[ibis-macro] Minutes from recent ibis-atm meetings
- From: Mike LaBonte <mike@xxxxxxxxxxx>
- To: IBIS-ATM <ibis-macro@xxxxxxxxxxxxx>
- Date: Thu, 8 Nov 2012 16:39:48 -0500
Minutes from recent ibis-atm meetings are attached.
Mike
IBIS Macromodel Task Group
Meeting date: 06 Nov 2012
Members (asterisk for those attending):
Agilent: * Fangyi Rao
* Radek Biernacki
Altera: David Banas
Julia Liu
Hazlina Ramly
Andrew Joy Consulting: Andy Joy
ANSYS: Samuel Mertens
* Dan Dvorscak
* Curtis Clark
Steve Pytel
* Luis Armenta
Arrow Electronics: Ian Dodd
Cadence Design Systems: Terry Jernberg
* Ambrish Varma
Feras Al-Hawari
Brad Brim
* Kumar Keshavan
Ken Willis
Cavium Networks: Johann Nittmann
Celsionix: Kellee Crisafulli
Cisco Systems: Ashwin Vasudevan
Syed Huq
Ericsson: Anders Ekholm
IBM: Greg Edlund
Intel: Michael Mirmak
Maxim Integrated Products: * Mahbubul Bari
* Hassan Raghat
* Ron Olisar
Mentor Graphics: * John Angulo
Zhen Mu
* Arpad Muranyi
Vladimir Dmitriev-Zdorov
Micron Technology: Randy Wolff
Justin Butterfield
NetLogic Microsystems: Ryan Couts
Nokia-Siemens Networks: Eckhard Lenski
QLogic Corp. * James Zhou
SiSoft: * Walter Katz
Todd Westerhoff
Doug Burns
Mike LaBonte
Snowbush IP: Marcus Van Ierssel
ST Micro: Syed Sadeghi
Teraspeed Consulting Group: Scott McMorrow
* Bob Ross
TI: Casey Morrison
Alfred Chong
Vitesse Semiconductor: Eric Sweetman
Xilinx: Mustansir Fanaswalla
Ray Anderson
The meeting was led by Arpad Muranyi
------------------------------------------------------------------------
Opens:
- Arpad unavailable for the 11/13 meeting. Since that also coincides with
the Asian Summit, the group agreed to cancel the meeting on 11/13.
--------------------------
Call for patent disclosure:
- None
-------------
Review of ARs:
- Arpad to send BIRD 121.2 draft 10 to Mike L. to post on the ATM website:
- Forgot after last meeting, was not done until this morning.
- Mike L. not available today so it was not posted.
- emailed to the group.
- done
- Fangyi to send "Redriver" BIRD draft and PowerPoint to Mike L. to post
on the ATM website:
- done
- Walter and Ambrish to clear up the language of BIRD 150 offline.
- done
-------------
New Discussion:
BIRD 121.2 draft 10 (Parameters for Data Management):
- Arpad: Are people ready to comment, or should we postpone the discussion?
- Bob: I have some minor points.
- Usage Rules for Supporting_Files:
- States, "... a file definition is redundant ...", but it is legal.
- Arpad: (changing document)
- new text, "...is legal, but redundant..."
- Bob: Definition for Supporting_Files:
- Remove "According to the IBIS specification".
- Arpad: done.
- Bob: Same sentence:
- Make "executable" plural, and change "its" to "their".
- Radek: actually, change "their" to "the".
- Bob: agreed.
- Arpad: done.
- John: Same sentence:
- Change "defined" to "declared".
- Arpad: done.
- Bob: Usage Rules for Supporting_Files:
- Remove "for some reason".
- Arpad: done.
- Bob: Example for DLL_Path:
- The (Value "NA") concerns me because NA has meaning elsewhere.
- Radek: Replace it with (Value "placeholder").
- Ambrish and Bob: agreed.
- Arpad: done.
- Arpad: I will save this as draft 11 and send it to Mike L. to post to the
ATM website.
Presentation on "Redriver" BIRD draft. (posted to ATM site, see ARs above)
- Mahbubul: Introduction to PowerPoint presentation
- Mahbubul, Ron, Hassan, Fangyi are collaborators.
- Fangyi: presenting.
- Redriver background slides
- Equalizer, no CDR.
- Nonlinear and Noisy, would break LTI assumption for end-to-end channel.
- Need to extend existing AMI capabilities.
- Bob: Are Tx and Rx blocks contained in separate IBIS models.
- Fangyi: yes.
- Fangyi: next page (page 3) of presentation
- Important point is that Tx Model is an analog output circuit, not a
typical D to A scheme.
- Arpad: If the Rx model generates a clock ticks array it is to be ignored?
- Fangyi: yes.
- Arpad: Repeaters with a CDR exist in practice, how are they handled?
- Fangyi: We define that as a "Retimer" not a "Redriver".
- Ambrish: The example shows a "Terminator" model for the Rx? Illegal?
- Fangyi: Yes, that should be "Input".
- We have a BIRD draft to make "Terminator" legal.
- Dan and Bob: Are these flagged as "Redrivers" in the .ibs file?
- Fangyi: No.
- Fangyi: next page - Simulation Flow Diagram.
- Ambrish: Are there any differences between these and a regular AMI file.
- Fangyi: Only difference is that Tx output analog model is not a D to A.
- Fangyi: The EDA tool provides a method for importing a "Redriver".
- The user imports the [Model]s into the "Redriver" device.
- Kumar: The repeater pin premise (BIRD 131) is more general.
- Ambrish: The user figuring out what a "Redriver" model is seems redundant.
- Walter: We have "Redriver" and BIRD 131's "Repeater".
- What's the difference between these proposals?
- Fangyi: The simulation flow is more detailed.
- Walter: Agreed, Agilent's flow is more specific.
- Walter: Does this generalize to multiple repeater packages?
- Ambrish: Agree with Walter.
- Modify the simulation flow of BIRD 131 based on this presentation.
- Walter: This "Redriver" is constrained to analog only Tx.
- There are repeater devices out there that use digital clocks.
- The same model maker is the doing Tx and Rx sides.
- Why not generalize this so model maker can do either type (of Tx).
- Fangyi: The signal that drives the Tx is different in a "Retimer".
- Walter: Redriver has a true analog waveform.
- If Rx has a clock, the only difference is the waveform going to Tx
goes from + or - 0.5V.
- Redriver - Rx returns analog.
- Retimer - Rx returns "digital." Jitter parameters would apply.
- Fangyi: With Retimer you need not use the Rx output waveform at all.
- Kumar: Repeater pin is more general.
- It just says left side, right side.
- Walter: I think we have something of a consensus on a good combination:
- Simulation flow of this proposal should be combined with BIRD 131.
- Ambrish: It's not necessary for that much detail in the flows.
- Walter: disagree
- Useful to show output of Rx GetWave is input to Tx GetWave.
- Walter: I think the flows should use the terminology from IBIS 5.1.
- See page 162 in IBIS 5.1.
- Fangyi: Which terminology is not consistent with 5.1?
- Walter: (asks Fangyi to share the BIRD itself)
- hAC1, for example, that terminology did not make it into 5.1.
- I think the wording is fine, but we need to be consistent.
- I like this terminology, but we should be consistent.
- Arpad: He's distinguishing between two outputs with hAC1 and hAC2.
- Isn't the hAC already consistent with 5.1?
- Walter: I don't think it made it into 5.1.
- I think we agree that we need consistent terminology.
- Fangyi: Agreed.
- Arpad: Can we make it an AR to combine BIRD 131 and this proposal?
- Ambrish: Please include me in the discussions.
- Fangyi: Maxim and Agilent will discuss this first.
- They will then contact Walter and Ambrish.
