IBIS Macromodel Task Group Meeting date: 05 Mar 2013 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: - None -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Arpad try to apply BIRD 158 shortcut language to his BIRDs 116-118 - In progress - Maybe we should schedule a vote for them - Ambrish email reflector regarding BIRD 147 status - It is ready for posting ------------- New Discussion: Interconnect - Michael M: Reviewed new EMD spec, now posted - No meeting this week - Expecting SI2 update for review next week Walter showed a slide about conversions to s4p - Walter: David Banas asked a question about conversions to s4p - In IBIS 6.0 David want to describe on-die interconnect as s4p - This is a surrogate model (someone from NCSU has call IBIS that) - The model in the top left takes a zero rise time stimulus - The diagram shows an equivalent RC circuit on the top right - The bottom left circuit can work correctly with a zero rise time - It has two s4ps, which can be combined into one on the bottom right - Fangyi: How does the waveform shaping happen? - The transition from top right to lower left might not work - Walter: It works like an op amp - I may not have the expertise to understand and explain it fully - Fangyi: There might be a one to many correspondence for that filter - Walter: The 1 to 2 insertion loss and the return loss at 2 are critical - David: I have to trim V-T tables to achieve high bit rates - Can an s4p incorporate the behavior of that? - Brad Brim had raised concern about that - I was cautioned about lumping these things into one s4p - Fangyi: I agree with that - This approximation is like a small signal analysis - The large signal operating point has to be chosen carefully - AMI is not always about linear - Superposition does not always hold, and it is more important than scaling - Todd: Know where you are in the operating curve is important - Once you have the right point an extracted touchstone model can be reasonable - Walter: The operating point will not be at either steady state - Radek: We are operating with a large swing, not a small signal - We are just basing an approximation on that point - Dan: Doesn't using a short rise time require large bandwidth? - Walter: Modeling the rise time requires NOT having full bandwidth - Todd: The s4p needs to go high anyway, otherwise the simulator handles the rolloff - David: All that matters is that the remaining energy above some point is minimal - Todd: Agree - Dan: I get nervous relying on extrapolation - Fangyi: We are modeling both the stimulus slope plus the RC delay - Walter: The rise times do not amount to much above Nyquist BIRD 147: - Ambrish: Suggestions were made to add sections to the backchannel BIRD 147 - Please comment on it AR: Mike post updated BIRD 147 BIRD 155: - Arpad motioned to untable BIRD 155 - David seconded - Fangyi showed New AMI API to Resolve Parameter Dependencies - slide 5: - Fangyi: AMI_parameters_in is the input - AMI_parameters_out is the new output - slides 9 & 10: - Fangyi described the flow - slide 8: - Fangyi: A new Usage Dep would be used for returned items - slide 11: - Fangyi: It easy easy to code FIR with this - slide 12: - Fangyi: An auxiliary file could contain the dependency information - Bob: This would have to be BIRD 155.1 - Radek: This could be used for Reserved_Parameters - Fangyi: You might use different TStonefiles for different corners - Arpad: I thought we would use dependency table parameter passing for [External Circuit] dependent parameters - Putting this in the DLL would take that feature away - Walter: The impulse response might be gathered in a separate simulation - The advantages given here seem to be mostly for the model maker - We should see what model makers think - Fangyi: This was proposed three years ago by IBM - Walter: Adge concluded that the .ami was the best place to put the data - This is more complicated - Radek: This is a simpler and more comprehensive solution - Mike L: Is Usage Dep needed? Those models will be unusable in existing tools - Radek: It will not be in the spec until 6.0 anyway - Fangyi: A workaround is to simply not change Usage in .ami files ------------- Next meeting: 12 Mar 2013 12:00pm PT Next agenda: 1) Task list item discussions ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives