IBIS Macromodel Task Group Meeting date: 17 March 2015 Members (asterisk for those attending): Altera: David Banas ANSYS: * Dan Dvorscak * Curtis Clark Avago (LSI) Xingdong Dai Cadence Design Systems: * Ambrish Varma Brad Brim Kumar Keshavan Ken Willis eASIC * David Banas Ericsson: Anders Ekholm IBM * Steve Parker Intel: * Michael Mirmak Keysight Technologies: Fangyi Rao * Radek Biernacki Nicholas Tzou Maxim Integrated Products: Hassan Rafat Mentor Graphics: * John Angulo * Arpad Muranyi Micron Technology: * Randy Wolff Justin Butterfield QLogic Corp. James Zhou Andy Joy eASIC Marc Kowalski SiSoft: * Walter Katz Todd Westerhoff * Mike LaBonte Synopsys Rita Horner Teraspeed Consulting Group: Scott McMorrow Teraspeed Labs: Bob Ross (Note: Agilent has changed to Keysight) The meeting was led by Arpad Muranyi. ------------------------------------------------------------------------ Opens: -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Michael M send list email about directionality issues. - Done. - Michael M send AMI Direction presentation to Mike L for posting - Posted now. - Walter and Randy produce C_comp BIRD. - Randy: Some progress, but we need to resolve issues about parameters. - Arpad to review IBIS specification for min max issues. - In progress. ------------- New Discussion: AMI Directionality: - Michael M showed IBIS-AMI and Direction Decisions - Michael M: There has been list discussion since this was sent yesterday. - slide 4: - Michael M: There are two options for I/O AMI models: - A: Unified or separate DLLs. - B: Separate DLLs only. - Ambrish: It is for the DLL to decide. - Radek: "A" is already permitted. - Michael M: They are in the wild now? - Walter: Michael M wants separate AMI files for the TX and RX. - We seem to want to have separate AMI files. - How would the tool know which parameters are for which if they are merged? - Arpad: "A" gives more flexibility. - slide 5: - Michael M: Two options for AMI files for I/O: - A: Unified or separate AMI files. - B: Separate AMI files only. - Walter: It's OK to have RX parameters in a TX AMI file. - Ambrish: We gain nothing by mandating a unified AMI file for an I/O. - Radek: Agree, I'm inclined toward "B". - Arpad: Does "B" imply only separate DLLs too? - Radek: A single value parameter would pass to the DLL but have no user options. - Walter: I said one DLL with 2 AMI files requires 2 Executable lines. - For an I/O the tool would have to figure out which Executable is TX and RX. - Michael M: We need clarity on this before we work on syntax. - John: The user could choose which Executable line to use. - Ambrish: They point to the same DLL. - John: A parameter is needed to say if an I/O is TX or RX. - Michael M: We have that in the proposal. - Radek: There could be separate [Algorithmic Model] sections for TX and RX. - Michael M: We need a foolproof association that is not ambiguous. - Ambrish: There is no way to know which DLL goes with which AMI. - Radek: The AMI can be checked for Reserved_Parameter keywords. - Michael M: This would be a check for conflicting parameter types? - Radek: Yes. - Michael M: I am proposing unambiguous parameters for this. - John is proposing an AMI file that uses only parameters not specific to TX or RX. - John: For an I/O a direction parameter is necessary. - Otherwise we have only a cross check. - Walter: We can't check the AMI against the DLL. - We can check the AMI against the IBIS that calls it. - John: We can't stop people from having any Model_Specific parameters they want. - Michael M: The parameter Walter asks for has been in the BIRD all along. AR: Michael M update AMI Directionality BIRD C_Comp enhancements: - Randy showed a C_Comp Model Enhancements BIRD draft. - Randy: This has [C_comp Model] and [End C_comp Model]. - It gives subckt info including the number of terminals. - It overrides any other C_comp keywords. - The "Parameter' sub-parameter should be consistent with the interconnect BIRD. - It just passes Corner values, no other formats. - Walter: We need just Corner values. - Typ-only should be allowed. - Randy: A Touchstone or ISS file can be used. - This tries to discourage using node 0. - It can be a parallel or series model. - Arpad: Can this be combined with a series C_comp? - In the series model the input could connect to "A_signal". - Are only those 4 reference voltages allowed? How about GND? - Randy: It uses the same connections a model has available. - Arpad: It would be best to limit to those four. - Randy: It can be used as a differential model. - There are rules about terminal counts. - It might be used for differential pins. - Mike L: There should be an example of that to be sure. - Walter: We could also use two single ended pins differentially. - Mike L: A protocol to say where true and complement are connected is needed. - Walter: That is defined in this BIRD. - Randy: The rules of precedence are explicit. - Walter: That is needed. PAM4: - Walter showed a draft PAM4 BIRD. - Walter: This allows a model to say if it is NRZ or PAM4. - The mapping of symbols to levels can be given. - The center threshold is 0 for NRZ. - The model can return threshold values. - The outer eye offset supports timing offsets. - Someone brought up duobinary. - David: We should not block that. - Walter: This doesn't block it, it just doesn't specify it. ------------- Next meeting: 24 Mar 2015 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives