IBIS Macromodel Task Group Meeting date: 02 May 2017 Members (asterisk for those attending): ANSYS: Dan Dvorscak * Curtis Clark Broadcom (Avago): Xingdong Dai * Bob Miller Cadence Design Systems: Ambrish Varma Brad Brim Kumar Keshavan * Ken Willis eASIC: * David Banas Marc Kowalski Ericsson: Anders Ekholm GlobalFoundries: Steve Parker IBM Luis Armenta Trevor Timpane Intel: * Michael Mirmak Keysight Technologies: * Fangyi Rao * Radek Biernacki Ming Yan Maxim Integrated Products: Hassan Rafat Mentor, A Siemens Business: John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff Justin Butterfield QLogic Corp.: James Zhou Andy Joy SiSoft: * Walter Katz Todd Westerhoff * Mike LaBonte Synopsys: Rita Horner Kevin Li Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: * Bob Ross TI: Alfred Chong The meeting was led by Arpad Muranyi. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: - Radek to send BIRD 158.5 to Mike L. for posting by Friday April 28th. - Done. - Walter to send BIRD 166.2 to Mike L. for posting. - Done. Bob Ross and Walter to create a BIRD 186.3 draft 2. - Done. It is actively being discussed in the Editorial Task Group meetings. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: - Arpad: Does anyone have any comments or corrections? [none] - Mike L.: Motion to approve the minutes. - Michael M.: Second. - Arpad: Anyone opposed? [none] ------------- New Discussion: BIRD 166.2 Redriver Flow: - Discussion: Arpad noted that he thought we had finished with this topic at the previous meeting. His recollection was that Walter was to create BIRD 166.2, which essentially rolled back the changes introduced in BIRD 166.1 and went back to the original BIRD 166. This could then be considered for inclusion into IBIS 7.0, and the new advanced proposal for additional IR outputs from the Tx and Rx could be added later. Curtis and Walter agreed with Arpad's summary of events, and Curtis noted that he had captured events this way in the previous week's minutes. Arpad, however, then noted that an email sent by Fangyi the morning of today's meeting had expressed concerns with BIRD 166.2. Fangyi noted that BIRD 166.2 reverted to BIRD 166, which was three years old. He noted that discussions when BIRD 166 was first introduced had come to the conclusion that it was limited. It relied on deconvolution in certain cases and this could be problematic. Walter noted that the basic redriver flow involved 4 devices (Tx1, Rx1, Tx2, Rx2). If all four of their models were Init() only, the flow in IBIS 6.1 was wrong, and the flow in BIRD 166.2 was right. Similarly, if all 4 devices' models were dual-mode (functional Init() and GetWave()), then the flow in IBIS 6.1 was wrong, and the flow in BIRD 166.2 was right. Walter noted that there were 81 possible flow combinations for a two channel redriver flow (4 devices, each with 3 possible model types: Init() only, GetWave() only, dual). Walter noted that IBIS 6.1 flows had problems (relied on convolution) even for some of the 9 flow combinations for a single Tx to Rx channel. Arpad again noted that he wanted to make sure we took an incremental approach in which any change we introduced would not have to be undone later. Radek noted that Walter's proposal was a partial improvement that was okay for many situations, but that we might have to abandon it later in favor of the final corrections. Walter disagreed and said that his approach was independent of adding the new IRs. He said that until the new IRs were added, one couldn't even start to worry about the special cases for which BIRD 166.2 had issues. Bob R. expressed concern about maintaining backward compatibility, and asked if a 7.0 model would not work in a later 7.1 flow. Walter said it would work fine because a 7.0 model would not advertise that it supported the additional IRs. Arpad said that while he liked some of the improvements in the advanced proposal from Fangyi, he had heard from Vladimir that it would take significant time for tool developers to implement the proposal. Therefore, he preferred the incremental approach (introducing BIRD 166.2 first) if possible. Fangyi said any proposed improvement should not be limited to only half the cases. Mike L. asked if BIRD 166.2 would make it harder for Fangyi to implement the advanced proposal. Fangyi said that the proposals went in different directions. Walter disagreed and said that for the basic cases his BIRD was correcting, the flow and IRs would be exactly the same in the two proposals. Walter motioned for a straw poll on whether the attendees would want ATM to recommend to the Open Forum that BIRD 166.2 be included in IBIS 7.0 (note: not a motion to make that recommendation, just a straw poll on where attendees stood on the subject). Bob Miller seconded. The results by member organization: ANSYS: yes Broadcom: abstain Cadence: abstain Intel: abstain Keysight: no Mentor: yes Micron: abstain SiSoft: yes Teraspeed: abstain Yes(3), No(1), Abstain(5) Those abstaining were interested in more information on the differences between the proposals or Fangyi's objections to BIRD 166.2. Arpad and Michael Mirmak asked if Fangyi or Walter could present a side by side comparison of proposed flows. Arpad mentioned slides 14 and 16 of this DesignCon presentation (http://www.ibis.org/summits/feb17/katz.pdf) as an example. Bob M. said that he wanted the BIRD 166.2 change introduced quickly, but he also wanted the advanced proposal. He noted that understanding exactly why Fangyi thought BIRD 166.2 would get in the way was important. Arpad, Bob M., and Walter asked Fangyi if he could demonstrate specific examples in which BIRD 166.2 would impede future progress. - Mike L.: Motion to adjourn. - Michael M.: Second. - Arpad: Thank you all for joining. AR: Walter and Fangyi to prepare information for the continued discussion of BIRD 166.2 and any effect it would have on future improvements. ------------- Next meeting: 09 May 2017 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives