====================================================================== IBIS INTERCONNECT MODELING TASK GROUP MEETING MINUTES AND AGENDA http://www.eda.org/ibis/adhoc/interconnect/ Mailing list: ibis-interconn@xxxxxxxxxxxxx ====================================================================== Next Meeting Wednesday, August 18, 2010 9 AM US Pacific Time Telephone Bridge Passcode 916-356-2663 4 948-5829 (for international and alternate US numbers, contact Michael Mirmak) LiveMeeting: https://webjoin.intel.com/?passcode=9485829 Agenda: - Attendees - Call for patents - Opens - MCP Documents Review ====================================================================== Minutes from August 11: Attendees: ---------- (* denotes present) Agilent - Radek Biernacki*, John Moore, Ken Wong Ansoft - Denis Soldo Cadence Design Systems - Terry Jernberg, Brad Griffin Cisco Systems - Mike LaBonte Green Streak Programs - Lynne Green Hewlett-Packard - Rob Elliott IBM - Greg Edlund ICT-Lanto - Steven Wong Intel - Michael Mirmak* Mentor Graphics Corp. - Arpad Muranyi*, John Angulo*, Vladimir Dmitriev-Zdorov Micron Technology - Randy Wolff*, Justin Butterfield* Sigrity - Sam Chitwood, Raymond Y. Chen, Tao Su, Brad Brim* SiSoft - Walter Katz* Teraspeed Consulting Group - Bob Ross* ======================================================================== No patents were declared. Bob Ross raised some brief comments about Touchstone 2.0, suggesting that the team consider releasing 2.1 once binary support is complete. He suggested that the original parser developer may shortly not be available. He added that any work be based only on current TSIRDs. Michael Mirmak asked whether any other changes were likely or desired, as only one TSIRD is now under consideration (binary). Brad Brim suggested a pole-zero format. Both Walter Katz and Bob Ross responded that this might best be addressed in a separate specification. Bob added that an additional delay term would be needed, while Walter suggested that optional data organization by framis be added. No objections were raised to the plan to distribute a new parser after the Open Forum vote on TSIRD2.1. Michael Mirmak stated that, per a decision in the IBIS-ATM Task Group, the IBIS-ISS specification would be discussed in the IBIS Interconnect Task Group from this point on. Meetings will alternate between IBIS-ISS and MCP development. Arpad Muranyi thanked the attendees. He also suggested that crosslinking of the IBIS-ISS documentation Between the ATM and Interconnect links be maintained. The team reviewed the IBIS-ISS rev. 0.5 document. Bob Ross asked whether the V-element was removed. Michael indicated not, but that a short element used a v-element with DC voltage set to 0. No other independent sources exist in the specification. The team discussed whether netlists per se are to be supported in ISS. Radek Biernacki asked whether the entire text of an ISS file is to be always encapsulated within a subckt. Michael responded that, in this version, not; other ISS files could be included without encapsulation. Radek noted that this might be dangerous due to node conflicts. Arpad noted that the IBIS-ISS specification is a subcircuit specification, not a full netlist specification, and everything was to be encapsulated in a subcircuit. Walter added that he assumed only file contents are subcircuits and that included files are flattened. Michael posed a simple question: if non-ISS files can include ISS content, can ISS files include non-ISS content? Walter answered no: ISS cannot include non-ISS. Contents must be flattened before parsing. Brad added that at some pint down in a hierarchy, the user can't include non-ISS content. Arpad agreed that non-ISS content shouldn't be allowed, because that would allow people to include transistor models too, which is not something we want to support. Michael suggested this is like IBIS calling a package file, though Bob suggested flattening changes this approach. Walter noted that parser command line options could select flattening or no. Arpad noted that the ideal source with S-parameter style analog buffer models used for AMI channel characterization could be modeled by IBIS-ISS this way. Bob suggested that the team follow the SPICE conventions of flattening the entire netlist before parsing. Walter, Radek and Randy Wolff all agreed that any included items must be ISS-compliant. Bob added that the included file does not have to be complete (encapsulated) but simply has to contain legal ISS syntax. Michael asked whether the team was calling all "ISS-legal" files the same as (encapsulated) "ISS-complete" files. Walter proposed that this was similar to having table data in separate files that were not self-contained. Per the discussion, Michael suggested that only the top-level calling file be "ISS-complete" but that any included file must contain only parsable, legal syntax. Walter added that documentation must distinguish between the single quote ` vs. open/close quote ' (not an accent). Michael asked whether passing parameters into subcircuits outside of ISS bounds was acceptable. Does this force other files to be ISS-compliant, at least in part? Walter suggested this was true. Arpad added that default subcircuit parameter names must be defined for the subcircuit definition. Michael responded That this effectively locks the parameter hierarchy to be local instead of global. Arpad agreed, due to the risk Of conflicts. Bob added that we note this implication at the end of the document. Walter suggested that the document can support either .subckt line parameter definitions or definitions of parameters within subcircuit bodies. Michael will show and add "parhier" option examples to the document. Walter noted that writing the document was hampered by how difficult it was to get a good explanation for W-elements based on table models out of the original documents. This area in ISS is weak and needs fixing. Radek added that W- elements should be a focus of the document. For next time, the MCP discussions will pick up again. Brad Brim will provide skeleton of an MCP spec using examples; He has key documents but wants time to clean them up before distribution. The files do not contain a complete subckt example, though the outline of one is presented. ------------------------------------------------------------------ The IBIS Ad Hoc Interconnect Task Group Mailing List Archives are available at: http://www.freelists.org/archives/ibis-interconn TO UNSUBSCRIBE: Send a message to "ibis-interconn-request@xxxxxxxxxxxxx" with a subject of "unsubscribe" To administer your subscription status from the web, visit: http://www.freelists.org/list/ibis-interconn