IBIS Macromodel Task Group Meeting date: 11 Dec 2012 Members (asterisk for those attending): Agilent: * Fangyi Rao * Radek Biernacki Altera: * David Banas Julia Liu Hazlina Ramly Andrew Joy Consulting: Andy Joy ANSYS: Samuel Mertens * Dan Dvorscak * Curtis Clark Steve Pytel Luis Armenta Arrow Electronics: Ian Dodd Cadence Design Systems: Terry Jernberg * Ambrish Varma Feras Al-Hawari Brad Brim Kumar Keshavan Ken Willis Cavium Networks: Johann Nittmann Celsionix: Kellee Crisafulli Cisco Systems: Ashwin Vasudevan Syed Huq Ericsson: Anders Ekholm IBM: Greg Edlund Intel: * Michael Mirmak Maxim Integrated Products: Mahbubul Bari Hassan Rafat Ron Olisar Mentor Graphics: * John Angulo Zhen Mu * Arpad Muranyi Vladimir Dmitriev-Zdorov Micron Technology: Randy Wolff Justin Butterfield NetLogic Microsystems: Ryan Couts Nokia-Siemens Networks: Eckhard Lenski QLogic Corp. James Zhou SiSoft: * Walter Katz Todd Westerhoff Doug Burns * Mike LaBonte 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 Ray Anderson The meeting was led by Arpad Muranyi ------------------------------------------------------------------------ Opens: - Arpad: There is one more meeting this year, Dec 18 -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Walter update BIRD 150 - Walter updated it and will send soon - Walter define Usage of dependency table parameters - Fangyi, Mahbubul, et. al., to discuss merging their "Redriver" proposal with BIRD 131. - Fangyi: Trying to schedule Maxim partners on this ------------- New Discussion: Interconnect report: - Michael M: Meeting held last week - Steve Pytel from ANSYS presented - Si2 may be able to disclose spec details in February - May have something for DesignCON summit Fangyi showed "New API to Resolve Dependent Model ...": - 2 - Fangyi: Vendors may want to hide dependency formulas from users - A new API can handle this - 3 - Fangyi described the AMI_ResolveDependentParam() function - 4 - Fangyi described the flow steps to use the new function - Tools are expected to use malloc/calloc to allocate memory - David: It is not good to separate allocate and free - Mike: AMI_Close() does the freeing - Fangyi: We would need a separate function - This is called before AMI_Init() - David: We should discuss the idea of this affecting the analog model - Radek: If the DLL allocates the memory then AMI_Close can free it - 5 - Fangyi listed benefits of the new approach - Ambrish: How do we know which are for the analog model and which for AMI? - Arpad: Is the input an exact copy of the AMI_Init() input? - Fangyi: Yes - David: Why not include this in AMI_Init? - Fangyi: Because the analog parameters may change - Radek: The impulse response can not be determined until this is resolved - Arpad: BIRD 116 is related, parameterization of ISS subckts - David: Did that pass? - Arpad: No - Fangyi: If there are no analog Model_Specific parameters it could be handled without exposing anything - Arpad: If the output of the Resolve function goes to Init the params must be InOuts - Unless we make an exception - Fangyi: If a param goes through Resolve it should not be Out - Init should not change it the second time - Bob: There are Info Reserved_Parameters that affect the DLL - Arpad: They are not passed to the DLL - Bob: The last slide should be included in the BIRD - Fangyi: About Radek's idea, the memory would have to be passed to Init - The model has not been allocated when Resolve is used - Ambrish: How does this help with analog parameters? - Walter: In a large number of AMI models, param settings affect impedance - Strength settings choose TOUCHSTONE files, for example - David: We decided to let users choose by [Model Selector] - Walter: IC vendors prefer to set registers and have those affect analog settings - David: We are leaning toward using [Model Selector] - Walter: That is OK for a simple TX with few settings - We see models with 128 settings - David: Does this also include non-analog settings like tap weights? - Walter: We have seen a model with 256,000 different settings including C_comp - MM: We see cases where analog model and AMI file are contradictory - What if it was pointed to from the pin list? - Some people are not clear about how much analog should be in the DLL - David: This seems like a headache for tool vendors - Walter: A TOUCHSTONE file is always required for analog models - The IBIS analog section doesn't solve the problem - Ambrish: We need to keep analog and AMI models separate - Walter: BIRD 116 assumes drivers in high or low state - SerDes models are not in those states - Ambrish: That problem should be solved first - Walter: BIRD 122 addresses it - The Wed meetings are discussing this - David: We see good results with stock IBIS analog - Arpad: The BIRD 122 circuit can be done with legacy IBIS - Walter: We do not have TOUCHSTONE - Ambrish: We can add that - Walter: That falls apart with redrivers - SerDes drivers have many states - Ambrish: It sounds like dependency should be moved outside of AMI - SPICE can handle it - Arpad: Walter's proposal is equivalent to BIRD 116 - Walter: That requires wrappers and D-A/A-D converters Michael M: When will we discuss repeaters? - These are tied to backchannel proposals - Arpad: Dependencies make it hard to schedule topics - Repeaters and backchannel have dependencies on each other - Walter: Specifications like 802.3bj need to tell us how to do training Arpad: We will table discussion of BIRD 150 and Fangyi's proposal for now Fangyi noted that his proposal does not have to apply to analog parameters AR: Walter send BIRD 150.1 draft 7 to list AR: Mike post BIRD 150.1 draft 7 to web ------------- Next meeting: 11 Dec 2012 12:00pm PT Next agenda: 1) Task list item discussions ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives