=============================================================================== IBIS INTERCONNECT TASK GROUP Mailing list: ibis-interconnect@freelists.org ================================================================================ Attendees from September 4, 2024 Meeting (* means attended at least using audio) ANSYS Curtis Clark, Juliano Mologni Broadcom James Church Intel Corp. Michael Mirmak* [Michael Brownell] Xiaoning Ye Keysight Technologies Ming Yan Marvell Steve Parker MathWorks Walter Katz* Micron Technology Justin Butterfield Siemens EDA Weston Beal*, Arpad Muranyi*, Randy Wolff* ST Microelectronics Aurora Sanna Synopsys Ted Mido, Edna Moreno University of Illinois Jose Schutt-Aine Zuken USA Lance Wang Michael Mirmak called the meeting to order. No patents were declared. During review of the minutes, Arpad Muranyi moved to approve. Weston Beal noted that the minutes contained an incorrect reference to September 3 rather than September 4. Arpad amended his motion to approval with this date corrected. Weston seconded. The minutes with this correction were approved without objection. Michael noted that his previous ARs were still open and being worked. Arpad noted that his groups AR was still open. Note that several edits were made directly to the document and are not noted here in detail. Michael asked about the keywords IBIS_file and the like: do the referenced files have to be placed in the same subdirectory? Arpad asked whether IBIS had a rule about referenced files being in the same directory or below as the referencing file. Weston replied that this was true, but that in this case, the relationship must be in reverse, upwards. Michael suggested this might be a security issue for parsing. Arpad asked whether we would call the other parser (e.g., ibischk), to check other formats. Weston replied that this would be more complicated than needed. Randy Wolff suggested the parser could check simple things, such as correct port numbering. Otherwise, yes, it would be dangerous to go up the directory tree. Weston added that the information was only for documentation purposes, not for compilation; at most, the parser could check the same folder, then issue a caution if the file was not found. Arpad suggested that using source code rather than executables in tools is difficult. Could the parser throw error messages when file permissions not available? Michael accepted the AR to contact local security experts to see if file traversal upward still represents a problem. [AR] Walter Katz suggested that the tool can look in the same directory and the one above, and not any further; the limit could be established there. Randy agreed, as did Weston, adding that this covers 99% or more of the cases. Michael noted that, in Example 1, Symbol_left etc. Short Cuts are not defined: Bogatin, IEEE, etc. are mentioned only in passing. Walter replied that notations like NFNF and NNFF are already used as part of commercial SPICE tools. Michael asked whether we could use alternative names. Arpad expressed dislike of the idea. Weston agreed, noting that this is redundant information and could be a comment. He asked whether these for net connections or symbol generation. Walter replied that the data could be used for the symbol and then for the net. Weston asked whether the parser should check for consistency; this is harder for the parser writer. Walter added that this was easier for the model-maker; one could use just NFNF without Net numbers. Having both helps verification. One could use *either* NFNF or Net left/right. Arpad replied that he generally doesn't like redundancy. Weston agreed, adding that the document could mention this approach, with the model-maker using a comment. Walter suggested to make it optional: either Net and Symbol_* OR NFNF/NNFF. Randy noted that mentions of both IEEE and Gonzalez are in IBIS 8.0 BIRDs. For Example 3, Michael asked about the meaning of the : character. Walter replied that it's a range. Weston added that, for Physical, the usage is XYZ. Group is a bus label, or a user-defined label. Walter noted that a probe may have three terminals on it. Imagine two probes, with each having a ground that attaches to a local point close to the probe point in that area (a coplanar probe). Colon is a separator in this instance. The period (".") is for a floating-point value (with leading or following digits omitted). "Top" is a layer name in this case and is not reserved. This is relative to a physical layout. "Physical" is context-dependent and not an arbitrary string. Weston commented that this is useful Michael noted that the document needs to include non-Port subparameters as part of examples (e.g., IBIS_file). Walter added that some context is set by IBIS_file, EMD_File, etc. He noted that items missing include the Layout_file and Schematic_file. Michael asked whether the document should use these to set context for Physical identifier. Arpad replied that they are optional, and therefore could be omitted. Physical should not be allowed if these file parameters are not present. Michael noted that Group and Reference are used in the same example. Can they be illustrated separately? Walter replied that they can only be used together. Group can be a bus label. Arpad asked whether it could be a physical pin. Walter replied that it could be. Arpad asked whether shouldn't we have + Physical and - Physical, for polarities. Walter suggested adjustments to Example 3 to show alternate Reference grouping schemes. Arpad moved to adjourn the meeting. Walter seconded. The meeting adjourned without objection. The next meeting will be held September 11, 2024. ================================================================================ 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 Tabled ARs: - Arpad to give an example of the physical connectivity needed for EMD automation.