Re: [vhdl-200x] RE: Modular types

Le 2014-07-24 11:20, Martin.J Thompson a écrit :
> Hi all, and thanks to Jim for the reminder J
> 
> I've summarised the past discussions on the implementation of a
> modular type. You can find a proposal here:
> 
> http://www.eda-twiki.org/cgi-bin/view.cgi/P1076/ModularTypes [1]

Thanks ! I added my 2 cents (I hope the formatting is correct)

> A number of questions are still outstanding…
> 
> First and foremost:
> 
> Do we implement as a "new type", or by "extending the operation of
> resolution functions"?
> 
> Andy, I'm not sure I've grasped all the implications of your proposed
> solution through the use of resolution functions. In particular -
> would it be possible to perform logical operations on modular types
> merely by overloading the resolution function?

I have no idea. My understanding of VHDL is too shallow for this so I 
won't dive.

> Also,
> 
> * Should logical, shift, etc. operators be allowed on
> non-power-of-2-modulus?

well, why not ?
As long as the operators are understood as base-2
AND the wrap-around is correctly applied afterwards,
it "should work" if it's needed.

> * Should assignment operate like Ada or in what has been described as
> "the expected way"?

I have not followed or understood this discussion,
what are those two ways ?

> * Should modular types *have* to go from 0 to something? Or would -1
> to 6 be acceptable? And would it allow logicals, shifts and rotates?

1) ideally, I see the modular type just like an integer with a 
user-defined range,
    so ranges that don't start at 0 are not ruled out, and could even be 
useful.
    It might be less practical to map/synthesise (adds protection 
circuits to preserve
    the range or representation) but i see no reason against.
2) see the previous question and answer : why not ? :-)

> Comments welcome!

as well as feedback on my edits on the twiki page :-)

> Cheers,
>  Martin

yg

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