[Coco] A faster "Real" CoCo Re: General Memory Question about speed

[jdaggett at gate.net]

On 15 Jan 2008 at 12:40, Mark McDougall wrote:

> jdaggett at gate.net wrote:
> 
> > Yes one could be based on that. There are several hurdles to
> > overcome in doing so. One would be do we base it on a given FPGA
> > development board or go for a dedicated board something like the
> > C-One. 
> 
> You can go a *long* way with the design before having to decide on
> target hardware. There are an increasing number of development boards
> with compatible resources which would serve as a prototype platform
> for any "Coco4" project.
> 
> I have successfully ported a number of projects to several different
> development boards with minimal changes. Perhaps the best choice to
> make up-front is the FPGA vendor (Altera/Xilinx). That simplifies the
> work somewhat, especially where FPGA memories and timing constraints
> are involved. The PLL/DCM is not so much of a problem.
> 
> Using a single generic 'core' the design can accommodate several
> hardware platforms using a top-level wrapper in conjunction with
> generics/constants if required. In fact, I find it even promotes
> better design practices...
> 
> Regards,

I agree in that a good start can be done prior to any board commital, development or 
dedicated. Also which vendor of FPGA? Still what constitutes a new Coco can be the work 
of one or a consensus of many. That is to be determined.

What I have started is a dedicated use of a Coco FPGA. More  a control application and 
embedded. Though peices maybe usable in other projects. I am not totally sure my coding 
style is "good general practice" but if it accomplishes what I want, then it will be acceptable 
to me. I am by trade a  hardware engineer and not software. So I think in hardware terms 
and use HDL as means of describing hardware. 

 
james