minutes Verilog-AMS Committee Meeting - 19 April 2007

From: Marq Kole <marq.kole_at_.....>
Date: Fri Apr 20 2007 - 05:06:57 PDT

All,

Here are the meeting minutes of the Verilog-AMS Committee Meeting of 19 April 2007

present:
Martin O'Leary - Cadence
Patrick O'Halloran - Tiburon
Geoffrey Coram - Analog Devices
Dave Miller - Freescale
Graham Helwig - FSTC
Marq Kole - NXP

agenda:
- Updated chapter 7 based on the multiple analog block that has been proposed and discussed (Marq Kole)

minutes:
An update (version 4) of the Multiple Analog Blcoks document was posted on the Verilog-AMS website, containing updates to the previous version as discussed during last weeks conference call.

After initially discussing the changes that had to be made to the draft LRM 2.3 for multiple analog blocks there was a growing unease amidst the participants about the support of concurrent analog blocks. It was felt that there were too many unknowns still, in particular regarding the concurrent assignment and variable acces for module-level variables in multiple concurrent analog blocks. There was a consensus that multiple concurrent analog blocks is a step too far for the current LRM.

Instead it was considered that the main purpose for multiple analog blocks in LRM 2.3 is to support the new generate construct and that it would be best to state that the behavior of multiple analog blocks would be defined by assuming that they had been concatenated in the order they appear in the module. This will still allow implementors to support concurrent evaluation as long as the behavior in that case is similar to what would happen if they were concatenated.

The reason was actually that the multiple analog blocks discussion document showed many areas that would be problematic to get consistent behavior for concurrent analog blocks. By defining the behavior in this way, the generate construct will work; there will be no limitations on things that can be modelled in the current language. The members present in the call felt that we should be proceeding cautiously. The driver for concurrent analog blocks will most likely come from the ability to offer improved performance based on sensitivity lists, but that area is still too unclear to base a standard on right now.

As it is the impact of multiple analog blocks on the LRM text will be quite modest.

- based on the discussion results and the decisions taken, the following will now be done:
  * Marq will make a new update of Section 7 of the LRM 2.3 to reflect the decisions, to be discussed next week.
  * Marq will propose changes for a new update of Section 6 of the LRM 2.3 to the section owner (Sri) to reflect the decisions, to be discussed next week.

For the next call it was suggested by Geoffrey that we discuss the examples on the Verilog-AMS website.

Changes to the descriptions of the atan2 and mod built-in math functions as reported by Paul Floyd will be made to the draft section 4 (merged_expr) document.

Geoffrey will update section 3 (merged_datatype) with the string variable as discussed in the review of section 11.

next call:
April 26, 2007, regular call times

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean. Received on Fri Apr 20 05:05:38 2007

This archive was generated by hypermail 2.1.8 : Fri Apr 20 2007 - 05:06:09 PDT