Echelon Neuron C - Manual (Page 221)

Neuron C Programmer’s Guide 209
For example, if a program has two input network variables, and has a single task
executed when either variable is updated, it is more efficient to code it as shown
in the After example, below. Likewise, use of a single when clause with the
nv_update_occurs event referencing just an array name is more efficient than
using multiple when clauses, one for each element of an array.
Before (compiles to 6 bytes of code):
when (nv_update_occurs(var1))
when (nv_update_occurs(var2))
{
...
}
After (compiles to 3 bytes of code):
// Use "unqualified" event to cover all variables
when (nv_update_occurs)
{
...
}
However, if you need to use specific nv_update_occurs events without the use of
the unqualified event shown above, the following guidelines can be used.
Consider a program that declares two network variables nviA and nviB:
network input SNVT_switch nviA, nviB;
The following code fragments are all functionally equivalent, because they all
respond to incoming network variable updates for either of these two network
variables. The first of these implementations is the least efficient of the three,
and the last one (equivalent to the Before example above) is the most efficient of
the three:
Variant 1 (compiles to 15 bytes of code):
when (nv_update_occurs(nviA) || nv_update_occurs(nviB))
{
...
}
Variant 2 (compiles to 9 bytes of code):
when (nv_update_occurs(nviA..nviB))
{
...
}
Variant 3 (compiles to 6 bytes of code):
when (nv_update_occurs(nviA))
when (nv_update_occurs(nviB))
{
...
}
Using the Link Map
Compiling an application for various models of Neuron Chips and Smart
Transceivers generally results in different application footprints, because a