IBIS 9/15/95 MINUTES

From: Bob Ross <bob@icx.com>
Date: Tue Sep 19 1995 - 16:39:00 PDT

 DATE: September 19, 1995
 
 SUBJECT: 9/15/95 EIA IBIS Open Forum Meeting Minutes
 
 VOTING MEMBERS:
 AT&T Global Info Solutions Dave Moxley*
 Cadence Design Sandeep Khanna, C. Kumar*
 Contec CAE, Ltd. Dileep Divekar*
 HyperLynx Kellee Crisafulli
 IBM Jay Diepenbrock
 INCASES Werner Rissiek, Olaf Rethmeier
 Intel Corporation Stephen Peters*, Will Hobbs*, Arpad Muranyi,
                                Derrick Duehren
 Interconnectix, Inc. Bob Ross*
 Meta-Software Les Spruiell, Mei Wong, You-Pang Wei,
                                John Sliney
 Motorola Ron Werner
 National Semiconductor Syed Huq*, Atul Agarwal, Cheng-Yang Kao
 NEC Hiroshi Matsumoto, Eldar Yazbashevz*
 Quad Design Jon Powell*
 Quantic Labs Mike Ventham*
 Tanner Research, Inc. Scott Wedge, Ed Miller, Peter Parrish
 Texas Instruments Roger Cline, Ben Andresen*
 Thomson-CSF/SCTF Jean LeBrun
 UniCAD Canada Ltd. Stephen Lum
 VLSI Technology Dick Ulmer*, Sung Oh*
 Zuken-Redac John Berrie
 
 OTHER PARTICIPANTS:
 ARPA Randy Harr
 Anacad Steffen Rochel
 Ansoft Henri Maramis
 Atmel Corporation Dan Terry
 Cadlab Ralf Bruning
 CFI Ron Christopher*
 Digital Equipment Corp. Barry Katz
 EIA Patti Rusher*
 High Design Technology Michael Smith, Dr. Ing. Cosso
 Hewlett Packard Tom Langdorf, Karl Kachigan, Henry Wu
 Integrated Silicon Systems Eric Bracken
 Intergraph Ian Dodd, David Wiens, Walter Katz
 IntuSoft Charles Hymowitz
 LSI Logic Corp. Satish Pratadneni
 Mentor Graphics Ravender Goyal, Greg Doyle
 Micron Technology Brian Johnson
 MicroSim Arthur Wong
 North Carolina State U. Steve Lipa, Michael Steer
 OptEM Engineering, Inc. Benny Leveille, Ken Ehn
 Pacific Numerix Paul K. U. Wang
 Symmetry Martin Walker
 Synopsys, Logic Modeling G. Bill Lattin
 Univ. of Illinois, Urbana Raj Mittra
 Zeelan Technology George Opsahl, Hiro Moriyasu
 (Independent) Bob Ward

 In the list above, attendees at the meeting are indicated by *.
 
 Upcoming Meetings: The bridge numbers for future IBIS teleconferences are as
 follows:
      Date Bridge Number Reservation # Passcode
      10/6/95 (303) 633-2197 1-308636 No pass code

 All meetings are 8:00 AM to 9:55 AM Pacific Time. We try to have agendas out
 7 days before each open forum and meeting minutes out within 7 days after.
 When you call into the meeting, ask for the IBIS Open Forum hosted by Will
 Hobbs and give the reservation number and passcode.
 
 NOTE: "AR" = Action Required.
 
 -------------------------------- MINUTES -------------------------------------
 
 INTRODUCTIONS
 Ron Christopher of CFI and working as an independent consultant on a ten-
 year roadmap for EIA standards introduced his work. He is working with
 CFI, EDAC and Semitech. His long range concern is EDA standardization
 of delay/verification and compatibility of related models including
 IBIS. He is also concerned about the acceptance of standards by major
 vendors. His documents and preliminary reports are on the Web at
 http://www.cfi.org/roadmap/roadMap-eii-report.html.
 

 EIA MEMBERSHIP AND TREASURER'S REPORT
 Patti Rusher reported the treasury is $8725.54. IBM and VLSI are members,
 as previously reported, and Tanner Research has also become a voting
 member.

 MINUTES REPORT, MISC.
 Will submitted some spelling corrections to the 8/25/95 minutes which have
 been made on the posted copy on vhdl.org. Also, Raj Mittra was incorrectly
 reported as present at the last meeting.

 MISCELLANY/ANNOUNCEMENTS
 C. Kumar reported that the 4th Topical Meeting on Electronic Performance
 of Electronic Packaging (EPEP'95) is being held in Portland, Oregon on
 October 2-4, 1995. More information can be obtained on a web site:
 http://www.emclab.umr.edu/news9.html.

 AR - Kumar post this on the reflector [Done].

 Jon Powell still needs logos from participants. Only a few have been
 received. Because of some vhdl.org computer difficulties, the message
 he sent may not have gone out.

 AR - Jon Powell send again the logo message [Done].

 PRESS UPDATES
 None.

 NEW MODELS
 None.

 OPENS FOR NEW ISSUES
 VI table and slow input discussion based on reflector questions.
 Ground bounce discussion based on reflector questions.
 Roadmap questions Ron Christopher is raising related to delay.

 EIA IBIS RATIFICATION (VOTES)
 Patti Rusher reported that she has not received the documents from Will.
 
 AR - Will Hobbs/Stephen Peters FAX all documents to Patti.

 Patti requires the rolled up standard, copies of the official responses
 and changes, and a cover letter. The package will be put together by
 EIA Engineering to be presented at the next EIA Board Meeting (EDAC) for
 review and ratification. This will then become EIA-656 and will also be
 forwarded for ANSI standardization.

 Will Hobbs will work on a generic press release timed at the official
 transition. The press release will list the official EIA members.

 Bob Ross will work with EIA (Cecilia Fleming) regarding ensuring all of
 the preface changes are included (that were not part of IBIS Version 2.1).

 FACE TO FACE MEETING
 Will Hobbs indicated that we typically meet sometime around December
 plus or minus a month. January is proposed as a meeting time. Intel is
 willing to host the meeting unless another company wishes to volunteer.
 A number of technical items will be discussed including monotonicity,
 package models, clamping models, and other items. In the past, these
 meetings have been very informative and have resulted in extensions to
 IBIS.

 WEB UPDATE
 Syed Huq reported that the EIA home page is up in preliminary form with
 a new search utility. IBIS can be accessed by searching on standards
 proposal 3527 or on (case insensitive) ibis. This currently points to
 the IBIS Home page area. Syed is going to put in the approved contents.
 When IBIS becomes EIA-656, then the search will can be done by this
 number. The web site is http://www.eia.org/eng.

 The EDN article appears in Word For Windows format on vhdl.org under
 /pub/ibis/documents. Syed will convert this and have it available on
 the Web site.

 Syed will post the FAQ's and will also send a copy to Bob Ross so they
 can be posted on vhdl.org.

 FAQ VOTE
 Q6: Can IBIS model SSO (Simultaneous Switching Output) ?
 This was posted on the reflector. The issue is that it is dangerous to
 imply that SSO is completely covered because there can be complicated
 interactions with external and internal chip and MCM parameters. However,
 IBIS does have a number of parameters that are useful for SSO analysis.

 To indicate that SSO can be handled by IBIS models to a limited extent,
 the original answer

 "IBIS as a model, has all the important parameters required to model an
 SSO event. ...."

 is changed to
 "IBIS as a model, has the key parameters required to model an SSO event. ..."

 This change was approved by committee vote, so FAQ Q6 can be posted.

 The issue of ground bounce generated some positioning discussion related
 to whether the primary mechanisms are external or internal to the chip.
 Usually the internal details are not known, nor is there any real desire
 to make them available. So the primary focus of ground bounce is at the
 board level. The SSO model being regarded as a practical approximation,
 but not an exact source of all information.
 

 MODEL USAGE TRACKING
 Syed Huq has been in contact with Michael Steer regarding setting up a model
 usage tracking system for vhdl.org. The information would be provided only
 to the providers of models for their models only. Michael Steer needs to
 respond on the problems and progress of this activity.

 AR - Michael Steer work with Syed and give status of model usage tracking
 project.

 GOLDEN PARSER UPDATE
 Version 9 has been distributed. A Test Matrix was distributed showing all
 know problems have been fixed. IBISCHK2 is ready for official (non-Beta)
 distribution.

 AR - Will Hobbs work with Paul Munsey to remove the Beta designation.
 The official Parser will be distributed to those who have purchased it.

 AR - Bob Ross provide Jon Powell with several IBIS test files to validate
 functionality of the Parser. The primary reason is to have test cases
 against which ibishck2 executables for different machines can be tested.

 AR - Jon Powell make executables for a number of machines and test them.

 AR - Jon Powell send-to/work-with Bob Ross to post these executables on
 vhdl.org. This should occur around the October 2 time frame based on
 peoples schedules.

 SPICE TO IBIS VERSION 2.1
 Michael Steer and North Carolina State University team have posted a Beta
 version on vhdl.org under /pub/ibis/s2ibis/s2ibis_v2.09. Bob Ross has
 done a prelimary review, and it functions well. There are a few typos
 in the documentation and the functionality still needs to be reviewed.
 A major difference is that the "*" and "**" control syntax is replaced
 by IBIS keywords and IBIS-like syntax.

 Syed Huq indicated that National has modified the original s2ibis code to
 be compatible with the internal Spice features.

 COOKBOOK
 Bob Ross reported no progress.

 NON-MONOTONOCITY
 Ron Christopher lead the discussion by indicating for a frame of reference
 that handling non-monotonic drivers seems to be a barrier to a major
 vendor (IBM) for adopting IBIS. The goal of IBIS was for a transfer of
 useful information while preserving proprietary process information. IBIS
 does allow the full characterization of non-monotonic VI data (with warning
 messages), but the real problem is that the actual internal mechanisms are
 guarded. Furthermore, performance information is not available. Both
 are needed for simulator vendors to actually process the data properly.

 Ron suggested that some independent vendor such as Zeelan could provide
 measurement based data to give characterization information. That would
 be a start, but not necessarily the total picture. Also, there may
 be problems of vendors suppling such sensitive chips for measurement
 for public distribution and the willingness of whoever does the measurements
 to distribute the information. Ron will pursue this path.

 The technical discussions regarding non-monotonic impact related to possible
 delay artifacts. Jon Powell indicated that attempts to reverse engineer
 the non-monotonic effects to provide a simulation model were to date
 unsuccessful. The simulation data did not match the actual data. So the
 internal mechanisms are probably needed.

 Other companies may have similar technology. However, TI's reference to
 GTL structures are probably not similar.

 Most likely the non-monotonic architecture will be more fully understood
 and could be implemented in IBIS. Sources of information may emerge
 through patents and pending patents, papers and other vendors. Vendors
 interested in using IBIS for these effects need to participate in the IBIS
 activities to supply the essential information so that EDA vendors can
 correctly architect the simulators. Ron, through his contacts could urge
 vendors to become active in IBIS.

 BIRD30 - PIN PROGRAMMABLE BUFFER STRENGTHS
 Arpad Muranyi had submitted BIRD30 in August, and it has already generated
 reflector discussion. Since Arpad was away on a prestigous internal technical
 conference, Bob Ross summarized BIRD30 and the reflector discussion. BIRD30
 introduces the new keyword [Model Selector] keyword. This allows a variety
 of buffers (different voltages, different strengths, different slew rates,
 different Model_types, even different technologies) to be assigned to a
 pin or group of pins. One direction on the reflector was to extend this
 capability to standardize "Classes" of capability. Then the selection method
 could be based on function. The consensus was that this complication was not
 warrented, would be difficult to standardize, and would still be subject to
 individual simulator characteristics.

 An opposite direction was to eliminate the "default" parameter and use
 the first entered model as the default. This would be simplier to implement.

 An intermediate position was to add a "discription" column associated with
 each model name. This discription column would guarentee a text entry
 (as opposed to relying on comment information) which could be viewed
 by simulators. While not standardized, the information would be reasonably
 descriptive. It could be used as the basis for interactive user selection
 of models for a group of pins. The consensus was to go with this approach.

 AR - Arpad Muranyi submit BIRD30.1 with the "default" parameter removed,
 with selection of the first model as the default, and with the a description
 column extension. This will be the basis for further discussion.

 EGG6 - CMOS/TTL DIFFERENCES
 Jon Powell introduced EGG6 discussion, also delayed from August. Basically
 there are differences in the way TTL and CMOS transitions occur. Several
 proposals have been sent regarding ways to detect TTL and CMOS type data.
 However, there are probably counter examples where any such detection
 scheme will fail. An alternative is to introduce into the [Model] keyword
 another sub-parameter designating TTL, CMOS, etc for technology
 differentiation. An automatic detections scheme may still be need if such
 a sub-parameter is not required. This egg is subject to more discussion
 related to how the transition methods of TTL and CMOS differ, possible
 sub-parameter markers, and possible automatic detection methods.
 

 BIRD27.1 - DEFAULT DIFFERENTIAL THRESHOLD
 Still Deferred, but will be submitted by Bob Ross in the future.

 BIRD28.2 PACKAGE MODEL ENHANCEMENT
 Agreement that the "Maxwell" capacitance should be the reference for the
 coupled matrix. Stephen Peters should submit BIRD28.3 with this change.
 This may generate more discussion and or lead to a possible vote.

 ROADMAP QUESTIONS
 Ron Christopher has some technical issues with IBIS which he has in a
 frame document on cfi/incoming/timezer. Since various participants
 do not have ftp or framework capabilitity, it is better to send this
 document in Postscript form on the reflector. Then the questions and
 concerns can be considered.

 AR - Will Hobbs get timezer, transform it, and send to the reflector as
 a Postscript document.

 NEXT MEETING:
 It is set on Friday, October 6, 1995.

 ==============================================================================
                                       NOTES
 
 IBIS CHAIR: Will Hobbs (503) 264-4369, Fax (503) 264-4210
             will_hobbs@ccm.jf.intel.com
             Modeling Manager, Intel Corp.
             2111 NE 28th M/S JF1-57, Hillsboro, OR 97124 USA
 
 VICE CHAIR: Jon Powell (805) 988-8250, Fax: (805) 988-8259
             jonp@qdt.com
             1385 Del Norte Rd., Camarillo, CA 93010
 
 SECRETARY: Bob Ross (503) 603-2523, fax (503) 639-3469
             bob@icx.com
             10220 SW Nimbus Ave, K4, Portland, OR 97223
 
 To become a voting member of the EIA IBIS Open Forum, send email to
 ibis-info@vhdl.org for instructions.
 
 If you want to join the e-mail reflector (ibis@vhdl.org), send e-mail to the
 IBIS secretary at ibis-request@vhdl.org.
 
 Check the pub/ibis directory on vhdl.org for more information on previous
 discussions and results. You can get on via ftp anonymous, "guest" login from
 telnet or dial-in (415-335-0110), or send an email request to the automatic
 archive server, archive@vhdl.org.
 
 "IBIS Spoken Here" placards are available from Jon Powell (jonp@qdt.com) for
 use at trade shows.
 ==============================================================================
 
Received on Tue Sep 19 16:44:07 1995

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:52:29 PDT