IBIS Macromodel Task Group Meeting date: 28 February 2023 Members (asterisk for those attending): Achronix Semiconductor: Hansel Dsilva Amazon: John Yan ANSYS: * Curtis Clark * Wei-hsing Huang Aurora Systems: * Dian Yang Cadence Design Systems: Ambrish Varma * Jared James Google: Hanfeng Wang GaWon Kim Intel: Michael Mirmak * Kinger Cai Chi-te Chen Alaeddin Aydiner Keysight Technologies: Fangyi Rao Majid Ahadi Dolatsara Ming Yan Radek Biernacki Rui Yang Luminous Computing David Banas Marvell Steve Parker Mathworks (SiSoft): Walter Katz Mike LaBonte Micron Technology: * Randy Wolff Justin Butterfield Missouri S&T Chulsoon Hwang Yifan Ding Rivos Yansheng Wang SAE ITC Michael McNair Siemens EDA (Mentor): * Arpad Muranyi Teraspeed Labs: * Bob Ross Waymo: Zhiping Yang Zuken USA: * Lance Wang The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: - Kinger to send out draft7 of the PSIJ Sensitivity BIRD containing changes reviewed and discussed in the February 21 meeting. - Done. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Arpad asked for any comments or corrections to the minutes of the February 21st meeting. Kinger moved to approve the minutes. Dian seconded the motion. There were no objections. -------------- New Discussion: Standard Power Integrity Model (SPIM) BIRD draft: Kinger reported that he had received no new feedback on draft9. Arpad asked whether the 40-character limit specified for various file name quantities was consistent with what was done in IBIS 7.2. Bob, Randy and Arpad said the 40 character limits defined for keywords that contain names, e.g., [Device SPIM Group] were consistent with existing usage. However, for file name keywords and subparameters, e.g., the File_TS subparameter of [SPIM Touchstone File], the 40-character limit should not be explicitly stated. Instead, as is done in IBIS 7.2, the file name quantities should rely on the global definition provided in Sections 3.1 and 3.2. Kinger said he would remove the explicit 40-character limit from the file name related quantities in the BIRD. Bob suggested that the file names in the Example sections be shortened to make the examples easier to read. PSIJ Sensitivity BIRD draft: Kinger and the group reviewed draft7. Kinger noted that the [PSIJ Sensitivity Group] Description and Usage Rules had been updated to incorporate language consistent with that used in [Interconnect Model Group] sections, per a suggestion from Arpad and Randy at the previous meeting. Bob expressed concern that name of the keyword [PSIJ Sensitivity Group] itself could cause confusion. He said in other contexts we use "group" keywords to reference groups of "sets". Arpad asked if naming it [PSIJ Sensitivity Set] would be more consistent with other usage in the specification. Randy agreed that it was more like a set, but he said he wasn't sure we'd want to use set in this context either. Randy asked if a [PSIJ Sensitivity Group] name would always correspond to a particular mode of operation (e.g., PCIe Gen4 or PCIe Gen5). Kinger said yes. He said, for example, a device might have three power rails. Different groups could specify the PSIJ Sensitivity for the rails differently for each mode. Bob said that [PSIJ Sensitivity Group] is essentially a selector. Kinger said he was open to suggestions for alternative names for [PSIJ Sensitivity] and [PSIJ Sensitivity Group]. Bob and Arpad said we might have to think about where the [PSIJ Voltage List] exists in the keyword hierarchy. Is it scoped by [Component] or within one of the PSIJ Sensitivity keywords? Kinger reviewed a presentation on PSIJ Sensitivity from last year's Asian IBIS Summits, and he described the overall goal. He said that traditionally SI and PI design and optimization had been done separately, and timing budgets for the SI simulation and noise budgets for the PI simulation were used independently. He said when systems were slower and the unit interval was longer, these types of targets and budgets could be defined and met at a reasonable cost. Now, however, with higher speed devices, designing systems based on simplistic noise and timing budgets can be very expensive or impossible. We need a more holistic approach to SI/PI design. SI/PI cosimulation would be ideal, but incorporating transistor level models, full PDN models, and an SI simulation sufficient to determine low BERs is almost impossibly time consuming. With the SPIM BIRD and the PSIJ Sensitivity BIRD, the goal is to bring PI and SI simulation closer together. The SPIM BIRD supports AC analysis and DC IR drop analysis for full PDN design and verification. In the future, the goal is to add SPIM model data to support transient simulation. At that point, SPIM could also help simulators compute the power rail noise in a transient analysis. The power rail noise from the transient PI simulation can then be combined with the PSIJ Sensitivity data to compute Jitter to be applied in the SI simulation. - Curtis: Motion to adjourn. - Randy: Second. - Arpad: Thank you all for joining. AR: Kinger to send out draft10 of the SPIM BIRD containing changes reviewed and discussed in today's meeting. AR: Kinger to send out draft8 of the PSIJ Sensitivity BIRD containing changes reviewed and discussed in today's meeting. ------------- Next meeting: 07 March 2023 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives