IBIS Macromodel Task Group Meeting date: 20 December 2022 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: - Regularly scheduled upcoming meetings that are cancelled: - December 27, 2022 - Holidays - January 03, 2023 - Holidays - January 31, 2023 - week of DesignCon ------------- Review of ARs: - Kinger to send out draft6 of the SPIM BIRD containing changes reviewed and discussed at the last meeting. - Done. Arpad had responded with more feedback, and discussion continues. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Arpad asked for any comments or corrections to the minutes of the December 13th meeting. Michael moved to approve the minutes. Ambrish seconded the motion. There were no objections. -------------- New Discussion: Standard Power Integrity Model (SPIM) BIRD draft: The group continued reviewing draft6 of Kinger's proposal and additional comments from others. Kinger thanked Bob, Randy, and Arpad for all the time they had invested in reviewing the proposal. He suggested adding Bob, Randy and Arpad as requestors. Arpad said he didn't object to being added to the list, but he didn't need to be added to the list based solely on providing comments. Kinger said the guidance had been important. Bob and Randy agreed with Arpad that the requestors are usually the original authors. Arpad suggested that any decision about the final list of requestors could be made later when the proposal is finalized. Kinger said anyone interested in being added to the requestors list should let him know. Arpad reviewed a point raised by Bob and Randy at the previous meeting. He said much of what was in the Introduction section of the BIRD should be moved into the Proposed Changes section to become the SPIM introduction portion of the IBIS specification itself. Introduction Section: Kinger mentioned that he had initially agreed with a suggestion from Arpad that SPIM could support a description of the PDN in IBIS-ISS subcircuit format as well as Touchstone S-parameters. Kinger said he still agreed with the addition of IBIS-ISS syntax support for a PDN description in resistance network format for use with DC IR calculations. However, for the AC impedance analysis support, he said he thought we should go back to language that only specifies support for Touchstone S-parameters. He said he thought existing terminal definition conventions in File_IBIS-ISS, for example the use of a single global reference terminal, were potentially in conflict with the goals and conventions of [SPIM Port List], which relies on local reference ports and the ability to cluster individual pins into ports. Bob expressed concern about the "Touchstone S-parameters" language. He said nothing in IBIS currently restricts the Touchstone file's contents to S-parameters. He suggested language such as, "Touchstone file, typically using S-parameters". Bob said the ibischk parser is unable to check to confirm that any included File_TS contains S-parameters as opposed to any other parameter type. Bob also said he wasn't sure that File_IBIS-ISS terminal conventions were necessarily incompatible with SPIM, but he would have to think about it more. Kinger said he would generalize port number specifications in the general discussion, for example OB_Stimulus_<1:8> should instead be OB_Stimulus_<1:N>, per a comment from Arpad. Arpad agreed but noted that specific numeric values can be maintained in the examples themselves. Arpad suggested that the image for figure 3 was dense and hard to read. He said we might want to change it to a white background instead of black background. Kinger agreed. With respect to Figure 4 and equations 1 through 7, Arpad noted the discussion of N+M ports (N stimulus and M observation). However, he asked whether this N+M failed to include the ports necessary to connect to the VRM. He said the equations might be correct for all the non-VRM ports, but he asked whether the text (above the equations) discussing the N+M port Z-parameter model should also include an additional number of ports for the VRM connections. Kinger said this was a good point. He said he would think about it, and he suggested a possible solution stating, "After connecting the VRM model, we have N+M ports." Solution Requirements Section: Kinger and Arpad discussed the third paragraph of the Solution Requirements. The second sentence, in particular, will need further editing for clarity. Bob said that the tree structure section should remain in text form for the time being. It's important documentation in that it gives a quick visual summary of the rules and hierarchy of the keywords. It's easier to modify in its current text form, and eventually it will be converted to the form used in section 3.3 of IBIS 7.1 and added to the Proposed Changes. Proposed Changes Section: Arpad had commented about the text appearing after the [End Chip SPIM Group] keyword. This text provides the introduction and usage rules for .spim files. Arpad said the first sentence describing the "local" (within the file) scope of SPIM definitions occurring within a .ibs file was okay, but he noted that we have the additional concept of [Component] scoping, and a .ibs file can contain multiple [Component]s. He asked whether we should discuss [Component] scoping in this section as well. Randy said the text in this section had largely been copied from the existing Interconnect Modeling section. Arpad asked whether the term "Chip" had ever been used elsewhere in IBIS. He said this might be a source of confusion. Kinger said that he originally thought of "Chip" as the packaged silicon composed of everything from the BGA up to and including the silicon. Arpad said that a [Package Model] is scoped within a [Component]. Randy asked whether Arpad thought [Component SPIM Group] would be better than [Chip SPIM Group]. Kinger then said that while he originally thought of SPIM as a model of packaged silicon to be delivered to a board designer, a SPIM model could just as easily be made for a module like a DIMM. So Component probably wouldn't be a good choice. Bob suggested something more generic such as [Power Distribution Group]. Michael suggested "Device" instead of "Chip". Arpad and Kinger thought "Device" might work, and Randy pointed out that there are only two keywords (and their end keywords) whose names include "Chip", so it wouldn't be a big change. Randy and Arpad suggested that we add language to clarify what "Device" could encompass. Kinger said that new usage rules for .spim files were modified clones of the .ims file usage rules. Kinger noted that one section describing File_IBIS-ISS, for which Arpad had suggested some changes, was an exact copy of existing text in the specification. Arpad said he hadn't realized this, and he agreed that we should maintain consistency with the existing language. Arpad wished everyone Happy Holidays and thanked them for their work on IBIS in 2022. - Ambrish: Motion to adjourn. - Curtis: Second. - Arpad: Thank you all for joining. AR: Kinger to send out draft7 of the SPIM BIRD containing changes reviewed and discussed in today's meeting. ------------- Next meeting: 10 January 2023 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives