Just how much does is slow down your "modern technology" device if, once every week it goes out and hits the external memory bus 24 times and generates a 240 ns cycle? The rest of the time it's running its internal cycles at 10 ns system cycle. Most of the MCU's in this family operate at 50 rather than 100 MHz, and, based on what you've said in THIS context, I doubt very much that you've ever taken time to determine whether you even need that last level of performance. I don't doubt you can do the arithmetic, since any 10-year-old can do it, but I do, based on your inane remarks, believe you've just said, "give me the fastest MCU that's avaiable, and maybe it will work." If that weren't the case, you'd already have known (a) that the F12x-series supports an external bus, and (b) that the external bus cycles can be extended to match the peripheral that's out there.

I don't believe you can come up with a circuit that operates reliably with your favorite MCU, even with HC logic without slowing down the MCU. Further, I doubt very much that YOU can come up with a circuit that provides the functions of an 8255, regardless of what sort of logic you have at your disposal. You seemingly can't even do the arithmetic that reveals how much a stretched bus cycle impacts the rate at which code is executed.

Anyone willing to do the simple arithmetic can make the 8255 work with the F12x-series of SiLabs parts without slowing down anything other than the external bus cycles that "talk" to the 8255's that are attached. Anyone except you, of course, since, in your case, it's just a matter of having your head in the sand. Why do you suppose the SiLabs guys provided these functions that extend the external bus cycles without slowing down the processor for any other cycles?

You're just demonstrating that you're prejudiced against what's OLD. Believe me, you'll get there soon enough!

RE