IBIS Macromodel Task Group Meeting date: 23 January 2018 Members (asterisk for those attending): ANSYS: * Dan Dvorscak * Curtis Clark Broadcom (Avago): Xingdong Dai Bob Miller Cadence Design Systems: * Ambrish Varma Brad Brim Kumar Keshavan Ken Willis eASIC: David Banas Marc Kowalski Ericsson: Anders Ekholm GlobalFoundries: Steve Parker IBM Luis Armenta Trevor Timpane Intel: * Michael Mirmak Keysight Technologies: Fangyi Rao * Radek Biernacki Ming Yan Maxim Integrated Products: Hassan Rafat Mentor, A Siemens Business: John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff * Justin Butterfield SiSoft: Walter Katz Todd Westerhoff * Mike LaBonte SPISim: * Wei-hsing Huang Synopsys: Rita Horner Kevin Li Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross TI: Alfred Chong The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - Arpad noted that the meeting scheduled for January 30th will not be held because most attendees will be at DesignCon. ------------- Review of ARs: - Walter to send BIRD189.5_draft15_v7 to the Interconnect list. - Done. It has subsequently progressed to draft16_v3. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: - Arpad: Does anyone have any comments or corrections? [none] - Mike L.: Motion to approve the minutes. - Dan: Second. - Arpad: Anyone opposed? [none] ------------- New Discussion: BIRD189.5_draft16_v3 - Arpad: [sharing draft16_v3, sent to Interconnect and ATM on January 19th] - The only major change in this version is the new text to explain the second Aggressor_Only figure. - It also establishes some rules for this case. The default will be to use the first Model in the first Set that contains the pin. [more detail below] - We hope to finalize this version in tomorrow's Interconnect meeting and submit it as BIRD189.5. - Discussion: Several participants had proposed changes. Arpad edited the draft real-time to create v4. Changes from Bob: - Pg. 29/30: - instances of "interconnect model" -> "Interconnect Model" - Under the Aggressor_Only section: - 4 instances of "model" -> "Interconnect Model" - Pg. 7: - Second to last bullet, starting with "A *_I/O pin_name which does not": - Break this bullet into two sentences at ", however". - Consider rewording the second sentence (Arpad added a comment for this). - Last bullet: - Bob noted that he thought the last phrase of the bullet, "... an Interconnect Model that contains a terminal with this pin_name." was redundant and should be removed. - Curtis said he thought it was not redundant. The bullet was trying to say that the model that will be used by default in this situation is the first Model that contains the pin, but it's the first Model in the first Set that contains such a Model. Mike and Arpad agreed, and Mike noted that the sentence is trying to describe something with multiple layers and is therefore complicated. - Arpad marked the sentence with a comment to review the wording. - Pg. 8: "General description of rail terminals" bullet: - In the first sub-bullet of this section, the word "single" may be redundant in "single signal_name or bus_label." - Arpad marked this with a comment to review "single". - In the next sentence, replace "If it" with "If the terminal". - In that same sentence, Radek noted that "these pin_names are shorted together" is improperly worded (pin_names are not shorted, associated terminals are shorted). Randy noted a grammar fix: "he case" -> "the case" (2 instances). Mike noted that we could remove the comment stating that Arpad should change the intro language for the Aggressor_Only figure. Arpad removed it. Curtis noted that at the previous ATM meeting he thought Bob/Arpad/Walter had considered making it illegal for the model maker to create a Model Group with one pin appearing in multiple models as Aggressor_Only but never as a victim (the case illustrated by Arpad's additional Aggressor_Only figure). Arpad said he hadn't been sure what the action plan was from the ATM meeting, but they had subsequently decided not to disallow it. They had instead described the first-Model-in-the-first-Set rule, and they had added text recommending that the model maker not create such a Group (see the last sub- bullet of the "I/O pin_name_rules" bullet under Usage Rules for [Interconnect Model Group]). Mike then noted that the last bullet of this new section said "...one or more Interconnect Models without the Aggressor_Only...", and the "or more" was probably incorrect. Arpad added a comment to review this wording. Bob noted that he thought the proposal still had technical issues. Arpad and Mike L. expressed surprise at this and noted that we had been working on reviewing final wording details. Bob said that he thought the fallback to legacy package modeling techniques has problems. He said we wouldn't know which legacy package model to use. Arpad noted that there had been extensive discussions on precedence rules. He noted that the only thing that is not allowed is mixing coupled [Define Package Model]s if the coupling overlaps with a BIRD189 Interconnect Model. That is, if pin N couples to pin N+1 and pin N-1 in the [Define Package Model], then it can't be used with a BIRD189 model that uses pin N+1 or Pin N-1. Mike agreed with Arpad's recollection and said the discussions had even changed the definition of legacy package model. Bob noted that the BIRD contained no statement of the rule Arpad described. Mike asked if Bob would take the AR to add this language. Bob agreed to look at it, but said he might prefer to ban fall back to [Define Package Model] since we don't even know what the coupling is (without the parser looking into the actual matrices) with an RLC matrix format [Define Package Model]. Arpad said he would send the modified draft to the Interconnect and ATM lists as draft16_v4. Mike L. agreed. Mike noted that before we can post this as an official BIRD189.5, we should deal with any comments still lingering in the draft. - Mike L.: Motion to adjourn. - Curtis: Second. - Arpad: Thank you all for joining. AR: Arpad to send BIRD189.5_draft16_v4 to the Interconnect and ATM lists. AR: Bob to propose language to address his concerns about fallback to legacy package models with BIRD189. ------------- Next meeting: 06 February 2018 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives