RE: [sv-bc] e-mail vote: closes Feb 5th

From: Steven Sharp <sharp_at_.....>
Date: Mon Jan 30 2006 - 14:50:53 PST
>From: "Bresticker, Shalom" <shalom.bresticker@intel.com>

>1255: The proposal does not make clear whether UDPs are like gate-level
>primitives or modules. Implementors need to tell us which is correct.
>Both 'ports' and 'terminals' are used with respect to UDPs. We need to
>make the terminology as consistent as possible.

UDP connections work like gate terminals.  They are scalar nets.  The
UDP instance can be declared with a drive strength, which will be
driven onto the output, unmodified by any intervening continuous
assignment.  They cannot be connected by name (even though the UDP
declaration appears to give them names like module ports).  If a
multi-bit expression is connected to an input, it will get reduction-ORed
to a single bit that gets connected, like a gate terminal.  This is
unlike a connection to a scalar module port, where the value will get
truncated.

In every way I can think of that terminals are different from ports,
UDP connections work like terminals.  Are there any other aspects that
you are concerned about that I missed?

I wouldn't be surprised if the term "port" was used in the description
of the UDP declaration itself.  After all, they look a lot like module
port declarations, so the term may have sneaked in.

Steven Sharp
sharp@cadence.com
Received on Mon Jan 30 14:51:14 2006

This archive was generated by hypermail 2.1.8 : Mon Jan 30 2006 - 14:52:03 PST