IBIS Macromodel Task Group Meeting date: 26 September 2023 Members (asterisk for those attending): Achronix Semiconductor: Hansel Dsilva Amazon: John Yan ANSYS: * Curtis Clark * Wei-hsing Huang Aurora System: Dian Yang Cadence Design Systems: * Ambrish Varma Jared James Google: Hanfeng Wang GaWon Kim Intel: * Michael Mirmak Kinger Cai Chi-te Chen Liwei Zhao Keysight Technologies: Fangyi Rao Majid Ahadi Dolatsara Stephen Slater Ming Yan Rui Yang Marvell: Steve Parker Mathworks (SiSoft): Walter Katz Graham Kus Micron Technology: Justin Butterfield Missouri S&T: Chulsoon Hwang Yifan Ding Zhiping Yang Rivos: Yansheng Wang SAE ITC: Michael McNair Siemens EDA (Mentor): * Arpad Muranyi * Randy Wolff Teraspeed Labs: * Bob Ross Zuken USA: * Lance Wang The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: Kinger: Provide BIRD223.1 draft2 with minor cleanup changes to Lance to post to the IBIS Open Forum BIRDs page as BIRD223.1. - Done. Michael: Send out draft8 of the [AMI Test Data] proposal including more feedback from the ATM meeting. - Done. Michael: Develop a full syntactically complete example demonstrating the [AMI Test Data] proposal. - In progress. Michael said he would work on this after we get the proposal itself more settled. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Arpad asked for any comments or corrections to the minutes of the September 19th meeting. Ambrish moved to approve the minutes. Michael seconded the motion. There were no objections. -------------- New Discussion: [AMI Test Data] proposal: Michael reviewed draft8, which had been sent to the ATM list prior to the meeting. He noted that the major changes were related to the waveform file Sub-Params, which are now Input_impulse_File, Input_waveform_File, and Golden_waveform_file. Michael noted that the proposal explicitly states that Input_impulse_file is required for all combinations of Tx or Rx and statistical or time domain. Michael said that Input_waveform_file is the time domain replacement for the original Stimulus_file (now removed), which had been overloaded and used for statistical and time domain cases. Input_waveform_file is required for time domain and illegal for statistical configurations. Michael noted that the Input_waveform_file definition describes the contents of the file for the Tx and Rx cases. For a Tx, the contents are a sampled waveform representing the "digital" state transitions. For an Rx, the contents are an analog waveform representing a stimulus convolved with the channel impulse response. Michael said that he had not created a table to describe the waveform file subparameters' contents for the combinations of Tx or Rx and statistical or time domain. He said he thought that with the new separate "impulse" and "waveform" input file subparameters, the descriptive text was now clear. He asked whether anyone thought a table was necessary. No one said they thought a table was necessary. Michael noted that this version explicitly states that the changes should be added after Section 10.4. Therefore, the changes will occur after the function signature descriptions upon which they rely. He said the changes will occur before advanced topics such as retimer flows, and they should contain no references to anything defined later. Michael asked whether everyone thought it was clear that this proposal applies to individual [Model]s, and it does not require a full Tx, channel, Rx, setup. Ambrish said that Input_waveform_file is analogous to the "wave" parameter in the AMI_GetWave function signature. He suggested that if the text made that connection to "wave", then it wouldn't have to provide its own explanation of the contents of Input_waveform_file. Michael agreed, and the group adopted language stating that the contents of Input_waveform_file are "equivalent" to the contents of the "wave" argument. Arpad suggested that we make it clear to the reader that there's no need for a Tx when testing an Rx model with this proposal. In response, the group added text stating that the contents of the Input_waveform_file for an Rx were equivalent to the "wave" input that would be presented to the Rx AMI_GetWave during a normal channel simulation that included Tx and Rx AMI executable models. Bob asked whether it was a good idea to include IBIS keyword descriptions in the AMI section. Michael said the new keywords relied on information about AMI function parameters and AMI Reserved Parameters, so the keyword definitions had to come after those sections. Ambrish said this raised an interesting point. Since this proposal is concerned with testing AMI models, perhaps it should be placed at the end of the AMI section, i.e., the testing portion comes last. Michael said there were two primary reasons he had introduced this proposal in the middle of the AMI chapter. First, the last subsection of the AMI chapter is just a list of all the AMI Reserved Parameters, and he thought these changes should go before that. Second, he had been frustrated at having to jump around to multiple places in IBIS 7.2 in order to fully understand and assemble summaries of the information upon which his proposal depended. Therefore, he had attempted to put the new section as close as possible to the sections on which it relied. Michael noted, however, that it may make sense to include this proposal a bit later than 10.4, so it would appear after all the flow discussions. Curtis noted some language about "number of columns" and aggressors that was not relevant to Input_waveform_file. Michael agreed that these were hold overs from the original Stimulus_file discussion and removed them. Ambrish asked whether the name of the [Waveform] keyword was descriptive enough. Michael suggested [AMI Test Configuration] instead. The group agreed to change [Waveform] to [AMI Test Configuration]. Bob suggested that a keyword hierarchy tree entry should be added for these new keywords. He also asked whether we should consider adding the associated "end" keywords (e.g., [End AMI Test Configuration]). Bob said that we might want a new table (or to add to an existing table) to define all the rules for what subparameters are required, illegal, optional under what conditions. Michael said he would add a table with this information. Michael said he would send out draft9. - Ambrish: Motion to adjourn. - Curtis: Second. - Arpad: Thank you all for joining. New ARs: Michael: Send out draft9 of the [AMI Test Data] proposal including more feedback from the ATM meeting. ------------- Next meeting: 03 October 2023 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives