====================================================================== IBIS FUTURES/COOKBOOK SUBCOMMITTEE MEETING MINUTES Date: June 2, 2005 Attendees: ---------- Cisco - Syed Huq, Zhiping Yang Green Streak Programs - Lynne Green Intel - Michael Mirmak, Arpad Muranyi Mentor Graphics - John Angulo Teraspeed - Bob Ross ====================================================================== Next Meeting: Thursday, June 16 1 PM - 3 PM US Pacific Time (tentative - Mirmak/Muranyi may not be available) Telephone Bridge Passcode 916-356-2663 2 738-7890 Agenda: 1 - 1:05 PM Opens 1:05 - 2:20 PM Futures Discussion - BIRD95 - BIRD99 - BIRD100 (?) - ICM changes update 2:20 - 3:00 PM Cookbook Discussion ====================================================================== Futures Minutes During Opens, Bob suggested that time be reserved for a discussion of DAC and Asian Summit presentations. BIRD95.5 -------- Syed and Zhiping summarized the latest proposed changed to BIRD95. Lance Wang has suggested pre-driver currents, particularly in the case of multiple buffer simulaitons, need to be included in the BIRD95 solution. In particular, non-ideal or multiple-load cases and load dependencies should be included/considered. Arpad's original SPICE model is not being used; an internal Cisco database is being simulated, but this cannot be released. Arpad suggested using an I/O buffer, not an output-only buffer, as the clamp currents could ease calculations. Zhiping will present on the multiple-buffer cases at the DAC summit. Ken WIllis and Lance, if available, will discuss algorithm options at the summit at well. The team agreed algorithm details should not be included in teh specification. Bob observed that BIRD95.5 does not include gate modulation effects. BIRD97/98 -------- Arpad summarized BIRD97 and BIRD98. He would like BIRD97 to be rejected in favor of BIRD98, which uses tables rather than a formula to represent gate modulation effects on other IBIS data. This will mask IP and be more conducive to data collection in a way familiar to the IBIS community. Bob suggests that clamp and on-die termination data in separate data tables. BIRD94 -------- No changes have been made since the last update. Bob accepted an AR to review the BIRD and provide feedback to Arpad. ICM -------- The team agreed, in general, that the ICM specification is ready for final readings/review and approval. Bob suggested that the dates in the document be updated before the next Open Forum reading. Michael mentioned a "peeve" of the use of "real" without definition; however, he did not demand an IIRD for this. Upcoming IBIS Specifications -------- Michael summarized the changes suggested by members, but not yet implemented or written as BIRDs. These include: - [Ext. Circuit] usage with [Model] (John corrected this) - ICM links to IBIS within IBIS models - user-defined measurements Bob stated that the ICM links are the most critical to include, as the 4.1 spec. strongly implies their existence. The current parser, however, does not check for these links and no specific syntax exists to support them. Bob proposed simply dropping the ICM link language from the current specification for 4.2, and reintroducing the language in 5.0. Michael suggested that BIRD authors in future specify their preference as to which version of IBIS should feature the BIRD text. This would not be binding on the Open Forum. ARs -------- Bob - review BIRD94 and provide feedback to Arpad et al. Michael/Arpad - create draft BIRD for removing ICM link implication in IBIS 4.1. ====================================================================== Cookbook Minutes Michael reviewed the latest cookbook edition, which simply "cleans up" text which was added by Arpad, specifically on differential C_comp and related issues. The team agreed that the cookbook is nearly ready to be issued. The remaining issues include correction of tables, addition of missing diagrams and some further text on split C_comp. Bob also suggested that references to "half-differential" buffers be removed or incorporated into the full- or pseudo-differential sections. Bob also mentioned removing "CAE" as an EDA tool reference. Some additional mention of the original intention of [Submodel] may be needed. The cookbook should be introduced for review within two IBIS Open Forum meetings. Michael will update the document and provide copies to the Cookbook Committee. Arpad will add the new C_comp text. Michael/Arpad - create split C_comp text and update cookbook for review.