event_handler_performance.txt

performance data collected: 

(this is for the tcs io board, with the restricted set of inputs/outputs
enabled: 40 outputs and about 15 inputs)


FOR COMPILATION WITH -Os
========================

processing 20 events of 0100..0104 (which have 8 matches in the registry)
~40 msec  (37-40 msec)
for 100:
174 or 212 msec (there is probably a rounding problem here, freertos cannot really measure sub-millisecond time)


processing 20 events that have zero matches
~2 msec (0-3 msec)
for 100 irrelevant events: 4-10 msec


processing events with one match in the registry:
100-140 msec for 100 events.



FOR COMPILATION WITH -O3
========================

processing events with 8 match in the registry:
160-200 msec for 100 events.

processing events with one match in the registry:
100-125 msec for 100 events.

processing events with zero match in the registry:
1-12 msec for 100 events.


AFTER FIRST ROUND OF OPTIMIZATION
=================================


processing events with 8 match in the registry:  (0100)
24-34 msec for 100 events.

processing events with one match in the registry: 
12-20 msec for 100 events.

processing events with zero match in the registry:
2-8 msec for 100 events.