Re: [myhdl-list] simulation semantics

[Geoffrey B. <geo...@in...>]

Sorry, Moore should be Mealy

On Thu, Jul 16, 2009 at 10:17 AM, Geoffrey Brown <geo...@in...>wrote:

> I've been creating little examples to see if I understand the semantics of
> myhdl.
>  I came across one where the behavior differs completely depending
> upon the initial value of a signal (no problem there), but the generated
> verilog
> is identical  -- see the attached and consider two alternatives:
>
>     t = Signal(bool(not state))
>     t = Signal(bool(0))
>
> In the first case, the simulation does what I expected -- the state
> alternates while in
> the second case its frozen at its initial value.  I can understand why this
> might happen,
> I'm just not sure it should.  The fundamental issue has to do with
> initialization of the
> simulation.   In the second case, the state is inconsistent with the
> combinational
> logic, but there is no event to trigger the necessary change.
>
> I was trying to understand how a Moore machine might be coded
> and simulated in myhdl.   I know this isn't always the design pattern of
> choice,
> but if you want to implement the handshake of communicating processes, this
> is what falls out.
>
> Geoffrey Brown
>
>