====================================================================== IBIS INTERCONNECT TASK GROUP MEETING MINUTES AND AGENDA http://www.eda.org/ibis/interconnect_wip/ (note new URL) Mailing list: ibis-interconn@freelists.org ====================================================================== Next Meeting Wednesday, January 19, 2011 9 AM US Pacific Time Telephone Bridge Passcode 916-356-2663 5 689-5118 (for international and alternate US numbers, contact Michael Mirmak) LiveMeeting: https://webjoin.intel.com/?passcode=6895118 Agenda: - Attendees - Call for patents - Review of Minutes - Opens - IBIS-ISS Draft 0.9 Review - MCP Discussion POSTPONED UNTIL NEXT MEETING (missing presenter) - Touchstone 2.x Status/Plans - Agenda for Next Meeting - Approve Next Meeting Dates: * Jan 26 * Feb. 2 (DesignCon conflict?) * Feb. 9 (potential for cancellation/alternate chair?) ====================================================================== Attendees: ---------- (* denotes present) Agilent - Radek Biernacki* Ansoft - Denis Soldo Cadence Design Systems - Terry Jernberg, Brad Griffin, Dennis Nagle, Taranjit Kukal Cisco Systems - Mike LaBonte* IBM - Greg Edlund ICT-Lanto - Steven Wong Intel - Michael Mirmak* Mentor Graphics Corp. - John Angulo*, Arpad Muranyi* Micron Technology - Justin Butterfield*, Randy Wolff* Sigrity - Brad Brim* SiSoft - Walter Katz* Teraspeed Consulting Group - Bob Ross* ======================================================================== No patents were declared. During opens, Arpad Muranyi mentioned some comments relayed to him from Lynne Green. These suggested having IBIS-ISS include more text noting its Derivation from HSPICE, to ensure no confusion regarding other SPICEs. The use of the term "reserved" regarding certain operators caused concern. Michael Mirmak summarized IBIS-ISS reserved words treatment. The reserved operators are not meant for future use but are instead disallowed to avoid conflicts with proprietary SPICEs when IBIS-ISS does not support the named functions. The next version will clarify the restrictions and remove the term "reserved". Arpad suggested that IBIS-ISS be finalized in time for presentation at the DesignCon IBIS summit. Arpad raised the question of wp and field solvers, which Michael clarified. Does the description regarding assumptions/requirements on W-elements in IBIS-ISS imply that SPICE tools must implement field solver functions? Radek Biernacki responded that no field solver requirement is implied and that no static R and G matrix is supported. Those tools that support the frequency-dependent R and G matrices should be able to support IBIS-ISS syntax without issue. Michael shared several syntax updates, including those posted on the e-mail reflector. These covered the ".include" and ".param" keywords, plus how single- and double-backslashes were intended to be implemented. At present, single-backslash support has been removed from the pending IBIS-ISS draft; further, $-delimited comments are disallowed. Arpad asked whether we need to cover .incl, .inclu and other .inc/.include variants We may not want each case spelled out one at a time in the specification. Michael suggested a variation, where a single sentence will express the rule that the character strings .include, .inc and .inc plus any number of characters in the sequence "l u d e" will all be accepted as synonymous. After some discussion, the team accepted this for both .include and .parameters. Radek asked about in-line asterisk support after double backslashes for line continuation. Michael took the AR to check. Michael mentioned that lines, tokens, expressions are limited to 1024 characters. Whether this applies to all operating systems or is more limited in Microsoft Windows is unclear. Michael to check. The current IBIS-ISS line continuation rules state that single backslash is not supported. Mike LaBonte pointed out that the single backslash character not supported means illegal and is therefore an error. This will be clarified in the next version. The team discussed whether the voltage shunt (DC=0) should be maintained or expanded to support more voltages. Arpad noted that DC termination is a potential application for a non-zero voltage source, though it has been proposed for use in buffer interfacing for IBIS. Bob Ross noted that passing voltages in through external nodes/ports would be the preferred method for interconnect purposes, rather than having the interconnect model define an internal voltage. Radek suggested that, if one needs a non-zero DC source, one needs a component with which to Associate it. The notion of a restricted element parameter violates concept of parameter. Bob added that both the F and G element syntax require the name of a voltage source as a sense element (for example, dependent source gain is in terms of current through another source). Mike LaBonte noted that one cannot guarantee LTI in a system with controlled sources. The team discussed whether this was an LTI limitation or a passivity limitation. Radek suggested simply issuing a warning with non-zero sources, to keep the specification Focused on interconnects. A motion was made and seconded to enable voltage sources to take non-zero values. Michael asked for objections; no objections were raised and the change was deemed accepted. Walter Katz shortly thereafter noted that he was unable to raise an objection due to an interrupting telephone call. He added that the G element can be used as a current source. Further, agreeing with Bob, he suggested that voltage sources should be passed in through a port reference rather than supported through separate IBIS-ISS elements. Further discussions will be held during the rev. 0.9 review. Michael noted that technical changes can still be made by the IBIS Open Forum after the document has been passed to it from the IBIS Interconnect Task Group. Brad Brim accepted an AR to prepare an update on the state of MCP and proposals for future work.