- Arpad: Okay then, we're about out of time. Any last comments?
- Arpad: Next week's meeting is canceled.
New ARs:
AR: Arpad to send the new draft 11 of BIRD 121.2 to Mike L. to post to the ATM
website.
AR: Fangyi, Mahbubul, et. al., to discuss merging their "Redriver" proposal
with BIRD 131.
AR: Fangyi, Mahbubul, et. al., contact Walter and Ambrish for further BIRD 131
discussions.
-------------
Next meeting: 20 Nov 2012 12:00pm PT
Next agenda:
1) Task list item discussions
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
IBIS Macromodel Task Group
Meeting date: 16 Oct 2012
Members (asterisk for those attending):
Agilent: * Fangyi Rao
* Radek Biernacki
Altera: * David Banas
Julia Liu
Hazlina Ramly
Andrew Joy Consulting: Andy Joy
ANSYS: Samuel Mertens
* Dan Dvorscak
* Curtis Clark
Steve Pytel
Luis Armenta
Arrow Electronics: Ian Dodd
Cadence Design Systems: Terry Jernberg
* Ambrish Varma
Feras Al-Hawari
Cavium Networks: Johann Nittmann
Celsionix: Kellee Crisafulli
Cisco Systems: Ashwin Vasudevan
Syed Huq
Ericsson: Anders Ekholm
IBM: * Greg Edlund
Intel: Michael Mirmak
Maxim Integrated Products: Mahbubul Bari
Mentor Graphics: * John Angulo
Zhen Mu
* Arpad Muranyi
Vladimir Dmitriev-Zdorov
Micron Technology: Randy Wolff
Justin Butterfield
NetLogic Microsystems: Ryan Couts
Nokia-Siemens Networks: Eckhard Lenski
QLogic Corp. * James Zhou
Sigrity: Brad Brim
Kumar Keshavan
Ken Willis
SiSoft: * Walter Katz
Todd Westerhoff
Doug Burns
* Mike LaBonte
Snowbush IP: Marcus Van Ierssel
ST Micro: Syed Sadeghi
Teraspeed Consulting Group: Scott McMorrow
* Bob Ross
TI: Casey Morrison
Alfred Chong
Vitesse Semiconductor: Eric Sweetman
Xilinx: Mustansir Fanaswalla
Ray Anderson
The meeting was led by Arpad Muranyi
------------------------------------------------------------------------
Opens:
- Fangyi: There is a conference next week
- Should we meet?
- Arpad: I will be there, but we still could meet
- We will meet next week
--------------------------
Call for patent disclosure:
- None
-------------
Review of ARs:
- Walter & Arpad determine latest BIRD 121 draft number
- Done
- Walter & Arpad clear up BIRD 121 draft language
- In progress
- Walter & Ambrish clean up BIRD 150
- In progress
- Walter update BIRD 123
- The latest is posted on the website
- Radek planned to check the clock PDF section
-------------
New Discussion:
Interconnect Task Group report
- Walter: The meeting tomorrow will be interesting
Arpad: Ambrish requested time to discuss BIRD 147
BIRD 121:
- Arpad: The question on whether to have trailing backslash remains
- John: We should always add a trailing slash
- Mike: That would preclude relative paths
- John: A path ending in ./ will work
- Walter: There are 3 options regarding back and forward slashes
- Walter added language regarding use of "." for relative paths
- Mike: DLL_Path is not stored in files, it is dynamically generated
- It should be in the format for the current OS
- Walter: Back slashes are a problem because they can be escaped out
- Mike: Should we rule out UNC paths, which begin with 2 backslashes?
- James: We should have a rule that the path is valid for the platform
- Arpad: Is the added sentence OK?
- John: It should state exactly that it should be '.'
- Bob: Is backslash allowed?
- Arpad: No
- Bob: This has to be compatible with AMI file checking
- Radek: This does not reside in the AMI file
- Fangyi: This parameter is not required
- What do DLLs do if it is absent?
- John: It should be provided if the DLL needs to find supporting files
- Ambrish: There is a Supporting_Files parameter proposed
- Walter: Supporting_Files may not all be needed by the model
- They could be Touchstone files needed for channel simulation
- There could be files for backchannel
- Ambrish: These should not be defined in the AMI file then
- John: This is tangential to whether DLL_Path is required
- Arpad: If the parser can't check DLL_Path, why bother?
- Walter: IBIS somehow has to provide for copying all required supporting files
- Ambrish: An AMI model can be run without package models
- Walter: We should not be deciding when DLL_Path is required
- Mike: If DLLs can't count on DLL_Path being provided they will not use it
- Walter: It should not be required if it is not needed
- Most DLLs may not need it
- Mike: It is not hard to always provide DLL_Path
- Walter: The model will request it by asking for an In parameter
- Ambrish: If the parser can't check it it will be ignored
- Radek: I agree with Walter
- Walter: This will be posted as revision 9
- Mike: Two revisions held by Walter have not been posted
- The last posted was 6 and this will be 7
- Arpad: The last one should be fixed if it say rev 8 inside
AR: Walter send BIRD 121.2 draft 7 to Mike for posting
AR: Mike check BIRD 121.2 draft 6 for correct numbering inside
Walter: I have sent a BIRD for assigning labels to lists
-------------
Next meeting: 23 Oct 2012 12:00pm PT
Next agenda:
1) Task list item discussions
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
IBIS Macromodel Task Group
Meeting date: 23 Oct 2012
Members (asterisk for those attending):
Agilent: Fangyi Rao
* Radek Biernacki
Altera: David Banas
Julia Liu
Hazlina Ramly
Andrew Joy Consulting: Andy Joy
ANSYS: Samuel Mertens
* Dan Dvorscak
* Curtis Clark
Steve Pytel
Luis Armenta
Arrow Electronics: Ian Dodd
Cadence Design Systems: Terry Jernberg
* Ambrish Varma
Feras Al-Hawari
Cavium Networks: Johann Nittmann
Celsionix: Kellee Crisafulli
Cisco Systems: Ashwin Vasudevan
Syed Huq
Ericsson: Anders Ekholm
IBM: Greg Edlund
Intel: Michael Mirmak
Maxim Integrated Products: * Mahbubul Bari
Mentor Graphics: * John Angulo
Zhen Mu
* Arpad Muranyi
Vladimir Dmitriev-Zdorov
Micron Technology: Randy Wolff
Justin Butterfield
NetLogic Microsystems: Ryan Couts
Nokia-Siemens Networks: Eckhard Lenski
QLogic Corp. James Zhou
Sigrity: Brad Brim
Kumar Keshavan
Ken Willis
SiSoft: * Walter Katz
Todd Westerhoff
Doug Burns
* Mike LaBonte
Snowbush IP: Marcus Van Ierssel
ST Micro: Syed Sadeghi
Teraspeed Consulting Group: Scott McMorrow
* Bob Ross
TI: Casey Morrison
* Alfred Chong
Vitesse Semiconductor: Eric Sweetman
Xilinx: Mustansir Fanaswalla
Ray Anderson
The meeting was led by Arpad Muranyi
------------------------------------------------------------------------
Opens:
- None
--------------------------
Call for patent disclosure:
- None
-------------
Review of ARs:
- Walter send BIRD 121.2 draft 7 to Mike for posting
- Done
- Walter & Arpad clear up BIRD 121 draft language
- Done, draft 8
- Mike check BIRD 121.2 draft 6 for correct numbering inside
- Fixed
- Walter & Ambrish clean up BIRD 150
- In progress
- Walter update BIRD 123
- Walter: Radek planned to check the clock PDF section
- Radek: In 5.0 we had the RjDj option, should be included here
- Walter's motion to submit the BIRD was delayed
-------------
New Discussion:
- Interconnect task group report:
- Walter: Last week was mostly a presentation by me
- There has been an email thread about what can be in a package
- We might make part of the model dependent of the die, part on the package
Arpad showed BIRD 121.2 draft 8:
- Arpad: Bob had asked for a change to the description field
- Added clarification "current working directory of DLL process"
- Ambrish: The description needs to be fixed to "where the DLL is located"
- Arpad made the change
- John: We should be clear that it is a process running the DLL
- Bob: Do we have to spell out which Supporting_Files are files and directories?
