Optimizing CRC16 routine.

[HenrikOlsson]

I am pretty sure I put the shift inside the loop

Absolutely, but shifting before checking bit0 is different from checking bit0 before shifting ;-)

I haven't yet tried Darrels version but I have looked at the ASM code it compiles to and as far as I can see it's now "longer" than what my current code compiles to. I'm actually starting to doubt it'll do the 9 byte test sequence in 825 cycles. Darrel, are you there?

Thanks!
/Henrik.

[Bruce]

Run Darrels' example through MPSIM with the stopwatch to see if it really takes only 825 cycles to process 9 bytes through his CRC routine.

Yours looks pretty darn tight as-is.