> On Fri, 24 Aug 2001, Martijn van Buul wrote:
>
> > I think Bogax means transfers from expansion mem to expansion mem. (I'm
> > not sure the original REU supported this. Actually, I think it didn't)
yup, that's what I meant
>
> It doesn't.
It doesn't support 16MB either (does it?), ie I didn't think you were
talking about strictly cloning the REC or REU, but rather a superset
of it's functionality.
For myself, I was planning on using 72 pin simms (32 bit's wide)
I'm not really familiar with the REU. (I just got one :)
Presummably the REU wraps transfers in memory, so (if that's so)
would you duplicate that in the 16MB space (for compatibility)
ie wrap within 64k banks or 512k banks or what ever the REU does?
>
> > These transfers could be completely internal - no reason to claim the
> > C64's bus, since it's not involved at all. In *that* case, an
> > interrupt could very well be handy.
>
> An interrupt might also be handy when using the comparison operation on
> the normal REU. A data compression application could enable the "verify
> error" interrupt when searching for longest substrings.
>
> > Furthermore, since you're not limited to the 1MHz bus of the C64,
> > you could speedup those transfers.
>
> Or you could provide a control interface to the REU clone to an external
> device. I wonder if the REC registers could be accessed from another
> processor during a block transfer.
>
> Marko
I'm going to bank/page the 16MB address space in to the processor's
address space (8>16)(or maybe 8>24, later)
So it would be possible to DMA, say, a peripheral's buffer to the
processor's address space, but I wasn't expecting it to be necessary.
bogax
Message was sent through the cbm-hackers mailing list
Archive generated by hypermail 2.1.1.