====================================================================== IBIS INTERCONNECT MODELING AD HOC TASK GROUP MEETING MINUTES AND AGENDA http://www.eda.org/ibis/adhoc/interconnect/ Mailing list: ibis-interconn@freelists.org ====================================================================== Next Meeting Wednesday, February 10, 2010 (No meeting on February 3) 9 AM US Pacific Time Telephone Bridge Passcode 916-356-2663 5 115-4363 (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 - TBD - Sparse Matrix Mapping proposal - Draft 10 review - Other topics to be determined ====================================================================== Minutes from January 27, 2010: 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, Raymond Y. Chen, Tao Su, Brad Brim* SiSoft - Walter Katz* Teraspeed Consulting Group - Bob Ross* ======================================================================== No patents were announced. Opens - None SPARSE MATRIX DRAFT 9 Walter Katz brought up Draft 9, and the changes were accepted. Bob Ross mentioned in the two examples that the argument for [Number of Sparse Labels] needed to be on the same line, and Draft 10 will have that change. SPARSE MATRIX LABEL RULES Bob Ross introduced the nomenclature for integer-labels by stating that they are restricted in the document to sequential integers. Several people wanted to generalize this into any string of characters. Brad Brim had shown this in previous presentations, and Bob showed Brad's e-mail arguments and his presentation to illustrate this choice. Meeting particpants expressed their views, and most favored character string labels or were neutral but leaning toward character strings. Radek Biernacki had also favored character strings or just using the colon character in past discussions. Bob and others indicated that there were no technical concerns regarding either choice. The case for using integers was that the integer provided a direct mapping to the network data complex column containing the data pair. This mapping can still be done by using the position of the character labels. The consensus was to accept character strings (or null character) along with the colon ":" at the end as a valid label. Since Brad advocated character strings, he accepted an AR to continue editing Draft 10 to revise the label rules, currently based on integers and integer-labels. Bob proposed changing the name "integer-label" tp "sparse-label" in the document. This had mentioned in previous meetings, and the name "sparse-label" follows more naturally as an argument associated with the keyword [Number of Sparse Labels]. The name "sparse-label" was accepted as a general editorial change to Draft 10 for Brad to make. Another of Brad's tasks is to remove existing rules based on integers or sequencing of integers. We discussed briefly other rules such as: A "sparse-label" is any text string that ends with a colon character. The colon character can serve by itself as a label (corresponding to a null character string before the label. No spaces (or other separators) are allowed for sparse-labels We may decide later to provided restrictions on the label content such as alpha-nunmerical characters only along with dash and underbar, but we also noted that the ending colon of a text string could be sufficient identification under the [Sparse Matrix Mapping] keyword. So other useful characters including commas and paranthesis could be accepted. No rules are needed concernging repeated label names since the position of the label under [Sparse Matrix Mapping] is the only information needed or processed by a tool. This also supports using just the colon character by itself as a label. Brad plans to edit Draft 10 and create rules for sparse-labels based on our discussion so far. This will serve as a base-line for further discussion. AR - Brad provide Draft 10 with markup to Bob with character-based sparse-labels for uploading/distribution before the next meeting. BINARY FORMAT Greg Edlund had sent a binary format example to the Interconnect Task Group mailing list, and Bob confirmed that all attendees received it. Walter commented and confirmed with Greg Edlund that the binary file should have also been named microstrip.s4p. For binary format, the number of ports needs to be known in order to process the binary content. The easiest way to do this in Touchstone 1.0 is to require the existing extension option of .sp when the file has binary content. A few more questions and suggestions were discussed including allowing the "0" following the binary command to be replaced with another character or string for big-endian/little-endian information. Bob also mentioned that the number of ports could be considered here, but this would change the structure of Touchstone. In response to Radek's question, the binary format cannot be used for a file that contains noise data information. Because the noise data addition is restricted to 2-port files, this is not a serious limitation. Walter mentioned that a publicly avaiable reader/writer similar to tschk2 or ibischk5 would be helpful. We will continue discussing the binary format and eventually generate a proposal (or TSIRD). NEXT MEETING Because most people will be attending DesignCon 2010 next week, the next Interconnect Task Group meeting is scheduled on Wednesday, February 10, 2010. Draft 10 will be reviewed, focusing mostly on the new sparse-label content.