================================================================================ IBIS INTERCONNECT TASK GROUP http://www.ibis.org/interconnect_wip/ Mailing list: ibis-interconnect@freelists.org Archives at http://www.freelists.org/archive/ibis-interconn/ ================================================================================ Attendees from May 2, 2018 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Intel Corp. Michael Mirmak* Keysight Technologies Radek Biernacki Mentor, A Siemens Business Arpad Muranyi* Micron Technology Justin Butterfield*, Randy Wolff* SiSoft Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* Michael Mirmak convened the meeting. No patents were declared. Justin Butterfield took minutes. Review of Minutes: - Michael called for review of the minutes from the April 25 meeting. Arpad Muranyi moved to approve the minutes. Mike LaBonte seconded. The minutes were approved without objection. Review of ARs: - Mike to send out the submission of the final BIRD189.5 to the Interconnect mailing list. - Mike reported this is done. - Bob Ross to prepare a BIRD189.6 draft with the proposed rail rule changes. - Bob reported this is done. Opens: - None Rail Simplification Discussion: Bob has both a PowerPoint presentation and new BIRD draft document with the proposed changes to discuss. Bob shared his slides first. He commented that for the BIRD, he mostly restored the changes taken out at a previous meeting. He does have a few examples, but did not include all the possible cases. The rationale for the changes is that it follows the nodal connectivity paradigm in SPICE. The simplification avoids unexpected blocked situations, avoids additional clarifications, and can simplify the parser. Walter Katz noted he has reviewed the presentation and document, and he supports the changes. Bob continued noting that all Interconnect Models without I/O terminals with only rail terminals are available for simulations with the proposed rule changes. If some terminals are not used in a particular interconnect, the tool would have to stub out the pins. And, PDN structures should not be left dangling by the tool. On slide 3, Bob outlined four types of Interconnect Models, including I/O paths only, I/O paths with reference terminals, I/O paths with PDN terminals, and PDN terminals only. He gave some examples of these in the following slides. Bob noted the purpose of the simplification is that it prevents some of the unintended consequences. We are forced to use workarounds such as separate PDNs and adding connections to decoupling models. On slides 7 and 8, we have two PDNs in the same Group. We have embedded PDN models in the Interconnect Model. For ball grid arrays, you may want to bring in models of sections of signals and power/grounds together as separate models for each section. This could support the W-element with different references at each end. On slide 11, Bob listed the current rules. On slide 12, he compared the much simpler rules he is proposing. He highlighted that the proposed changes would avoid complex checking and reduce confusion about what is allowed. On slide 15, Bob commented that people must understand the specification, the interconnect structure, and the IBIS-ISS or Touchstone files and their connections. Without the rule simplification some structures model developers might expect to create are blocked or have restrictions. With the proposed simplification, the model developer is expected to understand and correctly model the structure to avoid unintended path duplications. Walter commented that he understands what Bob is trying to do, and he supports it. Arpad stated, in general, he supports the idea, but he has some wording related concerns. BIRD189.6_v1_br Review: Bob shared the BIRD189.6_v1_br that he sent out. He said he intends to submit a motion to accept this as the new draft. Michael asked if it should be labeled as a draft1. Bob agreed this would be better. Bob noted, starting on page 8, he added a rule for interconnect models with only one interface for the decoupling model case. He also clarified the rule stating PDN models have two interfaces. On page 9, Bob simplified the first two rail terminal rules. And he added the pin, die pad, and pad only for decoupling circuits to the interface combinations list. He also added a rule that all PDN models can be brought in for simulation. On page 18, Bob clarified and added that only one interface can be used for Rail models. On page 23, he added decoupling models as an application. On page 24, Bob added that an Interconnect Model for rails can have terminals at only one interface. Mike suggested to add PDN only in parentheses to make it more consistent. Bob thought it could also be a decoupling model. Bob agreed that we can work on the wording. On page 25, Bob added pin rails only, die pad rails only, and buffer rails only to the list of what an [Interconnect Model] may contain. Walter noted that there are editorial changes need. He stated that he agrees with the changes and asked if we should go ahead and submit this or if we should work on the wording first. Bob noted that BIRD189.5 has been submitted and we need to introduce those changes. Michael suggested that we can introduce BIRD189.5 and work on the wording for BIRD189.6 in parallel. Arpad agreed. Walter thought in the next IBIS Open Forum meeting after this Friday we can introduce BIRD189.6. Bob suggested for anyone to submit any wording changes for a draft1. Arpad stated he needs to read it again. Bob mentioned that on page 34 there are rules that relate to connections in Interconnect Model Groups. He would like a simple a paragraph that states: Between two Interconnect Model Groups, same named terminals are assumed to be connected in a Group. This is true at all the interfaces. Mike asked if Bob means Sets. Bob clarified that he does mean same Group. Walter suggested for Bob to submit this text in an email. Arpad asked if we have any examples of using a decoupling model. Bob replied we do not. Walter stated that we do not, as it was illegal before. He agreed we should add one. Bob noted that there are still some redundant statements which he would like resolve. Walter stated that we should set a goal to have everything finished in 3 weeks and set the agendas for the upcoming meetings. Michael asked if Bob would be willing to write a decoupling example for Wednesday's meeting [AR]. Bob agreed. Michael asked if there are any objections to posting the proposed draft from Bob. Walter stated it was okay to post. Michael asked Bob to change the title to draft 1 and Mike to post it [AR]. Michael asked everyone to take a look at the wording of the new text. Michael noted he does not see any issues with introducing BIRD189.5 and stating there will be a BIRD189.6 to follow. Arpad agreed that it would be good to introduce the changes in BIRD189.5, as there are many changes. Bob stated he will work on some replacement text for page 34 [AR]. Michael noted that we are in a race with the BIRD158 replacement BIRD for gating IBIS 7. Walter suggested we need to set a deadline of the Interconnect meeting before the next IBIS Open Forum. Michael agreed we should take the next 2 or 3 meetings to finalize BIRD189.6 for submission to the IBIS Open Forum. Next Meeting: The next meeting will be May 9. Mike moved to adjourn. Arpad seconded. The meeting adjourned without objection. Task List BIRD189.5 editorial additions/changes to be completed: 1. Resolve inconsistency with using node 0 and node "0" by replacing with "node 0".