====================================================================== 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 September 6 Meeting (* means attended at least using audio) ANSYS Curtis Clark Cadence Design Systems Bradley Brim Cisco David Siadat Intel Corp. Michael Mirmak Keysight Technologies Radek Biernacki, Ming Yan Mentor, A Siemens Business Arpad Muranyi* Micron Technology Justin Butterfield*, Randy Wolff* SAE ITC Maureen Lemankiewicz, Logen Johnson Signal Integrity Software Walter Katz*, Mike LaBonte* Teraspeed Labs Bob Ross* University of Aveiro in Portugal Wael Dghais Mike LaBonte convened the meeting. No patents were declared. Justin Butterfield took minutes. Review of Minutes: - Mike called for review of the minutes from the September 1 meeting. Walter Katz moved to approve the minutes. Arpad Muranyi seconded. The minutes were approved without objection. Review of ARs: - Arpad and Walter to draft text for missing die pad to pin model order of precedence. - Walter noted that he would like to get a clean draft to work on this and on other issues. Opens: - Walter has a proposed version of the BIRD draft document with the [Interconnect Model Set Selector] keyword changed to [Interconnect Model Set Group]. BIRD189.5 draft 6: Mike noted that the changes discussed last time were made to BIRD189.5 draft 6 and it has been sent out and posted. He also added some comments on the remaining issues that still need to be addressed. Selector to Group Proposal: Walter stated that he started with BIRD189.5 draft 6 and accepted all the changes while retaining the comments. He made changes focused on eliminating [Interconnect Model Set Selector] and replacing it [Interconnect Model Set Group]. He commented that [Interconnect Model Group] is also a possible new name for the keyword. Walter noted that the EDA tool would use models from the [Interconnect Model Set Group] selected by the user. The group defines a list of [Interconnect Model Set]s to be used together in a simulation. The user selects one group for a given simulation. Walter commented that he prefers the keyword to be named [Interconnect Model Group] rather than [Interconnect Model Set Group]. He will revise the proposal with this change before sending it out. Bob Ross agreed with the name [Interconnect Model Group]. Walter noted a new rule was added. It states that "If an Interconnect Model Set name appears in different Interconnect Model Groups, then the file containing the two Interconnect Model Sets must be identical." Walter noted that he has made some changes to the examples to reflect the change to [Interconnect Model Group]. Example 1 did not have a [Interconnect Model Set Selector], so he added a group. Arpad commented that in the examples it might be good to add in different groups to select from. Walter replied that he did that in some cases. In example 4, there are 2 groups one for the Touchstone model and one for the ISS model. Bob commented that in the examples the [Interconnect Model Set]s were used, but the [Interconnect Model Set Selector]s were in the [Component]s. Walter stated that the Groups are used in the [Component] section as well. Arpad commented that it would be nice to have an example with more than one set. Walter agreed. He thought that in Example 4 we could add a PDN model for both Touchstone and ISS model groups. Mike commented that we are creating sets of sets called groups. Walter clarified that we should call it group of sets. Mike has some concern about the word "group". Arpad had an issue with the word "selector" since the user could select more than one entry in the list. There is also the issue of how to choose the sets and make sure there are no missing or conflicting pieces. The proposal to change to groups solves these issues. Mike asked if the user can select the name of the group. Arpad replied, yes, the user would select from a drop down menu or similar the name of the desired group. Bob commented he likes that when the user makes a selection, everything that it is intended to be in the simulation is in one group. He likes that you can reuse models for different groups. He asked why another level of indirection is necessary, as [Interconnect Model Set]s can do the same thing. The concern is that groups are unnecessary, as the sets could be complete. Walter commented that one of our requirements for the Interconnect BIRD was that the on die and package could be in different sets, and you could have different sets for power delivery and signals. Bob noted that we could satisfy the requirement by changing how sets work. Walter stated that the proposal to change to groups is a relatively simple change to the BIRD. Arpad stated that his view of Bob's concern is an efficiency of the syntax question that we do need to think about. Walter took an AR to send out the draft with the proposed change to groups for everyone to review. Walter stated that diagnostics could be done on groups to check for model completeness. Mike added that the groups proposal is a good way for the model maker to convey sets that work together. Arpad asked if we should wait for this proposal to be approved to start making other changes to the BIRD. Walter will send it out this draft, and he would like to review this and get closure on Friday. He will plan to make a motion. This change affects a lot of the document. He proposed to make a parallel draft to work on the other outstanding issues. Bob commented that there is a new issue regarding if A_gnd can be used in general to connect to node 0. Mike asked about how this would work inside IBIS-ISS. Arpad replied that there is no change to IBIS-ISS as it already has node 0 as ideal ground. Walter stated that the proposal is to allow node 0 as a terminal of an interconnect model. We need to make that decision then decide on the syntax. Arpad noted that we still need to discuss the order of precedence for missing models in the path. Bob advised against setting an order of precedence as we never intended to mix the models. Mike brought up requirement number 15 in BIRD189.5. This says that we cannot mix model formats. Bob stated this does not distinguish from buffer to pad or pad to pin. Walter thought that the EDA tools are not required to combine the models. If an EDA tool supports some mismatch of formats, that is not forbidden. Arpad has written a proposal on this issue for this group to review later on. Walter asked Mike to make a list of the remaining open issues. Arpad moved to adjourn. Walter seconded. The meeting adjourned without objection.