Verilog-AMS meeting minutes - 22 Sept 2010

  Date: 22 Sept 2010

Geoffrey Coram, Analog Device
Ian Wilson, Berkeley DA
Achim Bauer, TI
Ken Bakalar, Mentor
Graham Helwig, ASTC
Sri Chandra, Freescale
David Miller, Freescale

* Query on merging of SV and Verilog-AMS
   - Graham explained that he has done a very simple integration of
currently existing P1800 grammar with Verilog-AMS. Wherever there is
commonality merge the two BNFs.
   - No effort has yet been spent on actually looking at what
digital-only aspects of SV needs to be available within AMS at this point.
   - Query on interfaces; should it include continuous discipline to the
ports of the interface. This has not been looked into

* Ken raised the issue ASVA requirements and the fact that syntax might
change based on ASVA discussions
   - Currently discussions within ASVA committee are mainly on
mathematical solution; the committee is not yet looking at the syntax
(embedding problem)
   - The current merge done by Graham does not consider any ASVA
requirements
   - Will ASVA committee provide specific requirements for syntax
changes based on assertion enhancements to AMS committee (or) will they
also provide necessary changes to SV-AMS grammar?
   - Once ASVA requirements are clear and finalized need to get this
issue resolved (either specific requirements or affected BNF grammar
changes from the sub-committee)

* Discussion on genvar_iteration
   - can the genvar_iteration from SV be reused (current merger reuses
existing SV syntax)
   - Graham feels the SV BNF is a superset; need to confirm that it
works in the AMS context.
   - does it work?

* Note on cross keyword
   - cross is a keyword in SV as well as AMS tho used in different
productions without conflict.
   - Cross is used within SV when defining function coverage groups.

* Concatenation discussion
   - analog concatenation: this does not seem to be used in Verilog-AMS
at all. It is defined in BNF in Verilog-AMS LRM 2.3.1 but there does not
seem to be any reference/usage of this grammar within the language.
   - array literal vs concats: there does not seem to be a specific need
for analog_concatenation and seems like array literals can be used
instead (analog_concatenation was initially written specifically, so
that it contains only analog_expressions)

Regards,
Sri

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.