nct: Simplify and improve switching performance

Simplify the code and improve the NCT switching performance.
How much the performance improves depends on the use case:
The more threads in a NCT instance the better the performance
improvement.

For 10 threads an improvement around 3x can be expected in the NCT
code itself.

Signed-off-by: Alberto Escolar Piedras <[email protected]>

Author: aescolar