[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [f-cpu] "Tree"

On Thu, 10 Jan 2002, Ben Franchuk wrote:

> Juergen Goeritz wrote:
> > Only with that independent level you are really free to
> > move the border between hardware and software during the
> > development process to achieve a maximum performance of
> > the system in design.
> The problem is that both C and hardware HDL's are just shit
> for programing this stuff. Nobody has a good RTL language
> developed. I don't want to have * / in a RTL language if
> I can't specify addc and overflow and carry out from addition
> if I want it. I once had a old yellow book on RTL languages
> ( now lost ) from the days of punched cards. That language
> still was better I think than today's stuff as it was clear
> to read.

I agree completely! When I look at VHDL code it doesn't
hit my eye directly how it works. The same with C in
bigger designs.

I recently found some very old tool that has table driven
input. With this you see directly what the design does -
and the best of all, it really supports 'don't care' X
statements for inputs AND ouputs to provide the full
possibility range to the optimizers. That's especially
great for statemachines, decoders and such stuff.

> I don't like operator overloading because you are
> never sure what operator like '+' is. I would sooner type
> addc(...) than have '+' overloaded.

You are right again! When I recently analyzed that loading
control system (written in C++) it took most of the time
to figure out how they disarranged the normal C operators.
After that it showed up that the system wasn't a great
composition at all - lots of bottle necks. But it was very
complex to read. Probably a new kind of copy protection?


To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/