Minutes from the last 3 ibis-atm meetings are attached. Mike
IBIS Macromodel Task Group Meeting date: 09 February 2010 Members (asterisk for those attending): Adge Hawes, IBM Ambrish Varma, Cadence Design Systems * Anders Ekholm, Ericsson * Arpad Muranyi, Mentor Graphics Corp. Barry Katz, SiSoft * Bob Ross, Teraspeed Consulting Group Brad Brim, Sigrity Brad Griffin, Cadence Design Systems Chris Herrick, Ansoft Chris McGrath, Synopsys * Danil Kirsanov, Ansoft David Banas, Xilinx Deepak Ramaswany, Ansoft Donald Telian, consultant Doug White, Cisco Systems * Eckhard Lenski, Nokia-Siemens Networks Eckhard Miersch, Sigrity Essaid Bensoudane, ST Microelectronics * Fangyi Rao, Agilent Ganesh Narayanaswamy, ST Micro Gang Kang, Sigrity Hemant Shah, Cadence Design Systems Ian Dodd, consultant Jerry Chuang, Xilinx Joe Abler, IBM John Angulo, Mentor Graphics John Shields, Mentor Graphics * Ken Willis, Cadence Design Systems * Kumar Keshavan, Sigrity Lance Wang, Cadence Design Systems Luis Boluna, Cisco Systems Michael Mirmak, Intel Corp. * Mike LaBonte, Cisco Systems Mike Steinberger, SiSoft Mustansir Fanaswalla, Xilinx Patrick O'Halloran, Tiburon Design Automation Paul Fernando, NCSU Pavani Jella, TI Radek Biernacki, Agilent (EESof) * Randy Wolff, Micron Technology Ray Komow, Cadence Design Systems Richard Mellitz, Intel Richard Ward, Texas Instruments Samuel Mertens, Ansoft Sam Chitwood, Sigrity Sanjeev Gupta, Agilent Shangli Wu, Cadence Design Systems Sid Singh, Extreme Networks Stephen Scearce, Cisco Systems Steve Kaufer, Mentor Graphics Steve Pytel, Ansoft Syed Huq, Cisco Systems Syed Sadeghi, ST Micro Ted Mido, Synopsys Terry Jernberg, Cadence Design Systems * Todd Westerhoff, SiSoft Vladimir Dmitriev-Zdorov, Mentor Graphics Vikas Gupta, Xilinx Vuk Borich, Agilent * Walter Katz, SiSoft * Zhen Mu, Mentor Graphics ------------------------------------------------------------------------ Opens: - Arpad: We can change the order of items on the agenda if necessary -------------------------- Call for patent disclosure: - No one declared a patent. ------------- Review of ARs: - Arpad: Write a clarification BIRD to discuss accuracy issues related to the various AMI clock_tick algorithms in an IBIS-AMI DLL - TBD - Todd: Update the BIRD for IBIS S-parameter box based on feedback from discussion - No update - Arpad: Write parameter passing syntax proposal (BIRD draft) for *-AMS models in IBIS that is consistent with the parameter passing syntax of the AMI models - TBD - TBD: Propose a parameter passing syntax for the SPICE - [External ...] also? - TBD - Arpad: Review the documentation (annotation) in the macro libraries. - Deferred until a demand arises or we have nothing else to do ------------- New Discussion: Walter showed his Feb 7 summary email of AMI update BIRD changes: - Jitter should be handled in another BIRD - A separate BIRD could be written for analog models - Issues with the BIRD should be handled by email - Bob: The IBIS committee should not be burdened with resolving issues - We have to do a better job than the first time - Jitter has been stripped out, but needs to be replaced with something - Arpad: We have to decide where to handle jitter - Ken: It would help to have jitter changes in a separate BIRD - Arpad: It would be best to have jitter changes all in one BIRD - These BIRDs will go into the same IBIS release Walter showed the AMI update BIRD - Bob: There is not much change from the old text to the new - It should go quickly - Walter: Added impulse response definition - Added to tree structure explanation - Includes link to Wikipedia - Text fields unlimited, can contain linefeeds - Kumar: The strings might contain the entire tree of a model - Arpad: This makes newlines different if they are inside vs. outside quotes - Is it OK in descriptions? - Walter: Yes, it's easier than backslash-N in a flat string. - Mike: We might need the ability to escape quotes in strings - Bob: EDA tools are free to ignore linefeeds if they wish - Walter: Vector is changed to array - Arpad: In some places vector makes more sense - Walter: NA as used here may not be compatible with other uses - For example min=NA and max=NA means use typ values - Bob: Increment should not have typ - Walter: The increments are based on it - Randy: The example passed value=55 is not possible with delta=50 - Walter: For Corner, typ is probably the best Default value - Bob: Descriptions should not be too large to fit on a screen - Walter: Fangyi had recommendations for parameter passing - Kumar: This defies C conventions about string allocation and freeing - The AMI model developer is burdened with tracking the memory - Todd: If something is passed by accident the tool should ignore it - Walter: Should we require declaring things like GetWave_Exists as Boolean? - Mike: Should reserved parameters require type definition at all? - Walter: Some parameters can have different types - Bob: It should be optional, not banned - Walter: We should allow exponential notation for Ignore_bits - This can be a large integer - Bob: We never addresses the meaning of integer and float - Arpad: What if someone specifies 10.5 bits? - Bob: This can be a problem when passed to a DLL - Walter: This is covered - Walter changed the spec to allow 1e5 as an Integer, for example - Walter: Clock DCD and data DCD are separate - The existing BIRD covers clock DCD - The jitter BIRD can cover data DCD - Fangyi: It should be in a separate BIRD - Walter: Array is allowed at IBM's request - They produced non-compliant models but we can accommodate - Arpad: Can we create a new syntax that makes them compliant? - Walter: This would do that - There is an example of it in the BIRD - Ken: The value can be a file name, which can have anything - Todd: Then the software has no ability to provide a user interface - Some models do have external data - There does need to be a way to tell them where to find it - Ken: We should already have that now - Walter: Reading a file for every GetWave call is inefficient - Todd: We had Table in the original spec, but found flaws in it - These are conceptually the same thing - Walter: There was no ability to change values in Table - Arpad: So with Array we can set values by index? - Walter: No but each element can have a different type, like Range - This allows a GUI to manipulate it Arpad: Should this be merged with the other document? - Walter: AR: Walter send AMI update BIRD changes to Arpad for integration AR: Mike post updated AMI update BIRD draft Todd: Should we have a meeting next week? - We have winter break - Arpad: There will be no meeting next week Next meeting: 23 Feb 2009 12:00pm PT -------- IBIS Interconnect SPICE Wish List: 1) Simulator directives
IBIS Macromodel Task Group Meeting date: 23 February 2010 Members (asterisk for those attending): Adge Hawes, IBM * Ambrish Varma, Cadence Design Systems * Anders Ekholm, Ericsson * Arpad Muranyi, Mentor Graphics Corp. Barry Katz, SiSoft * Bob Ross, Teraspeed Consulting Group Brad Brim, Sigrity Brad Griffin, Cadence Design Systems Chris Herrick, Ansoft Chris McGrath, Synopsys * Danil Kirsanov, Ansoft David Banas, Xilinx Deepak Ramaswany, Ansoft Donald Telian, consultant Doug White, Cisco Systems * Eckhard Lenski, Nokia-Siemens Networks Eckhard Miersch, Sigrity Essaid Bensoudane, ST Microelectronics * Fangyi Rao, Agilent Ganesh Narayanaswamy, ST Micro Gang Kang, Sigrity Hemant Shah, Cadence Design Systems Ian Dodd, consultant Jerry Chuang, Xilinx Joe Abler, IBM * John Angulo, Mentor Graphics John Shields, Mentor Graphics Ken Willis, Cadence Design Systems Kumar Keshavan, Sigrity Lance Wang, Cadence Design Systems Luis Boluna, Cisco Systems Michael Mirmak, Intel Corp. * Mike LaBonte, Cisco Systems Mike Steinberger, SiSoft * Mustansir Fanaswalla, Xilinx Patrick O'Halloran, Tiburon Design Automation Paul Fernando, NCSU Pavani Jella, TI Radek Biernacki, Agilent (EESof) * Randy Wolff, Micron Technology Ray Komow, Cadence Design Systems Richard Mellitz, Intel Richard Ward, Texas Instruments Samuel Mertens, Ansoft Sam Chitwood, Sigrity Sanjeev Gupta, Agilent Shangli Wu, Cadence Design Systems Sid Singh, Extreme Networks Stephen Scearce, Cisco Systems Steve Kaufer, Mentor Graphics Steve Pytel, Ansoft Syed Huq, Cisco Systems Syed Sadeghi, ST Micro Ted Mido, Synopsys Terry Jernberg, Cadence Design Systems * Todd Westerhoff, SiSoft Vladimir Dmitriev-Zdorov, Mentor Graphics Vikas Gupta, Xilinx Vuk Borich, Agilent * Walter Katz, SiSoft Zhen Mu, Mentor Graphics ------------------------------------------------------------------------ Opens: - None -------------------------- Call for patent disclosure: - No one declared a patent. ------------- Review of ARs: - Arpad: Write a clarification BIRD to discuss accuracy issues related to the various AMI clock_tick algorithms in an IBIS-AMI DLL - TBD - Todd: Update the BIRD for IBIS S-parameter box based on feedback from discussion - No update - Arpad: Write parameter passing syntax proposal (BIRD draft) for *-AMS models in IBIS that is consistent with the parameter passing syntax of the AMI models - TBD - TBD: Propose a parameter passing syntax for the SPICE - [External ...] also? - TBD - Arpad: Review the documentation (annotation) in the macro libraries. - Deferred until a demand arises or we have nothing else to do ------------- New Discussion: Arpad showed the AMI Improvements BIRD: - Arpad: We should accept the changes and make the markup go away for readability. - Bob: We need to review some technical issues - For example, arrays - Why do we need labels? Walter lead a page by page BIRD review, accepting changes on the fly: - A new paragraph describes param passing to DLL: - Each param has only one value when it reaches the DLL - The user can select a value, or it can be automatic - Fangyi: We don't need to have NA values passed to the DLL - Bob: If the flow is Output we could not even declare a value for Model_Specific - Walter: OK - Walter: NA means no constraint - It also means Not Applicable for In and InOut parameters - Walter: Labels were initially used in Table - Now it is used for List - It is extremely useful - Walter added line that number of labels must equal number of list items - Walter: We could make Usage and Type optional for boolean params like GetWave_Exists - John: That's OK - Bob: The syntax checking gets complicated - Todd: Usage and Type boolean should never be used - Boolean should be regarded as the default type - Ambrish: For all parameters? - Todd: For all boolean parameters - Bob: There is mandated non-legacy support - It must be consistent with 5.0 Walter deleted "Float or" from Ignore_Bits, which must be Integer On Page 23: - Walter: Will propose changes to the definition of tree structure - Bob: We should not have a link to Wikipedia Walter removed Arpad's note about simulation times on page 26 - Walter: It can be an amendment to this BIRD Walter: I would like to get this BIRD out soon - Others can propose amendments after reading it - The final BIRD will be integrated into IBIS 5.1 - Jitter is non-trivial and will have to be separate - Arpad: Do we have a list of questions about the BIRD? - Bob: I have questions about requiring Default - Backward compatibility is the question - Walter: Any 5.0 file will be legal under this spec - Bob: An exception is GetWave_Exists - 5.0 does not allow format value - We decided this to get the parser out the door - Walter: Only for Tx_DCD - Ambrish: There is an exception for Tx_DCD - Walter: I will try to write some words to take care of this - Todd: The value can be omitted but the Default is True - Bob: We need to allow Default for legacy support - This can be constrained to 5 parameters Walter explained the new keyword Ignore_Time: - It is like Ignore_Bits but in units of time - Fangyi: This would have to be mutually exclusive - Walter: We could: - Take the longer time value - Prefer Ignore_Time - Make it an error to have both - Ambrish: What is the advantage? - Walter: It just an easier way to express it - Hard to express in bits when you have jitter, for example - Can be in UI - Todd: Time is independent of bit rate - Typically this is for receiver adaptation time, which is in bits - Giving a choice can be easier for the user - We decided not to propose Ignore_Time right now AR: Walter send updated AMI Improvements BIRD to Mike for posting Arpad: Again, do we have a list of technical issues? - We need a more constructive approach than repeated group passes through the BIRD - Mike: It would be best to have someone's punch list of items to discuss - Bob: I will have questions regarding arrays and NA - Arpad: We need to use email more for these discussions Next meeting: 02 Mar 2009 12:00pm PT -------- IBIS Interconnect SPICE Wish List: 1) Simulator directives
IBIS Macromodel Task Group Meeting date: 26 January 2010 Members (asterisk for those attending): Adge Hawes, IBM * Ambrish Varma, Cadence Design Systems * Anders Ekholm, Ericsson * Arpad Muranyi, Mentor Graphics Corp. Barry Katz, SiSoft * Bob Ross, Teraspeed Consulting Group Brad Brim, Sigrity Brad Griffin, Cadence Design Systems Chris Herrick, Ansoft Chris McGrath, Synopsys * Danil Kirsanov, Ansoft David Banas, Xilinx Deepak Ramaswany, Ansoft Donald Telian, consultant Doug White, Cisco Systems * Eckhard Lenski, Nokia-Siemens Networks Eckhard Miersch, Sigrity Essaid Bensoudane, ST Microelectronics Fangyi Rao, Agilent Ganesh Narayanaswamy, ST Micro Gang Kang, Sigrity Hemant Shah, Cadence Design Systems Ian Dodd, consultant Jerry Chuang, Xilinx Joe Abler, IBM * John Angulo, Mentor Graphics John Shields, Mentor Graphics Ken Willis, Cadence Design Systems Kumar Keshavan, Sigrity Lance Wang, Cadence Design Systems Luis Boluna, Cisco Systems Michael Mirmak, Intel Corp. * Mike LaBonte, Cisco Systems Mike Steinberger, SiSoft Mustansir Fanaswalla, Xilinx Patrick O'Halloran, Tiburon Design Automation Paul Fernando, NCSU Pavani Jella, TI Radek Biernacki, Agilent (EESof) Randy Wolff, Micron Technology Ray Komow, Cadence Design Systems Richard Mellitz, Intel Richard Ward, Texas Instruments Samuel Mertens, Ansoft Sam Chitwood, Sigrity Sanjeev Gupta, Agilent Shangli Wu, Cadence Design Systems Sid Singh, Extreme Networks Stephen Scearce, Cisco Systems Steve Kaufer, Mentor Graphics Steve Pytel, Ansoft Syed Huq, Cisco Systems Syed Sadeghi, ST Micro Ted Mido, Synopsys Terry Jernberg, Cadence Design Systems Todd Westerhoff, SiSoft Vladimir Dmitriev-Zdorov, Mentor Graphics Vikas Gupta, Xilinx Vuk Borich, Agilent * Walter Katz, SiSoft Zhen Mu, Mentor Graphics ------------------------------------------------------------------------ Opens: - Mike: Some recent unposted minutes files may be lost - Some files carried over through an operating system change are encrypted - Mike will try to recover the files -------------------------- Call for patent disclosure: - No one declared a patent. ------------- Review of ARs: - Walter to email latest BIRD draft to group - Done - Mike to convert Walter's BIRD to PDF and post to web - Done - Arpad: Write a clarification BIRD to discuss accuracy issues related to the various AMI clock_tick algorithms in an IBIS-AMI DLL - TBD - Todd: Update the BIRD for IBIS S-parameter box based on feedback from discussion - No update - Arpad: Write parameter passing syntax proposal (BIRD draft) for *-AMS models in IBIS that is consistent with the parameter passing syntax of the AMI models - TBD - TBD: Propose a parameter passing syntax for the SPICE - [External ...] also? - TBD - Arpad: Review the documentation (annotation) in the macro libraries. - Deferred until a demand arises or we have nothing else to do ------------- New Discussion: Bob: Advice given as IBIS chair: - Everything has a cost/schedule/performance tradeoff - Cost is restrained - Performance is key - The syntax must be technically robust - The description must be publication quality - We have to make rational technical decisions - But we don't have the needed clarity - We need to decide: - Time constraint - Feature set - Arpad: It's not that we are dissatisfied with Walter's work - Walter: It would be valuable to have a writer with technical competence - We are not quite ready for a parser developer to use it - EDA vendors need to make that judgment - We need to set a date to have it ready by - IC vendors are anxious to move forward on this - I'm preparing jitter and analog amendments - Bob: Disagree with most of those points - Walter: Not saying this document is ready for prime time - Bob: The process is not about EDA vendors - It must be specification quality - Walter: Agree with that - Bob: We should not bind to a schedule - It could be submitted today, but there would be much discussion - Walter: We should all make recommendations on how to go forward - For now we can discuss recent technical emails - Arpad: We probably agree the BIRD is not ready for submission - We can discuss hiring someone for technical writing - Mike: Maybe a developer could be hired to look at it - That was what derailed us last time - Bob: We can discuss this for 1/2 hour in the next open forum - Everyone has an interest in this - Arpad: Agree with Mike's idea - It would help to hear from outsider spec users Bob: I have received private concerns on this BIRD: - This will be a consideration for the editorial committee - Some features may have to be stripped off to make a cut line - Mike: Anonymous concerns might be OK, but not secret ones - Bob: For example, someone may be using the deprecated Table keyword - Walter: We have asked if anyone knew of anyone using Tables - That should be public - Arpad: Why would that have to be private? - Bob: This is a legitimate vendor, that's all I can say - Walter: Are they claiming the models will work in any EDA tool? - Bob: Any vendor can donate a technical writer's time - But the technical committee has ultimate control Walter showed amendment suggestions sent by email: - A new section tries to answer "What is a parameter?" - What know what a parameter tree is - A branch is a parameter if it has one of a given list of leaves - Ambrish: Can branches not be parameters? - Walter: DFE taps would not be parameters passed to the DLL - In the AMI file it would have keywords to make it a parameter - Ambrish: So "Tap" is not passed to the model? - Walter: I have an example below - Bob: It is good to clarify this Walter: Changes were made to Default - This handles the 5.0 spec completely - Bob: Can we have Value and Default together? Walter showed the txtaps example mentioned above: - Ambrish: Every branch is an AMI parameter - Mike: A branch then is a parameter if it has any leaf of a known type - Ambrish: Can framis for example have subparameters Walter: BNF can describe only a context insensitive grammar: - The AMI file is context sensitive - I provided a template AMI file with limited BNF-like syntax - Not sure if description of tap is OK - Bob: Agree BNF is not the way to go - This format should not be in the spec - IBIS is context sensitive - BNF can be helpful - It can't be the formal definition - It should not be in this BIRD - Walter: Not prepared to remove it - Arpad: Concerned that we don't have a good description without the BNF - Not ready to decide on this yet Next meeting: 9 Feb 2009 12:00pm PT -------- IBIS Interconnect SPICE Wish List: 1) Simulator directives