IBIS Macromodel Task Group

Meeting date: 08 aug 2006

Members (asterisk for those attending):
*Arpad Muranyi, Intel Corp.
*Bob Ross, Teraspeed Consulting Group
Todd Westerhoff, Cisco Systems
*Mike LaBonte, Cisco Systems
Paul Fernando, NCSU
*Barry Katz, SiSoft
Walter Katz, SiSoft
Ken Willis, Cadence Design Systems
Ian Dodd, Mentor Graphics
*Lance Wang, Cadence Design Systems
Kumar, Cadence Design Systems
Richard Ward, Texas Instruments
*Doug White, Cisco Systems
Sanjeev Gupta, Agilent
Joe Abler, IBM
John Shields, Mentor Graphics
*Michael Mirmak, Intel Corp.
*Randy Wolff, Micron Technology

-------------
Review of ARs:

- Ian invite a SystemC expert to speak to us about APIs
  - Ian not present

-------------
New Discussion:

Michael Mirmak said that the IBIS Futures committee has decided to
suspend meetings. They will form smaller ad-hoc groups for Touchstone
issues, etc. He said our group should keep doing what we are doing,
and futures group members would be joining.

Arpad suggested that maybe our group needs a name change, since the
macromodel library deliverable is essential completed and the group
seems to be driving toward a new charter. Suggested names:
- SerDes
- Advanced Technology
- Advanced Modeling

The Futures group met Thursdays, when to meet from now on?
We decided to retain the weekly Tuesday 3-4pm ET time slot.
Arpad suggested we mull this over one more week.

There was some recap of last week's meeting for Arpad's benefit.

Arpad suggested that before we discuss APIs we need to talk about the
compiler switch issue mentioned by Ian Dodd. Ian had suggested delivering
models as uncompiled C code, so that compilation would take place in the
target environment. Another possibility is to use a platform-independent
language such as Java. These tend to be interpreted, avoiding compilation
issues.  Yet, some can be compiled anyway to improve performance.

Lance Wang said that HSPICE has an API interface called CMI. Delivered
software usually supports Linux, Windows, and Solaris, and techniques
for insuring compatibility are well known.  However, delivering source
code is not good protection for IP. Arpad stated that it will be
difficult to support multiple platforms with compiled code.

Ian Dodd had proposed delivering encrypted source code. This avoids
compiler switch problems, but requires a compiler that supports
decryption.

We need more details on existing encryption proposals. Arpad said that
different AMS groups at different stages, but we want something now. He
posed the question: "Will tool vendors drive this or will standards
groups drive it?". Arpad is in the Verilog-AMS work group, but the
open encryption capability is proposed for Verilog (big D).
It was believed that this encryption standard extends to AMS anyway.

Bob Ross said the key question is if Cadence, Mentor, and Synopsys would
adopt these encryption standards. Even if there are tools that know how
to decrypt, how to encrypt? We need a standalone encryptor.
From the related EETimes article:
- Synplicity *may* have a commercial tool.
- If so, Arpad could look into buying it and try it out.

We still have no decision on API vs. AMS. Cadence says an API is
needed. Mentor says AMS will do just fine. More details on the Mentor
AMS approach would help. The Mentor presentation at the recent IBIS
Summit was about a model to perform JEDEC DDR2 measurements. It would
be good if Gary Pratt could present to our group. This might not involve
too much effort on his part.

Arpad mentioned that complete module documentation is the only missing
piece for macromodel library completion. Mike is having trouble finishing
it (priority issue). Arpad said he and others should be able to help.

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

Next meeting: Tuesday 15 Aug 2006 12:00pm PT