IBIS Macromodel Task Group Meeting date: 24 October 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 Alaeddin Aydiner 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. Ambrish Varma took the minutes. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: Alaeddin: Send out draft2 of the "AMI Ignore Block" proposal including feedback from the ATM meeting - done Michael: Send out draft13 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. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Arpad asked for any comments or corrections to the minutes of the October 10th meeting. Michael moved to approve the minutes. Walter seconded the motion. There were no objections. -------------- New Discussion: AMI Test Data proposal: Michael began by thanking Arpad for doing a detailed review of draft 13 of the AMI Test Data Support BIRD and said that the email Arpad sent to the ATM raised some important questions. Michael started by asking if there is any information coming out of the Init function other than the impulse response that the model maker want to test or compare? Ambrish said that only the Impulse Response can be compared. Michael wondered if that is the case, should it be optional or required for AMI Test Data for Type Time Domain (TD). He said he would like it to be required to determine if something goes wrong. Ideally, he said, everything should be an open book. Ambrish said (and Walter concured) that this should be optional for the TD case as there could be cases when the model maker may not want to publish the IP by giving the output IR from the Init function. Arpad mentioned that the specification describes methods to extract the filter from the Init function to be used in place of a missing Getwave function in the TD flow, that is the ultimate test (to be able to compare the Getwave output) and if the model maker wants to have test data for TD, they should provide the Golden_waveform_file and may optionally provide the Golden_IR_file. Walter said that if the model maker want to provide a working Getwave function for the TD flow - they can easily provide it in their models. Ambrish said that if the model only has an Init function, the model maker will only provide the Type Statistical Test Data - but the user may still want to use the model in the TD flow. Arpad wondered if we are trying to test the model or the tool. If Golden_waveform_file is also provided for Init only models, this will test the tool's capability to generate the correct filter extraction capability. The group went back and forth on why the Golden_IR_file should be optional or required - but agreed that it should be optional for type TD. Michael also mentioned that he would like to have some strong language in the BIRD to let the user know whether the model is supposed to be used in the Getwave(TD) flow or the Init(Statistical) flow. Michael also asked if there should be any mention of which OS/Architecture/Precision/Compiler the Test Data was created as otherwise the tool or the user have no information on how the data was created. Walter said that the Golden Waveform should be within some safe tolerance limit no matter where or how the Golden data was created. Arpad wondered whether we should provide some tolerance number in the BIRD/Spec. Ambrish said that tolerance should be left up to the user. Michael said that there should be some method to find which OS/Architecture/Precision/Compiler was used to generate the data and he will try to incorporate that in the next version of the draft. - Michael: Motion to adjourn. - Ambrish: Second. - Arpad: Thank you all for joining. New ARs: Michael: Send out draft14 of the AMI Test Data proposal including more feedback from the ATM meeting. ------------- Next meeting: 31 October 2023 12:00pm PT -------------