- Mike: The tool can figure that out easily
- Arpad: We should be clear about forward and backward slashes
- Radek: It does not say if the path
- John: The contents of directories should be copied
- Ambrish: That should be documented
- Bob: We need to say if paths are allowed
- Walter: We would not allow paths to items in subdirectories
- Radek: We should allow that
- Arpad: We should say that a subdirectory item in a copied directory just
copies the directory
AR: Walter add slash and directory clarifications to BIRD 121.2
Walter motioned to submit the BIRD 123 update in current form to Open Forum:
- Arpad: There have been no recent changes
- Radek seconded the motion
- Arpad showed the BIRD
- Ambrish: Is numeric literal correct?
- Bob: Default is optional
- No one objected to the motion
AR: Arpad send BIRD 123 draft 5 to Open Forum
BIRD 150 was tabled
Arpad: Fangyi mentioned repeaters
- We will schedule time for his proposal
Ambrish showed Back-Channel Support BIRD draft 3:
- Ambrish: This is now in the IBIS 5.1 document format
- It proposes the "Bits" type
- It has Training and Backchannel_Protocol parameters in the .ami file
- A number of other parameters are defined for the .bci file
- Bob: The Format of Training_Done is not right
- It should not have a Default
- Description has to be a string
- Mike: Is the .bci file written incrementally?
- Ambrish: It is written once
- Walter: It is read-only, supplied with the AMI file
- Bob: Does the .bci file have a version number?
- Ambrish: It could have [BCI Version]
- Walter described IBIS spec update scenarios
- It probably would be best to have [BCI Version]
- Ambrish: I will add this
- Walter: There are industry standards for backchannel protocol
- We should have real examples to work with
- Ambrish: There may be issues getting this from an IC vendor
- Alfred: I might be able to create an example
- Radek: We should allow newer .bci files with older .ami files
- Walter: The protocol could specify things only the tool needs to know
Arpad: We should finish the two BIRDs in progress then focus on backchannel
-------------
Next meeting: 23 Oct 2012 12:00pm PT
Next agenda:
1) Task list item discussions
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
IBIS Macromodel Task Group
Meeting date: 30 Oct 2012
Members (asterisk for those attending):
Agilent: * Fangyi Rao
* Radek Biernacki
Altera: * David Banas
Julia Liu
Hazlina Ramly
Andrew Joy Consulting: Andy Joy
ANSYS: Samuel Mertens
* Dan Dvorscak
* Curtis Clark
Steve Pytel
Luis Armenta
Arrow Electronics: Ian Dodd
Cadence Design Systems: Terry Jernberg
* Ambrish Varma
Feras Al-Hawari
Cavium Networks: Johann Nittmann
Celsionix: Kellee Crisafulli
Cisco Systems: Ashwin Vasudevan
Syed Huq
Ericsson: Anders Ekholm
IBM: Greg Edlund
Intel: Michael Mirmak
Maxim Integrated Products: * Mahbubul Bari
Mentor Graphics: * John Angulo
Zhen Mu
* Arpad Muranyi
Vladimir Dmitriev-Zdorov
Micron Technology: Randy Wolff
Justin Butterfield
NetLogic Microsystems: Ryan Couts
Nokia-Siemens Networks: Eckhard Lenski
QLogic Corp. * James Zhou
Sigrity: Brad Brim
Kumar Keshavan
Ken Willis
SiSoft: * Walter Katz
Todd Westerhoff
Doug Burns
* Mike LaBonte
Snowbush IP: Marcus Van Ierssel
ST Micro: Syed Sadeghi
Teraspeed Consulting Group: Scott McMorrow
* Bob Ross
TI: Casey Morrison
Alfred Chong
Vitesse Semiconductor: Eric Sweetman
Xilinx: Mustansir Fanaswalla
Ray Anderson
The meeting was led by Arpad Muranyi
------------------------------------------------------------------------
Opens:
Arpad: Should we cancel Nov 13 due to summit travel?
- Bob: No one from this group will be traveling
- We decided not cancel
Arpad: Should we cancel Nov 27 and Dec 3 due to holidays?
- We decided not cancel
--------------------------
Call for patent disclosure:
- None
-------------
Review of ARs:
- Walter add slash and directory clarifications to BIRD 121.2
- Done
- Walter & Ambrish clean up BIRD 150
- In progress
- Walter and Arpad update BIRD 123
-------------
New Discussion:
Interconnect Task Group report:
- Arpad: There was no meeting
Arpad: Should we discuss the new repeater BIRD while Fangyi and Mahbubul
are both here?
- Mahbubul: It would be better to have time to prepare
- An update could be sent ahead of next week
- Ambrish: Is this BIRD different from what we have?
- Fangyi: It is a new BIRD
- Arpad congratulated Fangyi on his EPEP best paper award
AR: Fangyi and Mahbubul send presentation to Mike for posting
BIRD 121.2:
- Arpad showed BIRD 121.2 draft 9:
- Radek: It may be missing part of the end
- Arpad: I will check
- Arpad: Parts of "Supporting_Files" are under Definition and other parts
under Usage Rules.
- John: The directory of the IBIS file is ruled out for Supporting_Files?
- Arpad: That was not intended
- Walter: That sentence was added to prevent adding unknown files
- Arpad: You don't have to use dot, the files can be added by name.
- John: Not sure dot should be prohibited
- Arpad: It could be dot slash
- Ambrish: There could be tons of files there
- Mike: The IBIS and AMI files would have to not be double copied
- Arpad: Should wildcards allowed?
- Mike: That might make it too easy to copy unintended files
- It would be best for the parameter to list all files explicitly
- Bob: Should the parser disallow asterisk?
- Radek: There could be cross OS issues
- Arpad: If a model is for one OS it should be safe to use a wildcard
- Mike: The rules for Perl globbing are platform independent
- Walter: We should keep it simple
- Either say nothing or disallow wildcards
- Arpad: It would be better to disallow
- Arpad added a clause prohibiting wildcards
- Ambrish: The characters are actually allowed in file names
- Mike: We might say "there is no expectation of wildcard expansion"
- A sentence was formulated
- Bob: Is the sentence disallowing "." sufficient?
- John: A directory could be ".mydir"
- Bob: Should we exclude names beginning with '..'?
- Walter: The intent was to support current directory and below
- Ambrish: How does the model maker know what is around the directory?
- John: These files might be somewhere down a directory tree of files
- Ambrish: The AMI file is supposed to be at the top
- Arpad: We could change the language to disallow copying the full
DLL_Path directory
- John: We can make it explicit that files must be same directory or below
- Arpad: Matlab might require files that are up and over
- Walter: Environment variables can be used for that
- We could specify the names of those
- Ambrish: That is OS specific
- Mike: Is it clear that the files have to be copied to the same tree structure?
- Arpad: There is a sentence covering that
- Walter: This is an issue for all of IBIS
- Every IBIS file has an implicit list of supporting files
- It is assumed that they are in the same directory
- Arpad: This file will be posted for review
- Hopefully we can vote on it next week
AR: Arpad send updated BIRD 121.2 to Mike for posting
-------------
Next meeting: 6 Nov 2012 12:00pm PT
Next agenda:
1) Task list item discussions
-------------
IBIS Interconnect SPICE Wish List:
1) Simulator directives
Other related posts:
