====================================================================== IBIS INTERCONNECT MODELING TASK GROUP MEETING MINUTES AND AGENDA http://www.eda.org/ibis/adhoc/interconnect/ Mailing list: ibis-interconn@freelists.org ====================================================================== Next Meeting Wednesday, March 24, 2010 9 AM US Pacific Time Telephone Bridge Passcode 916-356-2663 5 858-5870 (for international and alternate US numbers, contact Michael Mirmak) Live Meeting: http://tinyurl.com/yzo768q or: https://www.livemeeting.com/cc/sisoft/join?id=M462B6&role=attend&pw=P9%3D8%3BP%277b Agenda: - Attendence - Call for patents - Opens - Model Connection Protocol - Binary syntax proposal - April 8, 2010 Meeting?? ====================================================================== Minutes from March 17, 2010: TO BE FILLED IN Attendees: ---------- (* denotes present) Agilent - Radek Biernacki, John Moore, Ken Wong Ansoft - Denis Soldo Cadence Design Systems - Terry Jernberg, Brad Griffin Green Streak Programs - Lynne Green Hewlett-Packard - Rob Elliott IBM - Greg Edlund* Intel - Michael Mirmak Mentor Graphics Corp. - John Angulo*, Vladimir Dmitriev-Zdorov Micron Technology - Randy Wolff* Sigrity - Sam Chitwood, Brad Brim* SiSoft - Walter Katz* Teraspeed Consulting Group - Bob Ross* ======================================================================== No patents were announced. Opens - None BINARY ENCODING PROPOSAL Due to the fact that Walter Katz needed to leave early, we started with the Binary encoding proposal first. Greg Edlund introduced the draft document (designated here as tsbin1_draft1.pdf. Greg used pages from the tsird.doc as an initial template: http://www.eda.org/ibis/adhoc/interconnect/tsbin1_draft1.pdf Draft1 contained the basic binary encoding proposal with a two character argument BINARY as originally implemented by IBM where is D, F, or M for the precision selections, and is B or L for big/little endian choices. The draft showed one example for a Touchstone 1.0 type format with and without binary encoding. Several questions were asked, and also a private response from Walter was reviewed. Brad Brim asked whether this official format should be restricted to just Touchstone 2.1 files and above. There exists widespread of the existing Touchstone format, so the binary option within would just create industry incompatibilites. By including this in just Touchstone 2.1, the new feature would promote adoption of Touchstone 2.1 features and be consistent and have the binary option as part included. Another question is whether the originator of Touchstone has any concerns about bringing in the binary option at the lower level. Bob Ross commented that IBM could keep its original implimetation and also support a Touchstone 2.1 version with only a minor inconvenience. Brad also commented that we might consider a BINARY syntax where would be for frequency precision (such as when used locating a reasonance), for data precision, and for big/little endium. Greg agreed that this might be good choice because this eliminates "M" as a combined selection and keeps the options more general. He will ask if this is a concern within IBM. From Walter's list: ------------ 1. Some words need to be added regarding s2p and noise data a. I think Bob should do this because it needs to be coordinated with other changes to the Touchstone 2 document 2. Some words need to be added to indicate that the [End] statement is no to be added at the end of Touchstone 2 files a. Again Bob should make those changes when this gets integrated into Touchstone 2 3. I think the files (particularly the Binary file) needs to be a separate file referenced by this document 4. We already agreed to not supporting 128 bit floating point now, and that we could add it later if ever required. ----------- Regarding 1) and 2), we agreed to work with Greg Edlund to add the [Noise Data] limitation and state that [End] is not required for the Binary encoding. We clarified item 3) that Walter was proposing another example and not linkage to a separate file. The binary example still has hidden characters, and an alternative form might be displayed with a readable hex format. That way the one pad "byte 0" can be made visible in the specification to signal the start the binary encoding. Per the discussion at a previous meeting, we agreed to hold off on the 128 bit floating point expansion at this time in item 4). Brad's three-argument suggestion might be better for the future floating point selections.. Greg commented that the % BINARY sequence was a mistake. The % could be replaced by another character. Bob Ross stated that we have to pick something other than # and we might even consider making this a keyword. We will continue to discuss and evolve the binary format in future meetings. MODEL CONNECTION PROTOCOL (MCP) Brad first introduced the topic by reviewing a few slides from earlier presentations. In doing so he pointed out a few mapping issues such as pin names not exactly matching other pin name, and the physical locations of one device might have a different reference or have flipped coordinats for another device. Brad then introduced a draft specification. It was not ready yet to be released to the Task Group. Brad showed some suggestions for additions. Some of these might not be introduced yet. Because this is a preliminary overview, a lot of details are not recorded here. The first draft of the document proposal is expected to be made available prior to the next meeting after another editing pass. NEXT MEETING Model Connection Protocol and the binary encoding section are planned for the next meeting.