================================================================================ 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 February 21, 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 February 14 meeting. Mike LaBonte moved to approve the minutes. Randy Wolff seconded. The minutes were approved without objection. Review of ARs: - Michael to clarify the File_TS comments with Radek Biernacki. - Michael reported that Radek was able to clarify these comments in ATM. - Arpad to add the BIRD158.7 referencing discussion to the ATM agenda. - Michael reported that this was done. - Mike to send out BIRD189.5_draft17_v3. - Mike reported that this is done. Opens: - None. Review from ATM: - Michael asked if there were any ARs for this group from the ATM meeting. Arpad noted that the referencing issue was discussed, but no resolution was reached. There could be issues with different references between models. Two suggested approaches were discussed in the meeting. One approach is to force the reference to node 0 for all models, and the other is to allow different reference connections in an N+2 terminal approach with separate references for each side of an interconnect model. Arpad thought we are leaning toward the node 0 approach, but this has not been decided. The question remains for BIRD158.7, what should be said about the reference connection. And, for BIRD189.7, should we force the reference to A_gnd. Walter Katz clarified that this would only be for the File_TS shortcut, and in that case, we could remove the N+1 terminal for File_TS. Michael asked if it would be possible to remove the node 0 restriction in the future. Walter noted the only case needing to remove the restriction is for power aware simulations, where current flows from one terminal to another. He thought we could add another keyword to support this. Arpad noted we ran out of time to make a decision on this issue in ATM. Michael asked what changes would be needed to go with Walter's proposal of removing the N+1 terminal. Arpad stated that this would mean significant changes to the BIRD189.5 text. Mike wondered what the logic was to include the N+1 terminal to begin with. He also wondered if BIRD158.7 oversteps its bounds by showing the interconnect channel and its reference connections. Arpad stated that the drawing showing this is debatable if it should be included in BIRD158.7 or in a cookbook type document. Mike stated he will go through old minutes to determine why the N+1 terminal was added [AR]. BIRD189.5_draft17_v4 review: Michael stated on page 28 there was typographical error in the Terminal line rules last paragraph. Arpad suggested to also add the word "in" before the word "simulation". Michael noted on page 30 there is a comment from Bob Ross regarding the third paragraph in the Aggressor_Only section. Mike stated that this is a new requirement that if Aggressor_Only appears on one of the terminals it must be assigned to corresponding terminals at all interfaces. Michael asked if we are okay with this rule from a technical standpoint. Mike noted this is a parse-able rule. Bob stated that this forces Aggressor_Only to be set at both interfaces. Michael asked if this would be an error or warning. Bob thought it should be an error. Arpad stated that without this rule you could have conflicts between on-die and package models. Bob noted that any one Interconnect Model contains two interfaces. Mike suggested that we should retain the text that this rule is applied across terminals with the same pin_name. Michael asked if we are checking this across the Group and Set or the Interconnect Model only. Arpad asked about the case where you have separate Interconnect Models for the on-die interconnect and the package. Bob replied that the through rule applies to Aggressor_Only. If you have separate Interconnect Models in series for the on-die interconnect and the package, the pins will automatically be flagged as Aggressor_Only based on the pin_name setting for the other model. The other model is not required to be labelled as Aggressor_Only. Mike noted that on-die interconnect and package models may not be aligned one to one. If there is any Aggressor_only along the way, then the entire path has to be treated as a Aggressor_Only. In the last sentence of third paragraph in the Aggressor_Only section, Arpad suggested to use the phrase "terminal line". Michael added the phase "within any Interconnect Model" to the beginning of the sentence. Arpad suggested to remove the word "particular". Mike suggested to delete the first instance of the word "associated". Michael suggested for next time, we can discuss the N+1 terminal. Mike moved to post BIRD189.5_draft17_v4 as BIRD189.5_draft17. Arpad seconded. There were no objections. Mike to post the BIRD189.5_draft17 [AR]. Next Meeting: The next meeting will be February 23. Mike moved to adjourn. Arpad seconded. The meeting adjourned without objection. Task List BIRD189.5 editorial additions/changes to be completed: