[ibis-macro] Minutes for the 21 mar 2006 ibis-macro meeting

  • From: "Mike LaBonte (milabont)" <milabont@xxxxxxxxx>
  • To: <ibis-macro@xxxxxxxxxxxxx>
  • Date: Thu, 23 Mar 2006 10:53:22 -0500

Minutes for the 21 mar 2006 ibis-macro meeting are attached.

Mike
Meeting date: 21 mar 2006
Members (asterisk for those attending):
Arpad Murayni,
*Bob Ross,
*Todd Westerhoff,
*Mike LaBonte,
Paul Fernando,
Barry Katz,
*Walter Katz,
Ken Willis,
Ian Dodd
*Lance Wang

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

Mike make the website more outsider-friendly
- done

Todd contact Cadence about a new representative
- done

Mike finish documentation examples.
- in progress

Todd contact Ian
- done, waiting for reply

Arpad make buffers scalable
- done

Walter contact Gary Pratt about model data
- done, waiting for reply

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

How to enhance vendor participation
- There are plenty of new designs out there
  - New serdes buffer capabilities every month
  - Driver schedule won't do it, AMS needed
    - Need to implement state machines
- First step is to implement some of Gary Pratt's designs
  - Will be a test of ibis-macro
  - Probably written in VHDL-AMS
- Advanced buffers simulate VERY slowly in HSPICE
  - 1 day
  - 1000 states

Is there an IP protection problem?
- Even the control inputs might reveal IP
- Vendors might simulate HSPICE for us
- If NDA is required, this task group will probably be out in the cold.
- Who will make the macromodel, us or the vendor?

BIRD95 prototype macromodel
- Who could test this?
  - Sigrity would have to implement the SPICE equivalence library
  - Cadence has a gate modulation macromodel in KSPICE
    - Zhiping Yang wrote a BIRD95 model in HSPICE
    - Cadence has not been supporting BIRD95 as specified
    - Cadence has a RocketIO macromodel, but it is not public
    - Cadence has a public PCI-X macromodel (by request)
- Key is to get I/T curve data
  - Watch out: common SPICE models use node 0
  - But this should not affect power lead measrurement

AR: Mike find downloadable SPICE model and measure I/T

SiSoft may be able to do a serdes in ISPICE
- Either from Gary Pratt or a secret customer
- Either PCI-X or RocketIO

AR: Lance send PCI-X macromodel
AR: Walter test PCI-X macromodel in HSPICE and/or ISPICE

Walter will be unable to join the 28 mar 2006 meeting.

-------------
Next meeting: Tuesday 28 mar 2006 12:00pm PT

Other related posts: