Delay loops seem fairly common in the kernal - or at least timed out
sections of code. The optimisation need here is in code size, not speed!
For 2 cycle delays, costing 1 byte.. use a NOP
For 3 cycle delays, costing 2 bytes, use LDA $80 or similar
Rather than strings of NOPs, you can use the following...
pha
pla
Pairs of pha/pla cost only two bytes, yet use up 7 cycles each pair.
For longer delays. put a label on one of your RTS instructions
return RTS
now, whenever you want a 12 cycle delay you can go...
JSR return
that will only cost you 3 bytes.
Using combinations of the above lets you maximise cycle time, and minimise
byte usage.
Enjoy!
A
--
Archives (includes files) at http://www.biglist.com/lists/stella/archives/
Unsub & more at http://www.biglist.com/lists/stella/stella.html
