IBIS Macromodel Task Group Meeting date: 21 February 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 Cisco: Seungyong (Brian) Baek 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 Graphics: John Angulo * Arpad Muranyi Vladimir Dmitriev-Zdorov 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: - Bob Ross to submit BIRD 186.1 to the Open Forum. - Done. - Arpad to email the Open Forum regarding BIRD 186.1. - Done. - Arpad to email the Open Forum regarding BIRD 147.6. - Done. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: - Two sets of minutes needed review, as the February 7th minutes were not prepared and posted in time for the February 14th meeting. - February 7th minutes: - Arpad: Does anyone have any comments or corrections? [none] - Walter: Motion to approve the minutes. - Bob R.: Second. - Arpad: Anyone opposed? [none] - February 14th minutes: - Arpad: Does anyone have any comments or corrections? [none] - Mike L.: Motion to approve the minutes. - Walter: Second. - Arpad: Anyone opposed? [none] ------------- New Discussion: BIRD 187.2: - Discussion: Radek noted that the Open Forum had deferred the vote on BIRD 187.2 after he had pointed out some inconsistencies in a new sentence describing Format Increment: The Increment Format defines, for Usage In and InOut, a range of discrete integer values which can be swept by the EDA tool using a specified value ("delta"), where min <= typ <= max and delta is always positive. Radek noted that the text seemed to mistakenly imply the values defined by the increment were integers. In addition, "swept by the EDA tool" was a new descriptive phrase not used elsewhere and perhaps not desirable. Walter and Arpad asked Radek to take the AR to send an email to Michael Mirmak (and the ATM reflector) to make him aware of the issue, as he had not been on the Open Forum call and likely was not yet aware of the issue. BIRD 158.3: - Walter: [sharing a draft BIRD 158.4 from a private email with Radek] - Radek: I changed some information on packaging and requirements for the BIRD. - We should discuss it. - Walter: Some graphics have changed. - We've replaced the earth ground symbol with "reference." - I recommend we change the symbol to the triangle local reference symbol. - Ambrish: I've been waiting for the updated version to start my review. - Radek: There are some minor differences between this and 158.3. - One issue is a bit more important. It involves the question of whether the 4-port data file includes the package or not. I think users and model makers might want the flexibility of not including the package in the 4-port data file and providing the package model in an additional 4-port provided with the channel. - Discussion: Walter noted that intent had never been to include the package model in the 4-port data file. It has always been referred to as the on-die interconnect, and it was an alternate method for what could be in [External Model], which was also on-die. Walter then observed that a sentence in the introductory text that said "including the Tx package" was a mistake and was meant to say "not including". This might have caused the confusion. Radek noted that if we decide to pursue this BIRD we might want to elaborate on each possibility (including the package or not including the package). Walter said he felt that it would complicate the BIRD if it went beyond on-die interconnect. He noted that if a model maker wanted to include the package in the 4-port data file then it could be done, and any legacy location for the package model should contain a null package in that case. Arpad asked how the package model would be provided, if the approach of this BIRD is to bypass everything in the IBIS file. Walter clarified that the BIRD intends to bypass everything in the [Model], not entire IBIS file. So, package models could still be provided by the IBIS file. Arpad noted that if the BIRD makes it clear that it only replaces the [Model], then it shouldn't be hard to make this BIRD compatible with the other approaches to providing the on-die interconnect. - Ambrish: If you read in an IBIS model, and you could potentially get the same information about the analog model for the on-die interconnect from two places ([External Model] or the 4-port from this BIRD), does this BIRD make it clear how that's handled? - Discussion: Radek noted that this BIRD says the information from the IBIS [Model] is bypassed. So, both the traditional [External Model] approach and this BIRD's 4-port could be present at the same time. If a tool doesn't support this BIRD then it could use the [External Model] approach. Walter said that, as currently written, the BIRD would not obligate the model maker to provide the [External Model] equivalent. He said the BIRD was written for model makers who did not want to deal with [External Model]. Radek said this BIRD could be considered AMI only in that sense. Walter suggested that a tool could extract the information from the .ami file even for a non-AMI simulation if it wanted to. Walter and Radek noted the "in lieu of the analog model" statement as an answer to Ambrish's question. Ambrish noted that this statement was in the supporting documentation for the BIRD, but not in the text that would actually be part of the spec. Ambrish said that text in the spec should address any issues regarding multiple ways of specifying the same data and how the data should be used. - Editorial Discussion: Arpad asked that the terminals 2 and 4, labelled as "to channel", be changed to "Buffer terminal". Radek/Walter agreed. Bob R. said the parameter name "Tstonefile" should be changed to something else, such as "Ts4", since this BIRD is restricted to a four port model. Bob also noted that Touchstone format supports Y and Z data as well, so the BIRD should not say S-parameter data. Radek suggested this could be changed to Touchstone data. Bob noted that he felt this BIRD was restricted to the specific role in AMI simulations, since the Touchstone model in this BIRD had additional information folded in with the passive interconnect model. He noted that Touchstone files provided by this BIRD were restricted to one direction of flow. - Walter: I'll send the current draft to Radek, and he can incorporate the changes we've discussed. - Radek: Okay. - Arpad: Let's get this posted with today's changes, so we can all review it. Single Ended Applications of AMI: - Walter: I think we need Fangyi and Michael M. here for this topic. - Radek: Motion to adjourn. - Walter: Second. - Arpad: Thank you all for joining. AR: Radek to email Michael Mirmak and ATM regarding BIRD 187.2. AR: Radek to send out the modified draft of BIRD 158.4. ------------- Next meeting: 28 February 2017 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives