I have recently started developing code for a dot matrix display. Initially this occured with the IAR compiler and a MCS912DJ256B processor but have had to upgrade to the MCS12XD64 and codewarrior as the project progressed. In doing so, I have noticed that the same code that originally worked for generating PWM signals from the compare module (using IAR and DJ2566) went from 622us per period to 3.48ms per period (on XD64 using CodeWarrior)
I coded the original version based using the datasheet for the XD64 processor as I knew I'd eventually have to upgrade and this worked fine with IAR so cannot see how things are not working the same now.
To my knowledge I've written all the registers correctly and have double checked prescaler values etc. (no prescalers have been used and my 16MHz external clock is being divided down to having 4.9us instruction cycles against the expected 125ns.
Can anyone suggest something which I may have missed out that could account for this drastic change of execution speed?