IBIS 11/18/94 Meeting Minutes

Text item: Text_1

Date: Dec. 2, 1994

From: Will Hobbs (503) 696-4369, fax (503) 696-4210
         Will_Hobbs@ccm2.jf.intel.com
         XTG Modeling Manager, Intel Corp., Chairperson, IBIS Open Forum
         Intel Corporation
         5200 NE Elam Young Pkwy, Hillsboro, Oregon 97124 USA
         and
         Derrick Duehren (503) 696-4299, fax (503) 696-4904
         Derrick_Duehren@ccm.jf.intel.com
         Intel Program Manager

Subject: Minutes from IBIS Open Forum Meeting 10/28/94

To:
AT&T Global Info Systems Dave Moxley*
Anacad Steffen Rochel
Ansoft Henri Maramis
Atmel Corporation Dan Terry
Cadence Design Sandeep Khanna, C. Kumar*
Cadlab Ralf Bruning
Contec Dileep Divkar*
Digital Equipment Corp. Barry Katz*
EIA Patty Rusher
High Design Technology Michael Smith, Dr. Ing. Cosso
HyperLynx Kellee Crisafulli*
IBM Jay Diepenbrock, Joseph Flanigan
IBM-Motorola alliance Lynn Warriner, Hoa Quoc, John Burnett
INCASES Werner Rissiek, Olaf Rethmeier*
Integrated Silicon Systems Eric Bracken*
Integrity Engineering Greg Doyle
Intel Corporation Stephen Peters*, Don Telian, Will Hobbs*
                              Arpad Muranyi*, Derrick Duehren
Interconnectix, Inc. Bob Ross*, Scott Bloom
Intergraph Ian Dodd, David Wiens
IntuSoft Charles Hymowitz
Mentor Graphics Ravender Goyal
Meta-Software Mei Wong, You-Pang Wei, John Sliney
NEC Hiroshi Matsumoto
MicroSim Arthur Wong
National Semiconductor Syed Huq
North Carolina State U. Steve Lipa, Michael Steer
OptEM Engineering, Inc. Benny Leveille, Ken Ehn
PC Ware Paul Munsey*, Ron Neville*
Quad Design Jon Powell
Quantic Labs Mike Ventham
Racal-Redac John Berrie
Stanford University Randy Harr
Stree Co. Mon Seng(sp?)
Symmetry Martin Walker
Synopsys, Logic Modeling G. Bill Lattin
Texas Instruments Bob Ward
Thomson-CSF/SCTF Jean Lebrun
UniCAD Canada Ltd. Stephen Lum
Zeelan Technology George Opsahl*, Hiro Moriyasu

CC:
Intel Corporation Randy Wilhelm, Jerry Budelman,
                              Intel IBIS team

In the list above, attendees at the meeting are indicated by *.

Upcoming Meetings: The room and bridge numbers for future IBIS teleconferences
are listed below:
     Date Bridge Number Reservation #
     12/9/94 (916) 356-9999 406224 ** 2.1 RATIFICATION! **

All meetings are 8:00 AM to 10:00 AM PDT (16:00 to 18:00 UTC). 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.

NOTE: "AR" = Action Required.

------------------------------------------------------------------------------

Many thanks to Bob Ross for taking the notes for this set of meeting minutes.
                 :-)

Check-in, Intros, Announcements
Dileep Divekar is now representing Contec.

There were no corrections made to last month's minutes.

To date eight companies have paid for the Version 2.0. We need 10 to pay fully
for the Golden Parser development work.

New Agenda Items:
IBIS Information Technical Net
Relocating L_fixture
Update Cookbook and Overview documents
IBIS model revision level numbering system
(Later) Phase in of IBIS feature priorities.

Press Updates:
Will and Derrick have an outstanding AR to draft a press release with current
status. No progress made has been made, but they plan to have a draft ready
for discussion/approval at the 12/9/94 meeting.

Progress toward enlisting new IC vendors
Kellee has provided Xilinx with a set of models. Xilinx has agreed to release
models that they develop to the public domain. IBIS Model generation activity
is reported at Triquint Semiconductor and IBM. Bob Ross provided feedback to
Syed at National on Differential Models.

Golden Parser, 2.0
Paul and Ron both feel that the parser is close to completion.

The Beta Version "b" of 2.0 parser has been distributed by Will. Everyone
reported some difficulty in getting it compiled on non-DOS systems and there is
possibly one file name error in an INCLUDE statement (s.tmp_name should have
been s_temp file_name). Dave Moxley got it to compile on HP. Bob Ross
compiled it on Sun, it works fine. Stephen hasn't succeeded in compiling it,
and hasn't spent much time on it. Barry is in the same space-- typedefs and
include file problems.

AR - Everyone report to Paul and Ron any source code errors preventing
compilation. The Email address is in the README file.

Paul and Ron are getting an updated UNIX system and may be able to provide a
make file for future releases. Bob Ross and Kellee report old problems have
been fixed and no new problems have appeared. However, a Monotonicity warning
message is now printed for every line for a non-monotonic table. One message
per table is what was intended. Paul and Ron intend to fix this.

Kellee re-ran everything he had before, and everything that was broken was
fixed.

The new IBISCHK2 still did not work on the large IBIS file test case on a
DOS386 machine. However, Paul indicated that his revised code now works. The
issue of sizing was discussed and a number of large file test cases were
proposed for DOS systems with large files. Ideally, we should test the
following cases with ibis_chk on:

   100 - 20pin components, 1 model per component
   1 - 1000 pin component ....

The DEC Alpha chip has about 500 pins. Paul is doing these sizing tests.

It is reasonable to assume that components will have 1000-2000 pins, with 25
models in 5 years in the future.

Possible memory extension software was discussed, especially from Pfar Lap and
Rational. Cost of Phar Lap is a problem for free distribution. Watcom has an
extender, but it isn't too good. 5% of machines have this problem. Kellee can
compile a Win32s version that does not have memory problems. There are two
restrictions on win32-- need windows and win32s would have to be distributed
with the parser. It is free but is 1.5 meg, implying multiple disks.

Memory limitations could be solved on a company-by-company basis as needed.
Large IBIS design files can be partitioned into smaller files as a last resort.

A minimum machine was discussed. Code will work on 8088 - DOS systems, but
386, 4M, DOS 5.0 was presented as an acceptable minimum level. Some people
have less than 4M of memory on home systems.

AR - All. Stephen wants more people to do testing based on the Test Matrix,
especially in untested areas. The package model area has limited testing.

Stephen suggests that we should have a couple of tests: One 500-pin model with
25 models, and one with 100 separate components in a file. Kellee would like
to see a 100-pin model.

$ for additional features.
Monotonicity and incorporating new BIRDs were discussed in the context that the
IBISCHK2 project is growing in size. Paul and Ron indicated that they are
comfortable with the current changes, although they have taken more time than
anticipated. Some pending changes include:
o Monotonicity - already fixed (not sure if he implemented it correctly)
o DOS and UNIX file compatibility (control-M) Paul and Ron will investigate
  and report on impact of fix. (may be small, but could grow)
o Case sensitivity - not done, but may take time going though 50 files. This
  BIRD was later rejected.

Paul agreed to estimate the cost of implementing new BIRDs.

The Version number of the Parser appears as the first line. We would like the
first line to indicate this is a "BETA" version of the parser so that it will
not be confused with the final version.

AR - Arpad to notify Paul of first-line change.

AR - Intel and National. What are critical the features needed by vendors in
CAE systems supporting Version 2.1? Post them to the reflector before the
12/9/94 meeting.

BIRD20, Monotonicity Clarification and Error Reporting
APPROVED, but with the text revised in the last two paragraphs as shown:

   An IBIS syntax checking program shall test for non-monotonic data and
   provide a maximum of one note per V/I table if non-monotomic data is found.
   For example:
      "NOTE: Line 300, Pulldown V/I table for model DC040403 is
      non-monotonic! Most simulators will filter this data to
      remove the non-monotonic data." *

   It is also recognized that the data may be monotonic if currents from
   both the output stage and the clamp diode are added together as most
   simulators do. To limit the complexity of the IBIS Version 2.x syntax
   checking programs, such programs will conduct Monotonicity testing only on
   one V/I table at a time.

BIRD21, Minimum of 2 Entries in Waveform tables.
APPROVED.

AR - Bob Ross. Generate BIRD23 bringing language in line with V/I table minimum
of 2-entry limitations where the first and last entry must not be an "NA" in
the typical column.

BIRD23 approval/ratification is expected 12/9/94.

BIRD22, Subparameters not Case Insensitive.
REJECTED.

This is a good change for some future version, but making the case
insensitivity change would present a significant risk of error to bring the
IBISCHK2 parser in line with the specification because nearly all, if not all,
50 ibis_chk source files would have to be edited. The parser itself can be
used effectively to identify and correct any case sensitivity violations, so
while case sensitivity is an annoyance, it does not pose any limitation.

Note that BIRD16 was also killed/rejected.

IBIS General Session (face-to-face)
Will had earlier proposed that we have our next face-to-face meeting in early
December in San Jose to cover EIA affiliation and technology presentations.
After some discussion and realizing conflicts with vacation dates, the Summit
meeting will be scheduled in January.

We discussed the decoupling of Version 2.1 ratification from EIA specification
activities. The DIE committee meeting will be held Thursday and Friday,
December 15 and 16. Their agenda includes EIA affiliation. The probability is
low that the IBIS/EIA processes could go quickly enough to qualify for any ARPA
funding, even with a December meeting.

AR - Derrick arrange date/time/place for January meeting.

Possible technical sessions were discussed since some very interesting material
was presented last year. Possible topics include some work by Arpad on
equation-based formulation of min and max tables and also on some transmission
line measurement techniques, Stephen on alternative package models, and some
re-presentation of last year's sessions.

AR - All please submit to Will and Derrick:
     o What you would like to hear about
     o If you are considering making a presentation, what the talk is about
       and how long

L_fixture location:
L_fixture location, originally raised by Kumar, was discussed. General
consensus was that there was not a compelling reason to change IBIS at this
time. In fact, there may be some advantage with the current location regarding
simplified, distributed line model of fixture parasitics. The appropriate
mechanism for continuing this issue with the intention of modifying the
specification would be through the BIRD process.

Technical Information Network
Will is getting a lot of detailed, technical questions rather than just the
requests for IBIS information, and needs a process to respond. One process
suggested was to forward the questions to the IBIS reflector for anyone
interested to provide a response. New issues and concerns would then become
visible. Will also noted that IBIS has been discussed on the Signal Integrity
reflector, and it was suggested that questions could also be forwarded there.

Wrap-up, Next Meeting Plans
Our next meeting is 12/9/94. The major agenda item is ratifying IBIS Version
2.1.

AR - Bob Ross/Derrick, Send to Reflector and also Post on vhdl.org a Version
2.1 DRAFT containing all of the BIRDs resolved to Date. [DONE]

Also, on the agenda: Cookbook and Overview update, EIA affiliation, press
release (an old AR of Will/Derrick), naming structure, and revision levels.

AR - Arpad, submit a raw EGG on revision levels. (Note that IBIS 2.0 already
contains a recommendation.)

==============================================================================
                                      NOTES
If you know of someone new who wants to join the e-mail reflector
(ibis@vhdl.org), send e-mail to 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.
==============================================================================
Received on Fri Dec 2 14:46:27 1994