danil Kirsanov Ansys: Samuel Mertens * Dan Dvorscak Deepak Ramaswamy Jianhua Gu Cadence Design Systems: Terry Jernberg * Ambrish Varma Celsionix: Kellee Crisafulli Cisco Systems: * Mike LaBonte Stephen Scearce Ashwin Vasudevan Ericsson: Anders Ekholm IBM: Greg Edlund Intel: Michael Mirmak LSI Logic: Wenyi Jin Mentor Graphics: * John Angulo Vladimir Dmitriev-Zdorov Zhen Mu * Arpad Muranyi Micron Technology: Randy Wolff Nokia-Siemens Networks: * Eckhard Lenski Sigrity: Brad Brim Kumar Keshavan * Ken Willis SiSoft: * Walter Katz Mike Steinberger Todd Westerhoff 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: - None -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Bob check BIRD 127 for type clarification possibilities - In progress - May need to clarify float numbered indexing for taps - Ken draft language to clarify handling of multiple channels - Done - Ambrish draft BIRD for relationship between Type and Format (row 25) - No progress yet - Arpad follow up with Kukal on BIRD 121 - Done, may be more - Bob write a BIRD on correcting Table 1-3 in the spec. (Row 23 in the Task List). - In progress - Ken and Ambrish send examples of AMI model requiring array input - Done ------------- New Discussion: Arpad showed the new Crosstalk BIRD draft: - It has asterisks added to make it more readable - Ken: First column of matrix is primary channel, others are aggressors - Aggressor TX has to examine its own EQ first - Init must receive it's own channel impulse response - Init returns response to rewrite first column of matrix - RX may modify any matrix columns - Fangyi: Model makers will not know if it will be used as primary - Ken: EDA tools will have to manage that - Walter: This is not what needs to be done - The primary TX is connected to an RX - TX Init sets EQ for primary - Then it has to apply that to the other channels - RX then does almost the same thing - Arpad: This BIRD covers that - Walter: Not if it says TX can only change itself - Arpad: The tool will have to copy the responses - It knows which is primary and aggressor - This BIRD describes the matrix, not the flow - Fangyi: This is aggressor TX to victim RX? - Arpad: Yes - Radek: It won't work if one model has to do both - Walter: We should work this out in a limited email thread - Ambrish: Graphic representation would be good - Please include me - Radek: There could be parameters to identify the aggressor Walter showed "History and Usage of Out and InOut Model Specific AMI Parameters": - The need to present model state information has been declared in the past - A fundamental principle of AMI is every EDA platform gives the same results - Platforms may differ on choice of inputs to the models - May differ on output processing - Mike LaBonte explained in an email - Ken: Who said we could not write any outputs? - Walter: Ambrish - Scott: We have no way to define the meanings of the outputs - Without meaning we can't evaluate - This is incomplete from a specification point of view - Walter: Every param has a description - The user will know how to interpret - Arpad: Machines still can't interpret - Ken: It should be specified what will be done with it - Walter: The description will specify - Scott: Why object to more clarity? - Arpad: Maybe a perl script will process it - We can't leave it unspecified - We should at least say it should be put in a file - Walter showed an email replying to Fangyi - IBIS 5.0 allows values to be output - Models have info at Close - There is no other way to dump that info - Arpad: How would a tool plot string parameters - Ken: Models can write their own text files - That would be easier than parameter values - Radek: String output could have any data - This can be presented ready to be used - Ambrish: Could this be used for backchan? - Walter: Kumar came up with the same scheme independently - Arpad: We could specify a machine readable format - But the tool still will not know what to do with it - Walter: Each platform may differ on processing - We can't tell them what to do with it - The tool can only display a spreadsheet - The user must determine what to do with it - Mike: This idea enables innovation - The alternative is to wait until new keywords are in the spec - Parameter output may be more convenient than writing text files Arpad: We should spell out a little more how jitter param outputs are used: - Walter: The tool should use Init jitter but not GetWave - Arpad showed an email replying to Mike Steinberger - Radek: We don't need a universal rule - A rule for each parameter is OK - Walter: For new params we could cover it in a BIRD for 5.2 - Ambrish: We could have a rule that Model Specific outputs can't change results - Walter: That's OK ------------- Next meeting: 08 Mar 2011 12:00pm PT Next agenda: 1) Task list item discussions ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives