IBIS Macromodel Task Group Meeting date: 31 July 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 Arrow Electronics: Ian Dodd Cadence Design Systems: Terry Jernberg * Ambrish Varma Feras Al-Hawari Cavium Networks: Johann Nittmann Celsionix: Kellee Crisafulli Cisco Systems: Ashwin Vasudevan Syed Huq Ericsson: Anders Ekholm IBM: * Greg Edlund Intel: * Michael Mirmak LSI Logic: Wenyi Jin Maxim Integrated Products: * Mahbubul Bari 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 Sigrity: Brad Brim Kumar Keshavan Ken Willis 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: There was an AR from last week for me to create a bus package model example using BIRD 125 - Will be addressed in presentation today - Bob to propose a simpler way for addressing the needs of parameter passing under [External Model] and [External Circuit] - not done ------------- New Discussion: Arpad showed a presentation "Package Modeling in IBIS": - slide 2: - Arpad: Colored terms indicate new syntax - Michael M: You have to use node declarations? - Arpad: You don't have to, they are just for the extra, explicit nodes. - James: What if there are more die nodes than pins? - Arpad: A later slide covers that - IBIS assumes implicit pad names based on pin names - Explicit names are needed when there is not one-to-one mapping - This syntax overrides implicit names only as needed - Michael M: That may be a major discussion topic - slide 5: - Arpad: This shows two die pad nodes connected to pin 2 - Also two pins are connected to die pad_4 - slide 6: - Arpad: We did not originally define in IBIS how to connect power - Buffers are instantiated by column 3 of [Pin], so there can be only one - slide 9: - Arpad: We would need cascaded circuits here - slide 10: - Arpad: BIRD 145 proposes cascading models to solve the problem - We have to provide names for the buffer terminals - slide 11: - Arpad: This shows both BIRD 125 and 145 - This should answer some of the email questions - In BIRD 145 [Model Call] instantiates buffers - The legacy IBIS keywords match ports by pin name - The IBIS-ISS subckt call matches by position - In the [Circuit Call] portion ports are connected to [Node Declaration]s - Michael M: The objective is to make optimum use of existing package keywords? - Arpad: Yes, we should minimize changes - Michael M: For tool vendor minimal changes are good - Model makers might be in a different position - Almost no one uses [Define Package Model], so they would have to learn it from scratch - Arpad: Is [Define Package Model] unused because the syntax is so bad? - Michael M: Electrically long packages are hard to model with IBIS - The syntax is constraining - Swathing is not easy either - slide 12: - Arpad: This shows that arbitrary interconnections are possible - You can't connect two pins to one package port though - slide 13: - Arpad: This explains the previous limitation - There can be only one port number per pin - slide 14: - Arpad: Red and blue lines show power and ground connections - slide 16: - Arpad: A limitation here is the assumption of fixed pad topology - slides 18-24: - Arpad: These show a model sliding up and down the pin map - slide 25: - Arpad: This shows the syntax that does the sliding - We call port_map names instead of node names - Pins are assigned to nodes positionally - It works somewhat like [Model Selector] - The Position parameter selects which mapping to use - Parameter trees with dependency tables could be synchronized with the Position - Radek: Is there a continuation character if pin or node lists become long? - Arpad: The words "pins" and "nodes" indicate what is on the rest of the line - There could be more than one of each - That could be refined - Radek: Ports have two terminals, it would be best not to use that term for this - James: On slide 18 what is the difference between [Pin] and [Pin Numbers]? - Arpad: [Pin Numbers] is in the package and it has to repeat what is in [Pin] - James: How can slide 12 show a one-to-many relationship? - Arpad: The IBIS spec may allow a subset, which would allow this - If not we can change the specification - James: But is connecting one to many possible? - Arpad: That may have been a mistake - RAMBUS had a pad connected to two pins - James: It could be done inside the IBIS-ISS subckt Arpad: We could go through Walter's presentation: - Walter: There is not enough time now - Michael M: Will there be other IC vendor presentations? - Ray: One of the proposals submitted so far should be good enough - Randy: Agree with Ray, we should go over Walter's presentation - Arpad: My proposal does not include X/Y coordinates - Someone could propose something along that line - Walter: Are there any models using [Circuit Call] and other keywords together? - Arpad: Probably not, but there are a few [Define Package Model] - Ray: We have delivered hundreds of [Define Package Model] with RLC matrices - That format is running out of gas - Michael M: Anders might check the model library - James: It would help to have a real IBIS file example - Walter: Example models based on my example cases would be good AR: Arpad prepare example BIRD 125 IBIS file set ------------- Next meeting: 07 Aug 2012 12:00pm PT Next agenda: 1) Task list item discussions ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives