============================================================================== IBIS INTERCONNECT TASK GROUP Mailing list: ibis-interconnect@freelists.org ================================================================================ Attendees from April 9, 2025 Meeting (* means attended at least using audio) ANSYS Curtis Clark, Wei-hsing Huang, Juliano Mologni Arista Networks Jim Antonellis Broadcom James Church Intel Corp. Michael Mirmak*, Xiaoning Ye Keysight Technologies Ming Yan Marvell Steve Parker MathWorks Walter Katz* Micron Technology Justin Butterfield Siemens EDA Weston Beal, Arpad Muranyi*, Randy Wolff* Simberian Yuriy Shlepnev ST Microelectronics Aurora Sanna Synopsys Ted Mido, Edna Moreno University of Illinois Jose Schutt-Aine Zuken USA Lance Wang Michael Mirmak convened the meeting. No patents were declared. Michael reviewed the minutes from the March 19 meeting. Walter Katz moved to approve the minutes; Randy Wolff seconded. The minutes were approved without objection. Michael reviewed the minutes from the March 26 meeting. Arpad noted that the minutes contained the word "ompatibility" as well as a stray "t" character. Arpad moved to approve the minutes with these corrections; Walter seconded. The minutes were approved without objection. Arpad asked about the latest port-mapping draft; Michael noted that it was draft 11. Michael took the AR to send draft 11 to the reflector. [AR] Michael reviewed new language he is proposing for IBIS and IBIS-ISS regarding connections between IBIS, IBIS-ISS, and Touchstone files. On IBIS, Arpad expressed concern that EDA tools could have some checks or other verification that models are compatible. However, actual checks may not be possible in all cases. Michael suggested that it may be possible to wrap an incompatible model to make compatible with other models. The team agreed that ensuring connectivity and compatibility is really all up to the EDA tool. Arpad agreed that a caution message in the specification would still be useful. Walter expressed disagreement with the statement recommending avoiding adding node 0 inside subcircuits. Michael replied that SPICE hygiene very important, at least for circuit compatibility in his corporate experience, including avoiding node 0 except at the top level, and ensuring subcircuit parameter declarations are explicit. Randy agreed, noting it was a priority at his former employer as well. He added that there can be, in some SPICEs, a distinction between local and global ideal node 0. Arpad added that this distinction is mentioned in at least one of Sam Chitwood's IBIS Summit presentation. Walter replied that the W-element uses node zero as "local ground". This node is the place where return currents go. Arpad noted that local node 0 treatment is a different issue than passing node connections and parameters in subcircuit instances. Referencing is just a small part of the problem highlighted in the circuit examples featuring L-R circuits in the VSS path; you can have no node 0 and still have incompatible subcircuit definitions. Michael asked the team whether the EDA tool is still the only arbiter here of connectivity. Refinement on loop concepts in his proposed language is still needed. A separate statement regarding node 0 is also still needed. Walter: asked Randy about memory circuit cases with ~30 different VSS and VCC paths (connections) for DC analysis. This implies multiple elements. What are the datasheet DC, AC loading (for SPIM, stimulus or observation ports) requirements? He added that this kind of analysis would be incredibly difficult. Randy agreed, noting that no current profiles were shared, and he was not sure there ever will be. Walter pointed out how little we understand power integrity; getting the basic requirements is the major challenge of most companies. Arpad asked whether analyzing maximum currents was all that was currently being done. Michael outlined his revision suggestion for the port-mapping proposal, including removing the file references but keeping the Physical, Logical, etc. connections. This would also involve removing the EMD- and IBIS-specific language from Touchstone. Arpad suggested that the order in which the models were made and who creates them are critical considerations (e.g., for package interconnects). Walter noted that most of the intent of the model connectivity terms was to provide names for engineers to read, perhaps with automation support. Arpad stated that the purpose of the IBIS and EMD language was to say whether a port was connected to a signal pin, pulldown_ref terminal, etc. Michael replied that this could be smuggling contamination from IBIS back into Touchstone, which would have only niche appeal. The wrapper approach may be more appropriate. Randy commented that original port-mapping proposal was not renaming anything in IBIS; Arpad agreed with this concept, if these are strictly names. Michael added that parser cross-checking would be more difficult with the port-mapping proposal, as the IBIS and Touchstone (and possibly IBIS-ISS) parsers would have to interact. He added that SPIM is not mentioned in the draft but had been discussed as a fifth file type. Arpad replied SPIM already defines the port mapping independently. He agreed that the draft needed Physical naming, but not file references. Walter stated that the person putting the model together should put as much information into the port mapping section as possible for the user, not necessarily to support automation. Michael took the AR to send out his slides to the Interconnect reflector. [AR] He also took the AR to made modifications to the proposed text as suggested during the meeting by the team. [AR] Michael asked the participants to accept an AR: consider, and be prepared to state next time, what changes are still needed to the port-mapping draft to prepare it for a vote. [AR] Randy moved to adjourn; Arpad seconded. The meeting adjourned without objection. The next meeting will be held April 16, 2025. ================================================================================ Bin List: 1) Complete port naming proposal (Katz et al) 2) Complete/revise Touchstone 3.0 draft outline (Mirmak) 3) Complete ISS-IRD 1 Draft - enable cascading of S-parameters through W-element (Mirmak) - TABLED 4) ISS-IRD ISO/IEC