Hello all,
first of all thx for your help... 
Yesterday I did some extensive research and measurements with an oscilloscope...
I found out, that the digital read operations take a insignificant amount of processing time…
Serial communication on the other hand take more time – I am sending one character to the Windows application after each reference point, to visualize the progress with a progress bar.
My oscilloscope is an old one and not very accurate, but sending one character with 57k baud takes about 100-200µS.
However I found another issue.
Basically I am having my “G-Code” on the SD-Card and when I start the execution of the project, the data on the SD-Card is iterated and the steppers are fed with the data…
Each Second or so, I heard a strange click, which I could not explain for a while… After some calculations and research, I realized that the click happens when each 512th Byte were read from the SD-Card…
With the oscilloscope I found out, that on those spots (each 512th Byte), there were a gap, where the steppers were not triggered. The gap was about 1mS big…
So I guess, that this gap is from the buffering / data loading of the SD-Card.
I don’t know if this has an big issue regarding the performance of the C&C application – in terms of perfection it is not beautiful…
Does anyone have an idea how to make a workaround?
By the way, I will prepare and post the code for the stepper driver here, maybe some of you have an idea how to improve things… (I still have to translate the comments from german to english...)