All of this is utterly irrelevant when compared to the OS scheduling
quantum (typically between 10mS to 1mS), assuming the OP is running the
test systems on regular machines and not some fancy realtime bit of
hardware.
For the most part, if he had told us what the circumstance was, any one
of us could have delivered a more appropriate response. In one view, one
person and two mice could be 'clicked' 'simultaneously', FAIAP. In
another, if he really needs more simultaneous synchronization than he
himself could bring with two mice, then reliance on raw synchronization
that falls inside your timing region may not be enough either. So, yeah,
more would be needed.
Generally the lay person would not see that as a difference, and if his
application requires time stamps or other real time function, he would
need something more appropriate, like two systems integrated into a
chassis, being triggered by a third device installed into each of those
two systems. Even then there can be latencies between pieces of gear,
which have to be 'calibrated' against in the individual components.
This pretty interesting considering that the systems I currently plan
use a 10Mhz source and two GPS fed 10Mhz "switches" that then feed all
the remaining gear in the receiver system, keeping them all synch'd up.
The rest is just standard network switch gear. Fire up the analyzer
though and see that we know how to make the 'eye' look real nice.
Fun piecing together millions of dollars worth of equipment and
watching that and five hundred or so wires and cables come together to
allow a high bandwidth channel to the world, to be placed anywhere in the
world.
