However, a DMAC was not even on the ultra-long list, let alone the short list, of things I hope to examine.

There are lots of possibilities, but hobbling the entire device by adding something that doubles or triples the propagation delays into memory is not something I think will help anything.

Now, speeding up execution is a different story. Getting rid of the ultra-lame serial channel to counter-timer link, particularly by severing the link to the processor clock, is much higher on the list.

I also don't see any justification for preserving the SFR for a peripheral that's no longer there.

As I've said, it's a work in progress, and lots of things are under consideration, but DMA is not. After all, you have only 256 bytes of IRAM, and only half of it is drectly accessible. It would be a mess for the 1 ppm of applications for such a thing.

A super-805x is not really what I want, but it could develop. If it's fast enough, there should never be a need for a DMAC.

RE