RE: FW: New Mantis Issue #667 for V-1364

From: Steven Sharp (sharp@cadence.com)
Date: Tue Apr 19 2005 - 15:37:11 PDT

  • Next message: Steven Sharp: "Re: [sv-bc] potential command line option"

    Shalom wrote:
    >But I don't think any of this IS specified by the LRM.
    >It does not say anywhere that primitives have internal nets.
    >Primitives are equally thought of as abstract entities,
    >and I can implement it equally as a lookup table without going through
    >internal nets.

    I agree that the LRM does not specify that primitive terminal
    connections work like port connections (and they actually behave
    differently in one visible way, which is also not specified).
    However, if we do not assume that they work like port connections,
    then the LRM does not specify how they work at all.

    In practice, it is clear that they must work pretty much like port
    connections. You can write

    not inv(o, a & b);

    What does it mean to attach an expression to a terminal, if not the
    same thing as for a port? Something has to evaluate that expression
    and must do it combinationally, and there is nothing in the definition
    of a not-gate that will do so.

    >Anyway, can we have some specific new wording to vote on?

    We need to agree on the behavior before proposing wording. I am opposed
    to specifying that primitives unconditionally evaluate at time zero. In
    some situations, this would be incorrect. The rest of the time, it is
    just unnecessary, based on the assumption that terminal connections work
    like port connections. If you don't agree that the LRM is clear enough
    on that, then some additional wording may be needed there also.

    Steven Sharp
    sharp@cadence.com

    This archive was generated by hypermail 2.1.4 : Tue Apr 19 2005 - 15:14:11 PDT and
    sponsored by Boyd Technology, Inc.