Using Processor Expert Correctly

[FC]

I would like to know who use Processor Expert deeply in their embedded projects?  I just use the templates for interrupts and start up code.  It would seem to easy to use the tool for, lets say, SCI routines because the underlying hardware does not have to be thoroughly understood.

[alexod]

That is the point.  Processor Expert provides library functions for each flavour of CPU, and through point and click and form filling it makes life a lot easier.

I've done it the hard way on Motorola 1993-2000, and Cradle 2000-2007.  On returning to Freescale last year I found that beans made it easy to configure clock rates, interface speeds etc without having to know every bit in every register, and without having to know how the clock is distributed. It even manages clock dividers at different main clock speeds for low power sleep and doze modes.

I found the real power in a quick bring up of a board with a CPU which was new to me.  Usinging their get bit set bit functions and macros I could easily write more generic code above it.  Then, when 2 months later I was given a different board with a different clock and different CPU, I just had to create the beans on the physical IO I wanted, and it all magically worked.  I ported code in a couple of hours instead of a couple of days.

I am still using the STD Edn, so I have to code my own flash and fifo's, but I'm thinking that maybe $4,000 per CPU family would not be a bad investment if everything becomes hours instead of days.  The up front cost for fast cross platform development allows engineers to focus on applications, not low level drivers, and the reduced time to market means you can be first to market, and keep ahead of the competition.

You do still have to understand the hardware to do it all well, but you don't get bogged down in the nitty gritty.