=========
FCL tests
=========



test_fcl
========

0: sets up a network without learning where every layer
   has bandpass filters at its inputs. An impulse at the
   input generates different timed respones at the outputs:
   test_fcl_cpp_forward.dat

1: Network with one hidden layer and two neurons.
   The stimulus is repeated every 1,000 time steps.
   The input to the network is set to 1 between
   100 and 1000th step and the error between 500 and 600
   and then it repeats:
   test_fcl_cpp_learning.dat
   and can be plotted with:
   plot_learning.py

2: same as 1 but with filters. The impulse reponse is very
   short so the output looks identical to 1.



test_bandpass.cpp
=================

Generates an impulse response and writes it to impulse.dat



test_neuron.cpp
===============

Sets up a neuron with two inputs and sends in a "1" between
timestep 100 and 110.



test_layer.cpp
==============

Tests multi threaded processing in a layer. The output
should be identical with and without threads.


