Some final thoughts

Instead of conclusion...

It's relatively easy to design a fast 8 bit controller. All that has to be done is to follow the path well known from the big brothers, the 32 bit controllers. The short story is: analyze what "typical programs" mean, imagine a simple and fast instruction set, and implement it into a deep pipeline (by the way, for this topics, I recommend you "Computer architecture - a quantitative approach", by J. Hennessy and D. Patterson).

Then, why are the 8 bit controllers currently on the market so slow? The instruction set, CPI, max frequency for current 8 bit ucs are bad. In fact, they are so bad, that we must consider other factors than pure uc design to explain that. My guess is that market issues distructively interfere here. How is that, this could be another project's goal...

Now, looking back, I realize I should have cloned ARM7 instead. That should be a neat and easy design, because its instruction set is orthogonal and built with the pipeline concept in mind, unlike AVR. And it would have been more useful for the open source comunity.

Note

This documentation and a good part of the pAVR project itself are the result of the constructive criticism I received. I modifed the whole structure from bottom up to make it easier to digest. Open sourcing rules ;-) ...

Generated on Thu Mar 3 00:16:27 2005 for Pipelined AVR microcontroller (pAVR) by  1.3.9.1