====================================================================== IBIS FUTURES/COOKBOOK SUBCOMMITTEE MEETING MINUTES Date: September 8, 2005 Attendees: ---------- Cadence Design Systems - *Shangli Wu, *Lance Wang Cisco - *Zhiping Yang, *Vinu Arumugham Intel - *Michael Mirmak, *Arpad Muranyi Mentor Graphics - *John Angulo Micron - *Randy Wolff NCSU - *Ambrish Varma Sigrity - *Sam Chitwood Teraspeed - *Bob Ross ====================================================================== Next Meeting: Thursday, September 15, 2005 9 AM - 11 AM US Pacific Time Telephone Bridge Passcode 916-356-2663 1 724-9145 Agenda: 9 - 9:05 AM Opens 9:05 - 9:10 AM Cookbook Checkpoint 9:10 - 10:15 AM BIRD95 Discussion - Problems with C_comp Splitting (Mirmak) - C_comp Splitting Concerns (Wang, Wu) - Follow up on parasitics experiments 10:15 - 11:00 AM BIRD100 Followup ======================================================================== COOKBOOK --------- Michael Mirmak called for comments on Draft 7. None was offered. FUTURES --------- Michael reviewed his slides on BIRD100 issues. He particularly pointed out difficulties with interfacing [Pin Mapping] with his new syntax. Arpad Muranyi inquired how many buffers were supported by the new instance_name and pad syntax. Michael responded that no particular limit is enforced. Michael noted that the specification, as written, does not direct that [Pin Mapping] describe connections at the pin or at the internals/pads. Bob Ross and Arpad both pointed out that [Pin Mapping] connections within the device were NOT intended, even if not stated so. The team agreed that the fundamental choice for the future is whether [Pin Mapping] is used for all IBIS options or none at all. A partial mixture is not desirable. Arpad additionally noted that he had assumed POWER and GND pins would be directly connected to the buffer supplies. Michael responded by asking how the tool would know which supplies would connect to which buffer, particularly for [... Reference] keywords. Arpad suggested that the "tool would just miraculously" connect the supplies and references to the proper pin. Bob added that a direct short is assumed. Arpad noted his opposition to direct shorting of pins to power rails through [Pin Mapping]; package parasitics should apply. Michael added that [External Model] and [Pin Mapping] connections are not clear. How should A_puref connect to the pullup_ref column in [Pin Mapping]? John Angulo suggested that this connection is simply assumed. Bob confirmed that this was the fundamental assumption. Further, Arpad suggested that forks/joins within [External Model] are prohibited because [External Model] "lives" under [Model]. Michael concluded with a list of proposed directions for BIRD100, so that actual text could be written for Open Forum consideration and discussion. Bob opposed including an ICM parser within the IBIS parser. Arpad suggested that an ICM parser could scan the IBIS file linked to find pins, nodes, instances and ports required. This would require some significant rewriting of the ICM parser. Lance Wang briefly reviewed some calculations on how C_comp splitting might be taken into account in BIRD95. Zhiping Yang noted he disagrees with the written equations on current, if we are assuming current goes outside the device. Bob also expressed skepticism, suggesting the sign convention is inconsistent in the slides. Time ran out for further discussion. ARs ----- John Angulo - analyze what's being proposed, after Mentor discussion Sam Chitwood - analyze BIRD100 proposals, after Sigrity discussion Lance Wang - analyze BIRD100 proposals, after Cadence discussion